There are many good reasons for getting excited about highly automated vehicles, or HAVs, which is the acronym used by the National Highway Traffic Safety Administration (NHTSA). HAVs can make driving more fuel- and time-efficient. They can significantly reduce traffic congestion and emissions by driving a precise speed, minimizing lane changes, and maintaining an exact distance to neighboring cars. They can also increase accessibility and mobility for disabled and elderly persons.

There are many good reasons for getting excited about highly automated vehicles, or HAVs, which is the acronym used by the National Highway Traffic Safety Administration (NHTSA). HAVs can make driving more fuel- and time-efficient. They can significantly reduce traffic congestion and emissions by driving a precise speed, minimizing lane changes, and maintaining an exact distance to neighboring cars. They can also increase accessibility and mobility for disabled and elderly persons.

Sharing an HAV instead of owning is projected to dramatically reduce a household’s yearly transportation budget, which currently ranges between approximately $8,000 and $11,000 per car. HAVs carry promises not only in improved road mobility, and accessibility, but also in producing architectural and societal changes that can make mass parking spaces and personal car ownership obsolete in urban areas. Above all, HAVs can help improve road safety by preventing car accidents that cause more than 30,000 deaths/year in the United States alone, cost approximately $230 billion/year in medical and work loss costs, and are caused by humans 90% of the time.

Press articles in the 1950s and 1960s predicted that autonomous cars and “electronic highways” would become widely available by 1975. Major milestones in the use of new sensor, computation, and communication technology have recently reenergized the eagerness for HAVs. This first started with the 2005 “DARPA Grand Challenge”, where four different HAVs designed by teams of engineers from industry and academia completed a 132-mile trip across the Mohave desert in less than 7.5 hours with no human intervention. The 2007 DARPA “Urban Challenge” saw six teams autonomously complete a 60-mile course in an urban environment, while following traffic laws. Most teams used a combination of LiDAR, cameras, differential GPS, and computation power that is multiple orders of magnitude higher than what is typically needed for a commercial passenger vehicle. In 2009, Google (now Waymo) began designing and testing “self-driving” cars, which have since accumulated more than three million miles in autonomous mode.

Currently, most car manufacturers have HAV prototype systems and Google, Uber, NuTonomy have HAV pilot testing programs, including fully autonomous systems for public transportation, which, for now, are confined to segregated lanes and geo-fenced areas. Multiple Tier-2 supplier companies have emerged, which specialize in autonomous car technology. In early 2017, 36 companies were registered to test prototype HAV systems on public roads in the state of California.

However, in Figure 1 (for all figures, see inset photo, above right), Gartner’s “2016 Hype Cycle for Emerging Technologies” shows that HAV technology might be at the “peak of inflated expectations”, approaching the “trough of disillusionment”. Hype cycle curves are non-scientific tools that have been empirically verified for multiple example technologies over many years. Two example emerging technologies, commercial unmanned aircraft systems (UAS) and virtual reality, are included in Figure 1 for illustration purposes. The curve’s time scale may differ for each technology. One of many indicators of decreasing expectations on HAVs include a reduction in press coverage and the emergence of first negative news stories, in particular following the May 2016 crash of a Tesla Model S whose autopilot failed to distinguish a white trailer truck from the bright Florida sky. The Model S ran under the trailer causing its roof to be torn off and the operator to lose its life. The car kept going full speed on the side of the road through two fences until it hit a pole and came to a stop.

In parallel, until the end of 2016, Google was providing detailed reports of their self-driving car performance, which were designed to operate in real-world urban environments. These reports contain records of millions of miles driven autonomously, but also acknowledge “disengagements”, i.e., where the operator needed to take over control to avoid collisions. The data shows that HAVs are much more likely to be involved in collisions, even though these collisions are often of lower severity than in conventional human driving [HAVs typically get rear-ended because of their unusual road behavior] (see B. Schoettle, and M. Sivak, “A Preliminary Analysis of Real-World Crashes Involving Self-Driving Vehicles,” Additional Resources). Also, Uber’s autonomous taxis in Pittsburg have a reported rate of one disengagement per mile autonomously driven.

Moreover, the first fielded autonomous systems have revealed new safety threats. In particular, the technology’s functionality, as perceived by the human operator, does not always match the intended operational domain: for example, there have been cases of highway autopilots being used in urban areas and passing red lights without slowing down. In addition, human-machine interaction is at the heart of role confusion (is the operator or the HAV in charge?) of mode confusion (is the HAV in autonomous or manual mode?) and of the operator’s trust in this multimodal system. Misinterpretation may grow even wilder because a given functionality will not achieve the same level of performance across models and manufacturers, and operators may not be aware of the systems’ independently verified safety ratings. And, within the next few years, operators will be expected to anticipate hazardous situations and take over control. Thus, operating an HAV may require more education and different training than driving a car manually.

Current Safety Assessment Efforts
To focus this article, first consider the Society of Automotive Engineer (SAE) International’s classification of driving autonomy levels in Table 1 (see inset photo, above right). Under Levels 0 to 2, the human driver is responsible at all times, either for driving by himself, or for supervising the HAV in autonomous mode and taking control if needed. Under Levels 3 to 5, the system is self-monitoring and the driver is expected to take control, but only if requested by the system. Levels 0-4 provide partial automation under predefined driving modes and circumstances, whereas Level 5 is full autonomy.

The most advanced private car systems are currently Level 2, and pilot programs aim at achieving Level 3, although the mere presence of a kill-switch would imply that the system is actually Level 2. The transition from Level 2 to 3 is a remarkable leap that has significant implications on trust and comfort of human-machine interactions, on legal responsibility allocation between system and driver, and on technical challenges to overcome to guarantee passenger safety.

Over the past four years, the most publicized approaches to demonstrate Level 2 HAV safety have been experimental testing campaigns by Google, Tesla and Uber. Google’s approach to have HAVs drive millions of miles with minimal human intervention has been documented up until 2015. At this time, Google cars have autonomously travelled an impressive three million miles. Tesla’s autopilot is reported to have driven more than 130 million miles – on highways only – before it caused a fatality in May 2016.

In parallel, NHTSA reports about 3,000 billion miles travelled each year on U.S. highways by human drivers, with 30,000 deaths caused by traffic accidents; this corresponds to about one fatality in traffic accidents per 100 million miles driven in the U.S. But, this number accounts for incidents on all roads, in all weather conditions, and for all vehicle ages and types. Thus, a purely experimental, complete proof that HAVs match the level of safety of human driving would take about 400 years at Google’s current testing rate (of approximately 250,000 test miles per year), and would still take many decades if the testing rate increased exponentially. This is assuming that no fatalities occur during that time, that no major HAV upgrade is performed, and that the testing environment is representative of all U.S. roads. Thus, while an experimental proof is conclusive, it is not practical. Other, analytical, methods must be employed to ensure HAV safety.

Research Challenges In HAV Navigation Safety
Multiple technical aspects developed over decades for automated flying could serve as starting points for automated driving systems. Figure 2 shows research areas with overlap between aircraft (in blue) and car (in yellow) applications. Figure 2 is not intended to give a comprehensive list of all aspects of automation, but instead, it shows example technical areas that can be addressed using similar methods in aviation and automotive applications (in the green area). For example:

• performance standards set for software, communication, and electronic equipment are already being compared for aircraft versus cars in the NHTSA report by Q. D. Van Eikema Hommes, Additional Resources.
• the design of aircraft cockpit has been continuously improved over the past few decades, especially for highly-automated Unmanned Air Systems (UAS) with a remote pilot “in-the-box”; few car manufacturers envision futuristic car interiors where humans do not participate in driving, but as long as human-machine interactions are needed, lessons learned in cockpit design to avoid information overload are key. 
• while Automatic Dependent Surveillance-Broadcast (ADS-B) will be mandatory on all aircraft by 2020, a petition for proposed rule making has been issued to mandate Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) by the same date. (ADS-B is a situational awareness system for collision avoidance, through which aircraft share their positions with Air Traffic Control and with other aircraft.) 
• GNSS/INS navigation systems, which are extensively used in safety-critical aircraft navigation, are also being investigated for HAVs.
• overall safety standards also have similarities for aircraft and HAVs, which are discussed again below.

The focus of this article is on navigation safety. In aviation navigation, safety is assessed in terms of integrity (as well as accuracy, continuity, and availability, which are not discussed for brevity). Integrity is a measure of trust in sensor information: integrity risk is the probability of undetected sensor errors causing unacceptably large positioning uncertainty (See RTCA Special Committee 159, “Minimum Aviation System Performance Standards for the Local Area Augmentation System (LAAS), Additional Resources”). This top-level quantifiable performance metric is sensor- and platform-independent, and can thus be used to set certifiable requirements on individual system components to achieve and prove an overall level of safety.

The multiple separate efforts towards achieving Levels 3-to-5 HAVs reveal a compelling lack of coordination towards a common, uniform, quantifiable safety goal. Integrity can be used as an objective performance metric for open, transparent comparison and categorization across manufacturers. It can also provide a governmental regulating agency performance and testing standards for HAV certification, which would help accelerate the development, growth, and maturation of such HAVs, as displayed in Figure 3.

Moreover, the Federal Aviation Administration (FAA) has developed analytical methods to evaluate integrity. This provides the means to:

• quantify safety of existing multi-sensor systems under a variety of operating environments, thereby reducing the need for experimental testing
• allocate safety requirements to individual system components to achieve an overall target level of safety, thereby enabling design for safety 
• perform risk prediction, which is a key operational feature to enable hazard avoidance maneuvers

Several methods have been established to predict the integrity risk in GNSS-based aviation applications, which are instrumental in ensuring the safety of pilots and crew. As an example, Figure 4 illustrates a simplified definition of the integrity risk for aircraft landing applications. The aircraft positioning prediction is uncertain because of sensor measurement noise. An alert limit (AL) requirement box is represented around the predicted aircraft position. This AL is set by the certification authority, i.e., by the FAA in this application. Simply put, the risk of the actual aircraft position being outside the AL box is the integrity risk. (In practice, the most challenging part of risk prediction is to account for potentially undetected sensor faults, such as excessive GNSS satellite clock drift.)

Unfortunately, the same methods do not directly apply to HAVs, because ground vehicles operate under sky-obstructed areas where GNSS signals can be altered or blocked by buildings and trees. In general, the HAV environment is much more unpredictable than the aircraft’s, for reasons that include:

• a changing environment: traffic lights, construction, impact of rain on road adherence, sensor masking and occlusions,
• environmental diversity: intersection topography, road conditions, markings on ground, various traffic signs 
• road users that may interfere with HAV motion: other cars, trucks, pedestrians, bicyclists, etc. 
• comparatively large number of car manufacturers, equipment suppliers, and vehicle models, as well as with shorter model cycles than aircraft, causing wide variations in vehicle age and maintenance levels 
• non-uniform vehicle and road regulations at both the state and federal levels in the U.S. coupled with different international standardization processes.

Thus, HAVs require sensors in addition to GNSS, including laser scanners, radars, cameras, and odometers.

The parallel between aircraft and car applications in Figure 4 illustrates the significant challenge that lies ahead when bringing aviation safety standards to HAVs. It took decades of research and considerable resources to bring the alert limit requirement box down to 10 meters above and below the aircraft using the FAA’s GPS augmentation systems (the Wide-Area Augmentation System and the Local Area Augmentation System). For a car to stay in its lane, the alert limit requirement box must be an order of magnitude smaller, and has to maintain this level of safety in a more dynamic and unpredictable environment.

HAV Taxonomy
Creating a path to successful automated navigation requires an overall methodology to prioritize on imminently achievable objectives, and then expand to more challenging missions. First in this HAV taxonomy, a classification using six SAE autonomy levels has been presented in Table 1. This classification is further refined by segmenting a car’s trip into basic driving competencies, and by specifying the conditions under which a given HAV shall achieve these competencies. A similar classification was made in the early days of GPS-based commercial aircraft navigation safety analysis, where distinctions were made between different phases of flight, weather conditions, vehicle equipment, and airport infrastructure capabilities.

For example, in the early 1990’s, 40% of aircraft accidents were occurring during final approach and landing, and 26% during take-off and initial climb, which only represented an average of 4% and 2% of flight time, respectively. The FAA therefore concentrated their efforts on improving safety during these phases of flight. GPS augmentation systems were designed, with varying capabilities depending on airborne equipment and airport infrastructure, to guide the aircraft under the cloud ceiling, or to bring it all the way to touch-down. Similarly, the “first and last mile” are identified as the most challenging parts of HAV operations, whereas highway auto-drive systems have already been developed and implemented. In its 2016 Federal Automated Vehicles Policy, NHTSA identifies 28 HAV behavioral competencies, which are particularly challenging to meet in the first and last miles of a typical trip. These competencies are basic abilities that an HAV must have to complete nominal driving tasks; they include, for example, lane keeping, obeying traffic laws, and responding to other road users.

To better describe an HAV’s ability, the Federal Automated Vehicles Policy further specifies that basic driving competencies should be available under an HAV’s predefined Operational Design Domain (ODD), described by its geographical location, road type and condition, weather and lighting condition, vehicle speed, etc. The ODD captures the circumstances under which an HAV is supposed to operate safely.

Such classification is key to safety analysis. It can allow HAVs at different stages of their development to be simultaneously fielded, and for them to evolve by expanding their ODDs. The classification can also help in identifying geographical areas where improved road infrastructure is needed for automated operation, similar to airports requiring equipment for instrument navigation to deal with higher traffic density.

Furthermore, standards for electronic equipment, measured by Automotive Safety Integrity Levels, have been issued and compared with the aviation’s Design Assurance Levels (DAL). And, overall system safety levels have been codified, which in aviation account for both the severity and probability of occurrence of an incident, and in automotive applications account, in addition, for “controllability”, which is a measure of how likely an average driver is to maneuver out of a given imminent danger.

All of the above elements: (a) HAV autonomy level, (b) basic driving competency, (c) operation design domain, (d) vehicle electronic equipment, and (e) overall safety risk requirement must be specified to carry out a formal HAV safety analysis. Still missing from the HAV documents are clear guidelines, or example methods, on how to implement these safety requirements.

A Path Towards HAV Navigation Safety
When quantifying the safety of HAV navigation systems, such as in the example displayed in Figure 5, every component of the system including raw sensors, estimator and integrity monitor, and safety predictor, can potentially introduce risk. Unlike aircraft, HAVs require multiple and varied sensors to compensate for GPS signal blockages caused by buildings and trees. These sensor types must be integrated, and new methods to evaluate the integrity of multi-sensor systems must be developed. Furthermore, HAVs must have the ability to continuously predict integrity in a dynamic HAV environment.

In general, research on analytical evaluation of HAV navigation safety is sparse. For example, J. Lee et alia, Additional Resources use the concept of a “safe driving envelope,” but the approach focuses mostly on collision avoidance. The paper by O. Le Marchand, et alia, evaluates ground vehicle navigation, but shows an “approximate radial-error” of tens of meters, far exceeding the necessary sub-meter alert limit. A multi-sensor augmented-GPS/IMU system is used in the paper by R. Toledo-Moreo, et alia with “horizontal trust levels” of 7 meters to 10 meters, still an order-of-magnitude higher than the required HAV alert limit.

Multi-sensor integrity is addressed by M. Brenner, Additional Resources, but for a sensor combination specific to aviation and insufficient for terrestrial mobile robots. Other approaches to multi-sensor integration show promise, but do not provide rigorous proof of integrity. In fact, most publications use pose estimation error covariance as a measure of performance, which is understood as not being sufficient, but is the only metric currently available. Most critically, the metric does not account for fault modes introduced by feature extraction and data association, two algorithms commonly used in mobile robot localization (and discussed again below).

Unlike GPS, which gives absolute position fixes, IMUs, LiDAR, radar, and cameras provide relative displacements with respect to a previous time-step, or with respect to a map. Thus, measurement time-filtering is required, which makes integrity risk evaluation more challenging since past-time sensor errors and undetected faults can now impact current-time safety.

Example LiDAR Navigation Safety Evaluation
While safety quantification for GNSS and GNSS/INS has been rigorously performed for aviation applications, and is being researched for HAVs, navigation safety for LiDAR, radar, camera, and multi-sensor navigation is a widely unexplored research area. To provide a specific example on the research work that lies ahead, we have started developing safety risk evaluation methods for LiDARs. We selected LiDARs because of their prevalence in HAVs, of their market availability, and because of our prior experience. However, the techniques we are developing are general enough that radar, cameras, or any future sensor that returns range data can be substituted.

Raw range data must be processed before it can be used for navigation. One technique, visual odometry, establishes correlations between successive scans to estimate sensor changes in pose (i.e., position and orientation). These processes are highly computationally intensive, and have the same problems as other dead-reckoning techniques, such as wheel odometry over time. Thus, they can become inaccurate or cumbersome for HAVs moving over multiple time epochs. Although proprietary information regarding the use of visual odometry by HAV manufacturers is unavailable, the research literature suggests that it is only used for short time scale operations. A second class of algorithms provides sensor localization by extracting static features from the raw sensor data and associating those features to a map. This is typically done in two steps, as illustrated in Figure 6: feature extraction (FE) and data association (DA). The resulting information can then be iteratively processed using sequential estimators (e.g., Extended Kalman filter or EKF), which has been readily used in many practical applications.

There are several problems that the FE and DA algorithms are addressing. First, landmarks in the environment are unidentified, and their observations are not tagged in a manner similar to a GNSS satellite signal’s Pseudo Random Noise (PRN) number. Thus, the feature extraction algorithm must isolate the few most consistently identifiable, viewpoint-invariant landmarks in the raw sensor data. These features must be identifiable over repeated observations and distinguishable from one landmark to another. Features that are difficult to distinguish from each other can be found easily, but the possibility that the association is incorrect will greatly negatively impact the integrity risk.

Second, range data based on extracted features must match those features with those from a feature database or map. Data association algorithms accomplish this; however, incorrect associations commonly occur. These can lead to large navigation errors, as illustrated in Figure 6, thereby representing a threat to navigation integrity.

FE and DA can be challenging in the presence of sensor uncertainty. This is why many sophisticated algorithms have been devised. But, how can we prove whether these FE and DA methods are safe for life-critical HAV navigation applications, and under what circumstances? These research questions are currently unanswered. The most relevant publications on DA risk are found in literature on multi-target tracking. For example, in the paper Y. Bar-Shalom and T. E. Fortmann, an innovation-based nearest-neighbor DA criterion is introduced, which serves as basis in many practical implementations. The article by Y. Bar-Shalom, et alia, “The Probabilistic Data Association Filter,” provides a detailed derivation of the probability of correct association given measurements. However, this Bayesian approach is not well suited for safety-critical applications due to the lack of risk prediction capability, and to the problem of bounding the a-posteriori probability of association (a similar issue is encountered in the paper by F.C. Chan, et alia. Another insightful approach is followed in the paper by J. Areta, et alia). However, it makes approximations that do not necessarily upper-bound risks, hence do not guarantee safe operation, and it presents exact solutions that can only be evaluated using computationally expensive numerical methods, not adequate for real-time navigation. Also, the risk of FE is not addressed.

In response, we have been developing a new, computationally-efficient integrity risk prediction method to ensure safety of localization using LiDAR-based FE and DA. We have derived a multiple-hypothesis innovation-based DA method that provides the means to predict the probability of incorrect associations considering all potential landmark permutations. (For more details on these methods, see the following four papers in Additional Resources, Nos. 31, 49, 50 and 51.) We also determined a probabilistic lower bound on the minimum feature separation, which is guaranteed at FE, with pre-defined integrity risk allocation. The separation bound can be incorporated in an overall integrity risk equation. This new method was analyzed and tested to quantify the impact of incorrect associations on integrity risk. It showed that the positioning error covariance can be a misleading safety performance metric since cases were found where the contributions of incorrect associations to integrity risk far surpassed that of nominal errors accounted for in the positioning error covariance. In addition, the following key safety-tradeoff was illustrated: the more measurements are extracted, the lower the integrity risk contribution is under the correct association hypothesis, but the higher the other integrity risk contributions become because the risk of incorrect associations increases in the presence of cluttered, poorly-distinguishable landmarks. Finally, being surrounded by many landmarks increases the probability of continuous, uninterrupted navigation. The next step of this research aims at dealing with unmapped and non-static obstacles, and at quantifying the continuity risk of FE and DA.

Conclusion
Looking at the emergence of future HAV technology with the prior experience of aircraft navigation safety provides the means to scale up the challenges that lie ahead in the development of fully autonomous (Level 4 and 5) driverless cars. Many parallels can already be drawn between aviation safety requirements and early HAV standards and regulations. Still, the methods to fulfill these standards and regulations have to be established. If analytical methods are pursued, the following tasks need to be accomplished: (1) establish high-integrity raw sensor measurement error and fault models for non-GPS sensors; (2) develop analytical methods to quantify the safety risk of feature extraction and data association algorithms required in LiDAR, radar, and other pre-processing steps in camera-based localization; (3) design multi-sensor pose estimators and integrity monitors to evaluate the impact of undetected sensor faults on safety risk; and (4) derive, analyze, and experimentally implement integrity risk prediction in dynamic environments.

If these challenges are overcome, one will be able to quantify and prove the performance of an HAV’s navigation system — an essential part of safety. Proving navigation system integrity will also help give humans more confidence to trust HAVs, thus further developing the symbiotic relationship between humans and co-robots. Finally, as HAV technology progresses from driver’s aids such as active brake assist to full autonomous driving, this research is relevant now and will remain essential throughout the evolution of HAV technology.

Additional Resources
[1] Abuhashim, T.S., M.F. AbdelHafez, and M.-A. AlJarrah. Building a robust integrity monitoring algorithm for a low cost gps-aided-ins system. International Journal of Control, Automation, and Systems, 8(5):11081122, 2010.
[2] Ackerman , E., “Self-Driving Cars Were Just Around the Corner—in 1960”, IEEE Spectrum, September 2016
[3] Ackerman, E., “After Mastering Singapore’s Streets, NuTonomy’s Robo-taxis Are Poised to Take on New Cities,” IEEE Spectrum, 2016.
[4] Areta, J., Y. Bar-Shalom, and R. Rothrock, “Misassociation Probability in M2TA and T2TA,” J. of Advances in Information Fusion, Vol. 2, No. 2, 2007, pp. 113-127.
[5] Bailey, T., Mobile Robot Localization and Mapping in Extensive Outdoor Environments. PhD thesis, The University of Sydney, 2002.
[6] Bailey, T., and J. Nieto. Scan-slam: Recursive mapping and localization with arbitrary-shaped landmarks. In Workshop at the Institute of Electrical and Electronics Engineers Robotics Science and Systems (IEEE RSS), 2008.
[7] Bakhache, B., A Sequential RAIM Based on the Civil Aviation Requirements. In Proceedings of the 12th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GPS 1999), pages 1201–1210, 1999.
[8] Basnayake, C., M. Joerger, and J. Aulde, “Safety-Critical Positioning for Automotive Applications”, Inside GNSS Webinar, 2016.
[9] Bar-Shalom, Y., F. Daum, and J. Huang, “The Probabilistic Data Association Filter,” IEEE Control Systems Magazine, 2009, pp. 82-100.
[10] Bar-Shalom, Y., and T. E. Fortmann. Mathematics in Science and Engineering, chapter Tracking and Data Association. Academic Press, 1988.
[11] Bengtsson, O., and A.J. Baerveldt, “Robot localization based on scan-matching-estimating the covariance matrix for the IDC algorithm,” Robotics and Autonomous Systems, Vol. 44, 2003, pp. 29–40.
[12] Bonanni, R., “WAAS – LPV Airport and Aeronautical Surveys”, ANM Airports Conference, 2006.
[13] Bhuiyan, J., “Uber’s autonomous cars drove 20,354 miles and had to be taken over at every mile, according to documents,” available online here, 2016
[14] Blom, H.A.P., and Y. Bar-Shalom. The interacting multiple model algorithm for systems with markovian switching coefficients. IEEE Transactions on Automatic Control, 33(8):780783, 1988.
[15] Brenner, M., Integrated GPS/Inertial Fault Detection Availability. NAVIGATION, Journal of The Institute of Navigation, 43(2):111–130, 1996.
[16] Chan, F.C., M. Joerger, S. Khanafseh, and B. Pervan, “Bayesian Fault-Tolerant Position Estimator and Integrity Risk Bound for GNSS Navigation,” Journal of Navigation of the RIN, available on CJO2014, doi:10.1017/S0373463314000241, 2014.
[17] Chow, E., and A. Willsky. Analytical redundancy and the design of robust failure detection systems. IEEE Transactions on Automatic Control, 29(7):603614, 1984.
[18] Choukroun, D., and J. Speyer. Mode estimation via conditionally linear filtering: Application to gyro failure monitoring. AIAA Journal of Guidance, Control, and Dynamics, 65(2):632644, 2012.
[19] Clot, A., C. Macabiau, I. Nikiforov, and B. Roturier. Sequential RAIM Designed to Detect Combined Step Ramp Pseudo-Range Error. In Proceedings of the 19th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS 2006), page 26212633, 2006.
[20] ooper, A.J., A Comparison of Data Association Techniques for Simultaneous Localization and Mapping. PhD thesis, Massachusetts Institute of Technology, 2005.
[21] DARPA, “The Six Finishers of the DARPA Urban Challenge,” available online here, 2007.
[22] Defense Advanced Research Projects Agency (DARPA), “Robots conquer DARPA Grand Challenge,” Press Release, U.S. Department of Defense (DoD), 2005.
[23] Department of Transportation (DOT) National Highway Traffic Safety Administration (NHTSA) “Federal Automated Vehicles Policy: Accelerating the Next Revolution In Roadway Safety,” 2016
[24] Diesel, J., and S. Luu. GPS/IRS AIME: Calculation of Thresholds and Protection Radius Using Chi-Square Methods. In Proceedings of the 8th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1995), page 19591964, 1995.
[25] Dionne, D., Y. Oshman, and D. Shinar. Novel adaptive generalized likelihood ratio detector with application to maneuvering target tracking. AIAA Journal of Guidance, Control, and Dynamics, 29(2):465474, 2006.
[26] Diosi, A., and L. Kleeman, “Laser scan matching in polar coordinates with application to SLAM,” Proc. IEEE/RSJ IROS, 2005.
[27] Dissanayake, G., P. Newman, S. Clark, H. Durrant-Whyte, and M. Csorba. A Solution to the Simultaneous Localization and Map Building (SLAM) Problem. IEEE Transactions on Robotics Automation, 17(3):229–241, 2001.
[28] Dougherty, M., “Caltrans Leadership in Automated Vehicle Research,” Automated Vehicles Symposium 2017 (AVS2017), San Francisco, CA, 2017.
[29] Dragalin, V.P., A.G. Tartakovsky, and V.V. Veeravalli. The interacting multiple model algorithm for systems with markovian switching coefficients. IEEE Transactions on Information Theory, 45(7):24482461, 1999.
[30] Dragalin, V.P., A.G. Tartakovsky, and V.V. Veeravalli. Multihypothesis sequential probability ratio tests. ii. accurate asymptotic expansions for the expected sample size. IEEE Transactions on Information Theory, 46(4):13661383, July 2000.
[31] Duenas-Arana, G., M. Joerger, and M. Spenko, “Minimizing Integrity Risk via Landmark Selection in Mobile Robot Localization,” submitted to IEEE TRA, 2017.
[32] FAA, “System Design and Analysis,” Advisory Circular AC 25.1309-1A, 1988.
[33] FAA, “System Safety Design and Analysis for Part 23 Airplanes”, Advisory Circular AC 23.1309-1E, 2011.
[34] fars.NHTSA.dot.gov, “Fatality analysis reporting system,” Technical report, NHTSA, 2014.
[35] Federal Aviation Administration (FAA), “Automatic Dependent Surveillance-Broadcast Operations,” Advisory Circular AC No: 90-114A, DoT FAA, 2016.
[36] Federal Highway Administration (FHWA), “Vehicle Positioning Trade Study for ITS Applications”, FHWAJPO-12-064, 2012.
[37] Fenton, R. E., and K. W. Olson “The electronic highway” IEEE Spectrum, 1969.
[38] Forsberg, “NovAtel Establishes Advanced Research Partnership with Illinois Institute of Technology and the University of Arizona,” press release, 2016, available here.
[39] Gartner’s “2016 Hype Cycle for Emerging Technologies” available online here.
[40] Gertler, J., A survey of model based failure detection and isolation in complex plants. IEEE Control Systems Magazine, 8(6):3–11, 1988.
[41] Gitlin, J., “Prepare for the part-time self-driving car,” online at arstechnica.com, 2014.
[42] Google, “Google self-driving car testing report on disengagements of autonomous mode”, available online here, December 2015.
[43] Greenblatt, J. B., and S. Saxena, “Autonomous taxis could greatly reduce greenhouse-gas emissions of us light-duty vehicles,” Nature Climate Change, 5:860–863, 2015.
[44] Greiling Keane, A., “U.S. highway deaths decline for a fifth year, longest streak since 1899,” Bloomberg, Published December 08, 2011.
[45] Halsey III, A., and M. Laris, “Blind man sets out alone in Google’s driverless car,” The Washington Post, 2016.
[46] Hewitson, S., and J. Wang. Extended Receiver Autonomous Integrity Monitoring (eRAIM) for GNSS/INS Integration. Journal of Surveying Engineering, 136(1):13–22, 2010.
[47] Hype cycle curves available online here.
[48] International Organization for Standardization (ISO), “Road vehicles – Functional safety”, ISO 26262, 2011.
[49] Joerger, M., “Carrier Phase GPS Augmentation Using Laser Scanners and Using Low Earth Orbiting Satellites,” Ph.D. Dissertation, Illinois Institute of Technology, 2009.
[50] Joerger, M., M. Jamoom, M. Spenko, and B. Pervan, “Integrity of Laser-Based Feature Extraction and Data Association,” Proc. IEEE/ION PLANS 2016, Savannah, GA, 2016, pp. 557-571.
[51] Joerger, M., B. Pervan, “Continuity Risk of Feature Extraction for Laser-Based Navigation,” Proceedings of the 2017 International Technical Meeting of The Institute of Navigation, Monterey, California, January 2017, pp. 839-855.
[52] Joerger, M., and B. Pervan, “Quantifying Safety for Laser-based Navigation,” submitted to IEEE TAES, 2017.
[53] Kalra, N., and S. Paddock, “Driving to safety: How many miles of driving would it take to demonstrate autonomous vehicle reliability?” Technical Report RR-1478-RC, Rand Corporation, 2016.
[54] Kavanaugh-Brown, J., “Where the Research Meets the Road: Automated Highway Passes the Test ”, Government Technology, 1997. available here.
[55] Kelly, R., and J. Davis, “Required Navigation Performance (RNP) for Precision Approach and Landing with GNSS Application,” NAVIGATION, 1994.
[56] Lee, Y.C., “Analysis of Range and Position Comparison Methods as a Means to Provide GPS Integrity in the User Receiver,” Proc. of the 42nd Annual Meeting of The Institute of Navigation, Seattle, WA, 1986.
[57] Lee, J., B. Kim, J. Seo, K. Yi, J. Yoon, and B. Ko. Automated driving control in safe driving envelope based on probabilistic prediction of surrounding vehicle behaviors. Society of Automotive Engineers International Journal of Passenger Cars – Electronic and Electrical Systems, 8(1):207–218, 2015.
[58] Le Marchand, O., Philippe Bonnifait, Javier Ibaez-Guzmn, and David Btaille. Vehicle Localization Integrity Based on Trajectory Monitoring. In IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 3453–3458, 2009.
[59] Leonard, J., and H. Durrant-Whyte. Directed Sonar Sensing for Mobile Robot Navigation. Kluwer Academic Publishers, 1992.
[60] Li, Y., and Olson E.B. A general purpose feature extractor for light detection and ranging data. Sensors, 10(11), 2010.
[61] Lorden, G., Procedures for reacting to a change in distribution. The Annals of Mathematical Statistics, 42(6):18971908, 1971.
[62] Lu, F., and E. Milios, “Globally Consistent Range Scan Alignment for Environment Mapping,” Autonomous Robots 4, 1997, pp. 333-349.
[63] Madhavan, R., H. Durrant-Whyte, and G. Dissanayake. Natural landmark-based autonomous navigation using curvature scale space. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2002.
[64] Malladi, D. P., and J. L. Speyer. A generalized shiryayev sequential probability ratio test for change detection and isolation. IEEE Transactions on Automatic Control, 44(8):1522–1534, 1999.
[65] Maksarov, D., and H. Durrant-Whyte. Mobile Vehicle Navigation in Unknown environments: a Multiple Hypothesis Approach. In IEEE Proceedings on Control Theory Applications, volume 142, pages 385–400, 1995.
[66] National Transport Safety Board (NSTB), “Preliminary Report, Highway HWY16FH018,” Accident Report ID: HWY16FH018, 2016. available online here.
[67] Neville, K., and K. Williams, “Integrating Remotely Piloted Aircraft Systems into the National Airspace System,” Remotely Piloted Aircraft Systems: A Human Systems Integration Perspective, Wiley, 2017.
[68] Nguyen, V., A. Martinelli, N. Tomatis, and R. Siegwart. A comparison of line extraction algorithms using 2d laser rangefinder for indoor mobile robotics. In Proceedings of the Institute of Electrical and Electronics Engineers/Robotics Society of Japan International Conference on Intelligent Robots and Systems (IEEE/RSJ IROS), 2005.
[69] NHTSA “National motor vehicle crash causation survey: Report to congress,” Technical Report DOT HS 811 059, U.S. Department of Transportation, 2008.
[70] NHTSA, “Federal Motor Vehicle Safety Standards; V2V Communications, Notice of Proposed Rulemaking (NPRM),” DoT NHTSA, 49 CFR Part 571, RIN 2127-AL55, 2016, available online here.
[71] NHTSA “Assessment of Safety Standards for Automotive Electronic Control Systems”, DOT HS 812 285, 2016.
[72] Nikiforov, I., New Optimal Approach to Global Positioning System/Differential Global Positioning System Integrity Monitoring. AIAA Journal of Guidance, Control, and Dynamics, page 10231033, 1996.
[73] Nunez, P., R. Vazquez-Martin, J.C. del Toro, and A. Bandera. Feature extraction from laser scan data based on curvature estimation for mobile robotics. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2006.
[74] Othman, N.A., and H. Ahmad. The analysis of covariance matrix for kalman filter based slam with intermittent measurement. In Proceedings of the 2013 International Conference on Systems, Control and Informatics, 2013.
[75] Page, E.S., Continuous inspection schemes. Biometrika, 41(1-2):100–115, 1954.
[76] Parkinson, B.W., and P. Axelrad, “Autonomous GPS Integrity Monitoring Using the Pseudorange Residual,” NAVIGATION, Vol. 35, No. 2, 1988.
[77] Pfister, S.T., K.L. Kriechbaum, S.I. Roumeliotis, and J.W. Burdick. Weighted range sensor matching algorithms for mobile robot displacement estimation. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2002.
[78] Pfister, S.T., S.I. Roumeliotis, and J.W. Burdick. Weighted line fitting algorithms for mobile robot map building and efficient data representation robotics and automation. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2003.
[79] Radio Technical Commission for Aeronautics (RTCA), “Minimum Operating Performance Standards (MOPS) for Universal Access Transceiver (UAT) Automatic Dependent Surveillance – Broadcast (ADS-B),” RTCA, Washington DC, 2009.
[80] RTCA Special Committee 159, “Minimum Aviation System Performance Standards for the Local Area Augmentation System (LAAS),” RTCA/DO-245, 2004.
[81] RTCA Special Committee 159, “Minimum Operational Performance Standards for Global Positioning System/Wide Area Augmentation System Airborne Equipment,” RTCA/DO-229C, 2001.
[82] Reimer, B., “Revisiting the Topic – The Future is Autonomous Driving – But Are “We” on a Near Term Collision Course?” Automated Vehicle Symposium 2017, (AVS2017), San Francisco, CA, 2017.
[83] Röfer, T., “Using Histogram Correlation to Create Consistent Laser Scan Maps,” Proc. IEEE IROS-2002, Lausanne, Switzerland, 2002, pp. 625-630.
[84] Rogowsky, M., “The Truth About Tesla’s Autopilot Is We Don’t Yet Know How Safe It Is”, Forbes, 2016.
[85] SAE International, “Surface Vehicle Recommended Practice: Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles,” SAE Standard J3016, 2016.
[86] Schoettle, B., and M. Sivak, “A Preliminary Analysis of Real-World Crashes Involving Self-Driving Vehicles,” Report No. UMTRI-2015-34, October 2015.
[87] Sobel, M., and A.Wald. A sequential decision procedure for choosing one of three hypotheses concerning the unknown mean of a normal distribution. The Annals of Mathematical Statistics, 20(4):502522, 1949.
[88] Soloviev, A., D. Bates, and F. van Graas. Tight Coupling of Laser Scanner and Inertial Measurements for a Fully Autonomous Relative Navigation Solution. NAVIGATION, Journal of The Institute of Navigation, 54(3):189 – 205, 2007.
[89] Soloviev, A., Multi-Sensor Fusion for Navigation of Autonomous Vehicles. In Proceedings of the 26th International Technical Meeting of The Satellite Division of the Institute of Navigation (ION GNSS+ 2013), pages 3615 – 3632, 2013.
[90] Soloviev, A., C. Yang, M. Veth, and C. Taylor. Assured Vision Aided Inertial Localization. In Proceedings of the 27th International Technical Meeting of The Satellite Division of the Institute of Navigation (ION GNSS+ 2014), pages 2160 – 2173, 2014.
[91] Sukkarieh, S., E.M. Nebot, and H.F. Durrant-Whyte. A high integrity imu/gps navigation loop for autonomous land vehicle applications. IEEE Transactions on Robotics and Automation, 51(3):572578, 1999.
[92] Toledo-Moreo, R., M. A. Zamora-Izquierdo, B. beda Miarro, and A. F. Gmez-Skarmeta. High-Integrity IMMEKF-Based Road Vehicle Navigation With Low-Cost GPS/SBAS/INS. IEEE Transactions on Aerospace and Electronic Systems, 8(3):491–511, 2007.
[93] Tena Ruiz, I., Y. Petillot, D.M. Lane, and C. Salson. Feature extraction and data association for AUV concurrent mapping and localization. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2001.
[94] Thrun, S., W. Burgard, and D. Fox. A probabilistic approach to concurrent mapping and localization for mobile robots. Machine Learning and Autonomous Robots, 31(5):1–25, 1998.
[95] Thrun, S., W. Burgard, and D. Fox. A real-time algorithm for mobile robot mapping with applications to multi-robot and 3d mapping. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2000.
[96] Thrun, S., “Robotic Mapping: A Survey,” Exploring Artificial Intelligence in the New Millenium. Morgan Kaufmann Publishers Inc., 2003.
[97] Thrun, S., “National Highway Traffic Safety Administration (NHTSA),” keynote presentation, ION GNSS 2007, Fort Worth, TX, 2007.
[98] Van Eikema Hommes, Q. D., “Assessment of safety standards for automotive electronic control systems,” NHTSA Report No. DOT HS 812 285, Washington, DC, 2016.
[99] Waymo, “We’ve reached 3 million miles of selfdriving on public roads! That’s 1 million miles in just 7 months,” available online here, 2017.
[100] White, N. A., P.S. Maybeck, and S.L. DeVilbiss. Detection of interference/jamming and spoofing in a dgps-aided inertial system. IEEE Transactions on Aerospace and Electronic Systems, 34(4):12081217, 1998.
[101] Wikipedia , “Automotive Safety Integrity Level,” 2017. available here.
[102] Williams, S.B., G. Dissanayake, and H. Durrant-Whyte. An efficient approach to the simultaneous localization and mapping problem. In Proceedings of the Institute of Electrical and Electronics Engineers International Conference on Robotics and Automation (IEEE ICRA), 2002.
[103] Willsky, A. S., A Survey of Design Methods for Failure Detection in Dynamic Systems. Automatica, 12:601–611, 1976.
[104] Working Group C ARAIM Technical Subgroup, “Milestone 3 Report,” Technical report, EU-US Cooperation on Satellite Navigation, 2015.
[105] Yoshida, J., “Another Tesla Crash, What It Teaches Us,” EE Times, 2016.

The post Towards Navigation Safety for Autonomous Cars appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

Last year a problem with the GPS satellite timing signal triggered alarms and caused an unknown number of outages, and in Europe earlier this year the fledgling Galileo signal crashed due to unspecified ground facility issues.

A government report commissioned by Innovate UK, along with the UK Space Agency and the Royal Institute of Navigation, entitled “Economic impact to the UK of a disruption to GNSS”, comes in the wake of troubling incidents for GNSS operators, both the United States and Europe.

Last year a problem with the GPS satellite timing signal triggered alarms and caused an unknown number of outages, and in Europe earlier this year the fledgling Galileo signal crashed due to unspecified ground facility issues.

“We wanted to know the economics behind a loss of GNSS, and if there are innovations in the GNSS market we should be investing in, perhaps addressing GNSS vulnerability or new technology integration,” said Andy Proctor. “Understanding the economics of a GNSS worst-case situation has not been done in the UK before.”

Proctor, who chairs the UK Government PNT Group, commissioned the UK GNSS vulnerability report for Innovate UK, an executive non-departmental public body sponsored by the Department for Business, Energy & Industrial Strategy.

“Innovate UK is a Government Agency,” Proctor said, “a non-departmental body, which means we work in a cross-government way, talking to all departments regularly. We invest, mostly via grants, into UK businesses to stimulate economic growth, unlock R&D and market barriers and address market failures.”

Of crucial concern to Proctor and Innovate UK is the fact that while GNSS is a widespread technology, the full extent and nature of its use, as well as the resilience of its users to disruption, has not yet been understood.

The lead researcher and writer of the report is Greg Sadlier, Divisional Director, Space, at London Economics.

“We do lots of GNSS work,” Sadlier said. “We actually lead a consortium that does research for the European GNSS Agency (GSA) market analysis report. So, we do a lot behind the scenes in Europe space and also in the UK space policy environment.

“Given the substantial use of GNSS in the UK, the question was do we need to worry about resilience and, if so, to what extent.”

What is Vulnerability?
Although the report is not strictly concerned with defining the possible causes of a major GNSS outage, who or what the possible culprits might be is pretty clear.

For example, there is space weather; when the solar winds are acting up, satellite signals propagating through the Earth’s atmosphere can be profoundly affected. Indeed, the solar weather scenario was the basis for the chosen duration of the envisaged GNSS crash described in the report.

“For the study,” Proctor explained, “I decided upon a five-day duration, as it links in with scenarios in the national risk register, the space weather impact reports such as that from the Royal Academy of Engineering, and also the UK Government Space Weather Preparedness Strategy.”

But, he said, the GNSS signal could also be deliberately attacked. Terrorists can buy or build a jammer that is powerful enough to affect large areas of a major city from a publicly accessible location. Indeed, with a simple multi-frequency jammer, now easily available, any person can knock out all L1 to L5 bands, meaning GPS, Wide Area Augmentation System (WAAS), Galileo, EGNOS, and the rest. Finally, satellite components and/or ground-based systems can fail.

“It should be noted,” Sadlier said, “that the overall impact of an outage of GNSS is not necessarily independent of the source of the disruption: e.g., a severe natural space weather event causing a loss of GNSS may also cause an outage of other (satellite) services, including communications, broadcasting, meteorological, earth observation, as well as power supply.”

Stark Terms
The report finds that the UK could lose £1 billion per day (about $1.263 billion) if GNSS were to go down. And such a crash would cause more than just financial losses, as everyday essential activities would also be affected, including emergency care and mass transportation. A lack of GNSS would hit navigation hard, but would also affect multiple industries that need it for mapping, tracking and timing.

The report explains how GNSS is used, what part it plays in a variety of systems, as well as how resilient those systems are in the case of GNSS disruption across 10 application domains: road, rail, aviation, maritime, food, emergency and justice services, surveying, locationbased services (LBS), other infrastructure, and other applications.

What was probably already clear to some and what will be alarming to many others is the finding that all critical national infrastructures in the UK rely on GNSS to some extent, with communications, emergency services, finance, and transport identified as particularly intensive users.

This vast reliance on GNSS has developed over decades, based on widespread assumptions about availability and continuity. GNSS is also a primary input for transport, including road, air, maritime, and rail transport, as it is in agriculture, surveying, and for the legal professions.

The UK space industry derived an estimated turnover of £1.7 billion (about $2.15 billion) from PNT services in 2014- 15, supporting 4,000 jobs, while sectors generating a total of £206 billion (about $260 billion) in gross value added (11.3% of UK GDP) are supported directly by GNSS. But the crucial role played by GNSS in national infrastructures means that an even wider range of economic activities is underpinned by GNSS indirectly. Proctor also noted that there were many areas where the impact of a GNSS disruption was difficult to monetize, so that the final estimates arrived at in the report are likely to be low.

“We always expected the transport sector to be heavily impacted,” Proctor said, “but most surprising to me was the level of reliance of the maritime sector. I just didn’t think it would be so great, and this is something that sector in particular should take a look at and consider.”

Indeed, in the maritime navigation sector, GNSS is now widely treated as the sole necessary navigation solution. Virtually all traditional and even more recent back-up systems have simply disappeared, such that all other means of navigation have been replaced by GNSS.

The “Vulnerability Community”
The report will certainly be welcomed by the so-called “vulnerability community”, a loosely connected band of determined individuals that includes the likes of Dana Goward of the Resilient Navigation and Timing Foundation, who has been trying to ram home the GNSS vulnerability message for years.

The “Vulnies” also include eminent personalities such as Professor David Last, Past-President of the Royal Institute of Navigation, and even the venerable Brad Parkinson, “Father of GPS”. All of these and other figures have appeared at industry and policy events with messages not so much of doom and gloom, but of beware and prepare.

They believe it is perfectly right to point out the potential vulnerabilities of satellite-based navigation, so that the widening array of critical GNSS-supported operations can be appropriately safeguarded.

EU Heeds the Call
So far, no government seems to have moved very far towards answering the call for a GNSS back-up system. The European Union (EU) is no exception. Indeed, until very recently, anyone trying to get a serious answer from the European Commission (EC) on the question of GNSS vulnerability might have assumed the Commission hadn’t given the issue much thought at all.

The EU’s avoidance of questions about GNSS vulnerability is probably understandable, if nothing else on a human level. After more than 20 years of bleeding, sweating and crying tears on the road to an operational Galileo system, the last thing those folks will want to hear is “‘Oh, by the way, Galileo is not resilient enough so we need to look for something else”.

Attention is sometimes diverted by talk of the Galileo Public Regulated Service, the vaunted, military-like PRS. But it is generally expected that only a small proportion of Galileo users will have access to the PRS, and while it may be more robust, it will certainly not be immune to a wide-scale GNSS outage, either natural or man-made.

But the EU’s reluctance to look GNSS vulnerability square in the face may be changing. Inside GNSS recently reported, as per unnamed sources, that the Commission is funding a study in support of a European radionavigation plan, and that the study discusses the need for resilient PNT and looks at using terrestrial systems as well as space-based signals. This again will be music to the ears of the vulnerability squad.

Meanwhile, the European Space Agency (ESA), it appears, is also taking a broader perspective in its new navigation endeavors, describing a PNT effort, not a GNSS one, with a strong emphasis on hybrid systems.

In a recent conversation, one highly placed source within ESA said the Agency is “very conscious” of the vulnerabilities of GNSS, including Galileo. “The more these systems are used the more vulnerable they are,” our source told us. “I think we are still at an early stage in terms of market penetration. The number of users is still very low compared to what it will be in the future. There is a growing awareness, and we are at the correct stage to start implementing solutions to address vulnerability.”

Proctor said he too senses an increasing understanding of the vulnerabilities of GNSS across the EU. “There is still a lot of awareness to raise I believe, as GNSS has become proliferate and often embedded in systems sometimes without risk managers being aware.”

Back-up Options
Sadlier referred us to a full range of possible back-up systems considered in the report, from clocks and sextants to determine position at sea, or the use of old paper maps on the road, to more modern technologies such as radar systems.

“The aviation sector can make use of a number of existing back-up systems,” he said, “but there is currently no ‘universally applicable’ alternative to GNSS for the case of positioning and navigation, and many of the traditional means of navigation might not be readily available or useable, depending on the individual application.” True enough, my sextant went missing years ago.

For timing applications, Sadlier said, loss of GNSS can be mitigated by using adequate oscillators in the GNSS timing receiver that can hold time for a certain holdover period, ranging from a few minutes to many months. However, higher quality equipment with longer holdover periods is more expensive. Hence, loss of the GNSS signal will still affect sectors relying on its timing capabilities.

Of Course, eLoran
It seems any talk of GNSS vulnerability inevitably leads to the topic of eLoran. The two seem permanently linked. So much so that some have wondered whether the vulnerability “scare” isn’t just a pretense for the “eLoran folks” to get their pet technology funded.

On the other hand, it is just as possible that eLoran is constantly being put forward because it is in fact the best single back-up option. As reports of movement towards establishing eLoran as a potential back-up system continue to circulate in the United States, a glance in Europe’s direction also reveals a number of old Loran C navigation sites that could support an eLoran service as a back-up for GPS and Galileo.

Proctor said he was careful not to try to tip the scales when commissioning the report. “Yes, it’s true, a lot of people seem to have been talking about eLoran lately. But when we commissioned the report we didn’t lead Greg [Sadlier] in any way in terms of which potential back-up systems should be favored. As I have consistently said, I do not see eLoran as a cure-all for every case.”

And STL?
Another technology currently making waves is called Satellite Time and Location (STL). The Satelles company is using existing low-earth-orbit Iridium satellites, normally used for communications, to deliver a powerful signal for accurate and resilient positioning, navigation and timing that works anywhere, including indoors.

The STL signal is about 1000 times more powerful than GNSS signals, and it has some built-in cryptography elements, making STL easier to “hear” in difficult locations and harder to jam or spoof, compared to GNSS.

However, as with all other options, STL has its limitations. As Satelles’ Senior Radio Frequency Hardware Systems Engineer Stewart Cobb explained to us last December in Noordwijk, “STL works a lot like the old transit system where you watch a satellite go overhead and you take a series of fixes and between them you figure out your position. With GPS you need four satellites to get a fix, but generally you can see 10 or 12 so you can get a fix almost instantaneously. Basically, with STL it’s going to take longer to get a precise fix.”

Looking at the array of solutions examined in the report, Sadlier said, “The most applicable mitigation strategies for the largest number of applications are eLoran and STL. These high-availability services could mitigate many of the detriments in the maritime sector, and while the accuracy is insufficient for container stacking and autonomous cranes, the ability to schedule port operations and reduce downtime would help keep ports open.”

The cost of resurrecting eLoran to a usable level, he said, would be on the order of £50m over 15 years (or about $65.1 million). The cost of STL is still unclear at this early stage in its development.

Proctor also suggested the best solution is likely to involve a combination of technologies. “The combination of eLoran and STL likely would give the broadest coverage in the event of an extended GNSS outage,” he said.

The report identified Omnisense SP500 and Locata as possible preferred solutions for localized applications that require high levels of accuracy.

“Timing applications have been found to be resilient to a five-day outage of GNSS,” Sadlier said, “but one could implement eLoran, STL, Locata or freely-available Network Time Protocol (NTP) servers as a source of timing for low accuracy applications. If higher accuracy is required, Precision Time Protocols (PTP) or time-over fiber networks, like NPL Time, are two alternatives.”

Obstacles of the Political Kind
Proctor said there are three key target audiences for the report. First is the GNSS community itself. “This is a real evidence-based report,” he said. “As such it is a resource for the industry in question. There is new factual information here, real figures about the real world.”

Second, he said, is the infrastructure operators, the users. “It’s for all of those people and organizations whose equipment needs GNSS to function. This applies all the way down in the supply chain to the general public, people who are already dependent on GNSS and may not even know it.”

Finally, there are the policy makers, who need to understand and who may be in a position to say yes or no to funding initiatives.

“Personally, I believe there can always be more awareness of the benefits of GNSS and also the vulnerabilities associated with using it,” Proctor said. “Perhaps the blockages are where GNSS has become cheap to procure and implement, so assessing the costs of using an additional technology to back it up, when it rarely fails, is a difficult sell.”

The UK GNSS Vulnerability report is one of two PNT-focused studies recently commissioned in the UK, the other being a high-level Blackett Review, both of which will provide a well-rounded picture of the PNT-related economic and technical challenges faced by the UK, including critical infrastructure dependencies.

The post Answering the Call for a GNSS Back-up appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

The GPS community and Virginia-based Ligado are weighing new and upcoming test results as the standoff over interference with satellite navigation services enters its seventh year.

The dispute centers on the company’s now modified proposal to build a terrestrial wireless network supported by frequencies originally allocated for satellites. Though there had been a move some years earlier to augment the satellite services with ground stations the company’s first plan envisioned some 30,000 high-powered ground terminals.

The GPS community and Virginia-based Ligado are weighing new and upcoming test results as the standoff over interference with satellite navigation services enters its seventh year.

The dispute centers on the company’s now modified proposal to build a terrestrial wireless network supported by frequencies originally allocated for satellites. Though there had been a move some years earlier to augment the satellite services with ground stations the company’s first plan envisioned some 30,000 high-powered ground terminals.

Ligado’s frequencies neighbor the band relied upon by GPS and tests in 2011 showed the proposed network would overload the vast majority of GPS receivers. The Federal Communications Commission (FCC) put the project on hold early in 2012 and the company filed bankruptcy not long thereafter. It reorganized, emerging from Chapter 11 late in 2015 and soon changed its name from LightSquared to Ligado. It modified its network plan, but deep concerns over GPS interference continued.

Since exiting bankruptcy Ligado has been pressing the FCC for approval to build its network. Last year it poured more than $1.5 million into lobbying, according to filings with the Lobbying Disclosure Act database, and it funded studies of its own aimed at convincing the FCC to pull the broadband plan off the back burner.

The Results Are In
Ligado hired two technical organizations — Roberson and Associates, a telecommunications engineering firm, and the federal government’s National Advanced Spectrum and Communications Test Network (NASCTN).

Roberson released a study in May 2016 that, despite efforts to put the outcome in the best light, showed that Ligado’s broadband signals would impact several different GPS receivers. In December 2016, a summary of another Roberson test on signal reacquisition times, posted on the FCC docket by Ligado without comment, also showed impacts to some receivers.

Ligado used Roberson’s research to press for a change in the way interference is assessed. They had the firm measure position error as well as the carrier-to-noise-density ratio (C/N₀). Changes in the signal-to-noise ratio are the internationally accepted metric for determining levels of interference. Ligado asserted that position error was a better choice — a proposal that has, so far, failed to gain traction.

“Roberson’s use of measures of harmful interference other than the established standard of a 1 dB increase in the noise floor is not appropriate, as the record before the FCC amply demonstrates,” said Jim Kirkland, general counsel and vice-president of Trimble in a statement issued in response to the May 2016 Roberson test results. “Precision GNSS products like Trimble’s control important processes such as construction machine guidance at a level of precision that cannot be reliably observed by Roberson’s tests.” See here.

Using position error would favor Ligado’s signals and allow damaging levels of interference to GPS receivers, said Logan Scott, a GPS signal expert and a consultant specializing in radio frequency signal processing and waveform design for communications, navigation, and other systems. GPS receivers can give accurate position information even when experiencing serious interference, he explained in an interview last year. As long as they are tracking satellites they can provide fairly accurate positioning information right up to the point where the interference causes them to lose their lock on the satellite.

Though Ligado continues to press for a position error standard at least one expert thinks the matter is settled.

Tim Farrar, a technology consultant specializing in the satellite industry who has followed the Ligado saga closely, believes the firm has failed to persuade decision makers that position error is the better metric.

“They’ve lost that argument,” Farrar told Inside GNSS.

NASCTN Reports
Ligado’s project with the NASCTN also incorporated position error. NASCTN is a national network of federal, academic, and commercial test facilities formed by the National Institute of Standards and Technology (NIST), the Department of Defense (DoD) and the National Telecommunications and Information Administration (NTIA).

NASCTN’s self-described mission to accelerate deployment of wireless technologies raised doubts about its impartiality when the tests were announced — doubts that sharpened when it became clear Ligado, and not NASCTN, was crafting the research questions. Among the issues the project was designed to examine was the “preference” for using changes in C/N₀ over other measures, including position error, to evaluate interference.

The work got off to a controversial start with a test plan that was developed with limited input and then sprung on the GPS community — which then had just a few days over a weekend to provide feedback. NASCTN said at the time, however, that “to preserve its neutrality,” it would provide robust test methodologies and validated measurement data but “not make policy recommendations based on this data.” Moreover, NASCTN told Inside GNSS that it would “not make any policy-related interpretations associated with the test data,” although it “will likely make recommendations about testing methods.” See here.

The group released a 428-page report in February, describing its work and the tests it developed. As promised they are also making the data available. A request form for the data CD and a link to the report are here.

“They did a very competent job,” said Scott, who noted that their conclusions, as promised, were focused on the tests themselves. “Their conclusion was that you needed a high level of automation to conduct these tests.”

Damages
Scott reviewed the report and found the data in its charts was “pretty much consistent” with the Roberson tests, the tests done by the Technical Working Group (TWG) in 2011 that first clarified the levels of interference and the initial results of the Adjacent-Band Compatibility (ABC) Assessment now being conducted by the Department of Transportation (DoT).

“Basically what they’re showing is, you start transmitting high-power next to GPS, you’re going to damage national infrastructure,” Scott told Inside GNSS.

Scott’s perspective is in stark contrast to the interpretation put forth by Ligado, in a February 24 letter to the FCC. (Interestingly, the letter described the NASCTN project simply as a “Government study” and did not mention the firm’s sponsorship.)

“This comprehensive 428-page study that involved 1,476 hours of testing,” they wrote, validates the conclusion reached by the major GPS companies over the last 14 months: An LTE network operating within the specifications proposed in Ligado’s pending FCC applications will not harm the performance of GPS devices.”

Scott vehemently disagrees with the interpretation. “The report does not say that and the report actually provides quite a bit of evidence that (the network) will cause harm. How they reached this conclusion from this report is just beyond me because the data in the report very clearly shows damage.”

The difference in views appears to hinge on Ligado’s continued assertion that position error is the appropriate yardstick for interference.

“The government study found no impact on the position and timing accuracy of many GPS devices when exposed to mid-band LTE signals at significantly higher power than they would be under Ligado’s proposal,” the Feb. 24 letter said. The charts in the study, they continued, “illustrate that the study confirms Ligado’s prior submission in the record that a 1 dB-Hz decrease in the carrier-to-noise-density ratio (C/N₀) is not the appropriate standard for assessing harm to GPS receivers.”

Charts 6.24 and 6.28 from the study, however, show several of the receivers losing satellite lock as the signal power increases though it is not possible to determine which devices those are. The study lists the receivers tested, but it does so in such a way that specific results cannot be attributed to a specific piece of equipment.

Scott pointed out that the geographic impact of the network’s signals, based on the study’s data in Figure 6.24, could be substantial.

“So what kind of range is associated with that at the powers you’re talking about?” said Scott. “You’re talking about something of the order of 5 kilometers. So, you turn on a Ligado base station and 5 kilometers away there are high-precision receivers that are dying. And I don’t just mean degrading, I mean dying. Dead. Unable to operate.”

Scott explained that his conclusion is not based on NASCTN data, as NASCTN did not measure for such impacts, but on data from the TWG — the cadre of researchers who first looked at the LightSquared/GPS interference issue in 2011. Those researchers conducted tests at the same signal power that Ligado is now proposing, he said. To compare the two sets of data he annotated a chart based on TWG part 4 measured signal strength with boxes based on data from the NASCTN report. High-precision “receivers could experience total failure at ranges > 4 kilometers,” he said.

Scott noted also that there are reference receivers among the devices tested and that position error measurements would not reflect damaging interference to those receivers.

“The measure of a reference receiver is not its location performance; it knows exactly where it is,” Scott said. “The measure of its performance is how many satellites it’s tracking and how accurately it’s tracking the satellites. Looking through the test results I’m seeing all kinds of degradation. I’m seeing situations where they just flat out kill the receiver.”

The Adjacent-Band Assessment
The results from what, potentially, could be the most impactful study, however, are yet to come.

The ABC Assessment is a two-phase effort to determine the power limits by frequency, that is the interference tolerance masks (ITM) needed to protect both existing and future GPS receivers. The ABC Assessment is looking at a range of receivers and was developed with significant community input, including input from Ligado, which has been unhappy with the fact that it relies on C/N₀.

The first round of results, released in October, confirmed that cell phone receivers were the least sensitive and high-precision receivers the most sensitive, which tracks with earlier tests. Interestingly the masks developed for certified aviation receivers, which would seem likely to be restrictive, were not enough to protect other kinds of receivers. See here.

The ABC Assessment is “very conservative and therefore going to be pretty difficult [for Ligado] in terms of the power limits that are going to be imposed on the transmitters,” said Farrar. “…Obviously Ligado would, presumably, hope the FCC would just agree to the limits that Ligado proposed previously. But [Ligado’s proposed limits] are not likely to be viewed particularly positively [if the DoT is much more conservative].”

As of press time, the second round of data from the ABC Assessment was set to be released March 30. More details on the Global Positioning System Adjacent Band Compatibility Assessment can be found online here.

The post Ligado Test Results Roll In appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

The train control segment remains one of the most important components of the railway market today, accounting for €12.7 billion, or about $13.5 billion dollars, out of the world rail market, which amounts to about €150 billion (about $159.3 billion), according to the latest studies by UNIFE (The Association of the European Rail Industry). In the railway domain, this control segment also is one of the most important in terms of industrial and technological innovation, in particular in Europe with the development and the continuous improvement of the European Rail Traffic Management System/European Train Control System (ERTMS/ETCS) standards.

In the last two decades the ERTMS/ETCS has been developed by the European Union (EU) to not only enhance safety and increase efficiency, but more importantly to enhance cross-border interoperability and foster the development of a unique European rail market that is characterized by a high number of very different legacy systems in each country working with widely different regulations.

Early Objectives
The first aim of the development of the market pillar. It didn’t prohibit the possibility of developing a high-quality standard for the train control system that could be implemented all over the world. At the end of 2016, some 90,000 kilometers of ERTMS lines have been contracted in the world with about 12,000 vehicles representing a de-facto ERTMS/ETCS standard was, then, focused on the single standard and a laudable benchmark in terms of safety and interoperability. Figure 1 shows the architecture of the ERTMS Level 2, the dominant solution being promoted today. 

The European regulations called for an implementation and migration plan for the new standard requiring that all the core TEN-T network be equipped with the ERTMS/ETCS technology by 2030, and the entire comprehensive TEN-T network by 2050. It also provides that no significant improvement be implemented on the existing legacy systems, in order to further incentivize the transitions from these systems to the new European standard.

Some European countries, going beyond the European regulations, have already decided on a full and accelerated transition from their own legacy systems to the new standards on the whole national networks. This is the case in Belgium, Denmark, Luxembourg, Sweden, Switzerland and Norway (these two last countries decided for a network-wide implementation of ERTMS/ETCS even while not belonging to the EU).

Due to the high quality of the European system, ERTMS/ETCS specifications are becoming the standards for most of the world train control systems, even those outside of Europe. In some cases the system was chosen as developed by the EU, while in other cases the ERTMS/ETCS specifications have been adapted to the local framework. A network-wide development of the ERTMS/ETCS technology has been decided by the authorities of some countries outside Europe, including Algeria, Libya, Australia, China, India, South Korea, Saudi Arabia, and Taiwan.

A study performed by Bocconi University (Milan, Italy)estimates that the ERTMS/ETCS technology potential market could rise up to a quarter of the world market by 2030 and to half of the world market by 2050.

Despite the great success of the ERTMS/ETCS specifications, the system does have its limitations, in particular in terms of cost or economical sustainability especially for its deployment on local, regional and low traffic lines.

The ERTMS (European Rail Traffic Management System) consists of two main components: the ETCS (European Train Control System), that is a standard for in-cab train control, and the GSM-R, the GSM mobile communications standard for railway operations, that requires a dedicated telecom network whose costs range between €40,000-60,000 per kilometer (about $42,140-63,658) and whose technology is reaching the obsolescence.

An interesting alternative to GSM-R is represented by the so-called IP-based bearer independency accepting the use of public networks instead of dedicated networks. Another innovation proclaimed since the early stages of ETCS’s conception — but not given serious consideration until recent years — is the use of GNSS to localize the train. Both of these components are now being studied around the globe by the train control industry in order to decrease costs and foster the system deployment through innovations. For more details, see the Additional Resources section near the end of this article.

New technology “game-changers” such as IP-based telecom bearers and satellite localization were introduced at an ERTMS stakeholders meeting at the Innotrans fair in Berlin (September 20, 2016) where the new ERTMS Memorandum of Understanding (MoU) was signed (Figure 2).

Current ETCS specifications require that the train is localized by its odometer whose error is reset intermittently through balises, georeferenced points located along the line. The innovation consists of introducing the concept of a virtual balise that would be functionally equivalent to the physical balise, an electronic beacon deployed along a railway track at a georeferenced point (See ). This method prevents changing the current specification and the system architecture proposing a continuous localization system through satellite, and then assuring interoperability between current technologies. Figure 3

The detection of the virtual balise would be performed by an on-board virtual balise reader providing the absolute train location to the train control system, while the odometric subsystem measures train speed and distance travelled, then provides a relative positioning with respect to the last detected physical balise.

The main advantage of this technological innovation consists in the savings related to the not-installed physical balises and to the related maintenance that could become increasingly costly for the infrastructure manager, particularly on those lines that are not easily accessible. The need for providing a specific on-board virtual balise reader does represent an increased cost that has to be valued against the potential savings of the ground side.

Reference Platform
As a reference, the ERTMS platform is designed to operate with virtual balises — tasked with the same function of physical balises for the train localization — and uses public GSM/LTE/Satellite networks instead of the GSM-R dedicated network. Studies and research activities are being developed to assess the autonomy of the virtual balise concept and to identify operational conditions under which the virtual balise concept can replace the deployment of physical balises along the lines.

This is the case of railway lines with long tunnels or with a high complexity in terms of forks and junctions, that may require the presence of some physical balise in complement to the use of the GNSS localization.

Furthermore, the generation of virtual balises implies a proper interface with a GNSS Augmentation network and the ERTMS system to over bound the positioning error according to the safety requirements of ERTMS.

Figure 4 shows the architecture studied to implement the virtual balise functionality in Figure 3. The interface with the ERTMS Radio Block Centre (RBC) is the Track Area Local Server (TALS) introduced for processing the augmentation messages supplied by either dedicated and/or public augmentation networks including European Geostationary Navigation Overlay Service (EGNOS) and Wide Area Augmentation Sysztem (WAAS). The augmentation messages, after being processed by the TALS, are transmitted to the GNSS receivers of the trains through the secure telecommunication link of the ERTMS system. Multiple constellations (i.e., GPS + Galileo) and other localization means based on telecom signal of opportunities can enhance the availability and resiliency of train localization.

The train localization is provided by the virtual balise reader by processing the GNSS signals of two independent receiver chains with a track data base describing the geometry of the rail segments. Furthermore, coherence checks (independently from GNSS) are provided by the ERTMS diagnostic means to achieve the high Tolerable Hazards Rate (THR = 0.5 * 10-9 / hour) required by the ERTMS, which otherwise could have not been provided by the receiver itself. When the train passes on a point where a virtual balise is located, it identifies the location and transmits it to the RBC. The RBC then compiles information from the interlocking and trains in its control area and sends movement authorities and other information to individual trains, taking into account a safe distance to the train ahead.

On this basis, the train on-board unit (OBU) is able to calculate its braking distance and optimal speed. The train’s OBU is composed by the ETCS standard modules, a new TLC module able to transmit and receive GSM-R, GSM, LTE and satcom signals, and a physical balise reader in case some physical balise is needed on the ground side to ensure interoperability on other lines equipped with the standard ERTMS solution.

Economical Analyses
An impact analysis aims to evaluate the conditions of economic conveniences from the introduction of a new satellite technologies solution for the evolution of the ERTMS system. A quantitative analysis has been carried out within the ERSAT EAV project based on the Cost-Benefit Analysis (CBA) methodology and focused on regional and local lines in Europe. The total market in Europe has been conservatively estimated to be 28,000 line kilometers where different legacy systems are installed, many outdated and in several cases still manually operated. The goal is to replace these older systems with the ERTMS based on GNSS and multi-bearer telecom providing the new investments are economically sustainable.

The CBA methodological framework took into consideration the following cost elements provided by RFI and DB Netz, the railways Infrastructure Managers (IM) in Italy and Germany, and by Ansaldo STS, which is delivering the first GNSS-based train control system in Australia:

• the economic effects are assessed at the system level, from the society’s point of view;
• the economic effects of the project scenario are evaluated in differential terms with respect to an alternative scenario, which is the upgrade to the traditional ERTMS;
• the analysis is applied considering a time horizon defined taking into account the life cycle of the main investment and it uses a discounted flows technique.

Under these circumstances the CBA shows that the potential benefits far outweigh the costs of introducing a GNSS-based ERTMS with advantages depending on the status and complexity of the single line. Since cost savings are due to the reduction of physical balises and telecom infrastructures, a proper sensitivity analysis has been made to evaluate the impact of maintenance costs of balises and the availability of GSM-R infrastructures (Figures 5-6).

Ground equipment costs slightly more for IMs (the owner and operators of the railways network) because of additional components in the RBC (namely the TALS and the track database). For Railways Undertaking (RUs) — the train owners that provide the service to passengers, the required investment in the OBU is also higher in the project scenario — the layout of OBU in the project scenario includes: standard modules, (including ETCS), a VBR (Virtual Balise Reader), a BTM (standard Balise Transmission Module since physical balises are still needed in some critical points of the network – e.g., level crossings), and a MAR (Multipath Access Router) which selects the public telecom bearer including the GSM-R.

The main benefits for IMs derive from the reduced need for physical balises, which translates into savings during both the investment and the operating phase.

Collateral benefits also can be assessed for RUs due to the possibility of optimizing the train driving, reducing both the time spent at the start of a mission (when the train needs to meet a physical balise for its position to be safely assessed) and the events of unnecessary brakings and slowdowns deriving from failures in the balise-BTM communication.

However, in the project scenario, the accuracy and availability requirements are assumed to generate (at least in the first part of the time horizon) the need for a further level of augmentation of the satellite signal being provided by a satellite-based augmentation system (SBAS), which represents an additional cost for IMs.

In terms of the telecommunication part, the project scenario benefits from removing the need to invest in GSM-R while making use of already available public/commercial infrastructure and more affordable services.

All in all, the GNSS-based solution proves to be especially convenient because of relevant savings in both capital and operating expenses, with final Benefit/Cost Ratios well above 1 in all case studies examined.

However, most of these savings are captured by the IMs. The business case for RUs is a more complex issue since both VBR and BTM are necessary to ensure a train operates with traditional balises and virtual balises as well. Nevertheless, in new green field lines it is possible to eliminate the BTM, thus reducing the investment for RUs. The defined scenarios do not look ahead to impacts which directly benefit the balance sheets of RUs, but some indirect “collateral” benefits have been identified. Some of these benefits have been estimated within the CBA, while others cannot be estimated quantitatively because they depend on several factors linked to the policy and strategies of other players, but are worth mentioning.

Additionally, some of the gains that can be highlighted also affect the convenience of the project at the system level and are capable to lead it to higher Benefit/Cost Ratios.

• TLC costs and infrastructure charges. Currently the fee for TLC services used by RUs is generally included in the infrastructure charges paid by RUs to IMs, as IMs are the providers of GSM-R services. In the project scenario, the value chain of the telecommunication services would change, and the prudential assumption applied in the CBA was that it would be the RUs to pay direct fees for TLC service to the TLC and satcom provider. However, this means that the project scenario includes operating TLC costs for RUs that the baseline scenarios does not consider – since no assumption on a corresponding decrease of infrastructure charges was adopted. As a matter of fact, the decrease of infrastructure charges as a way to partially transfer the benefits of the GNSS-based scenario from IMs to RUs is a possibility that – while difficult to estimate – has to be taken into account as a favorable option for the railway system as a whole. Even if this was limited to the amount of TLC costs (so that the shift from the baseline scenario to the project scenario was cost-neutral for RUs in terms of telecommunications), the benefits for the business case of RUs would be relevant, putting their Benefit/Cost Ratio at 1.14.

• Additional costs of OBUs and Virtualization of all balises. The main disadvantage for RUs, and the main item which makes their business case weak is the need to upgrade the OBU at a higher cost than the ERTMS scenario would imply. In fact, there are three sensitive assumptions in the project scenario: (1) the OBU will still include a BTM. This derives from the fact that physical balises are still necessary today for allowing accurate positioning of the train in certain critical points of the network; (2) the unit cost of MAR; and (3) the unit cost of VBR. The combination of these assumptions implies an additional cost that can increase the cost of OBUs of up to 16%. By reducing the additional cost of OBUs in the project scenario (compared to the baseline scenario) it would be possible to further contribute to the business case of RUs, moving their BCR above 1.32 in the theoretical case where OBUs for the GNSS-based ERTMS would not cost more than in the traditional ERTMS scenario.

At the system level, this solution is beneficial even under prudential assumptions. However, a major issue is avoiding the need for an additional ground-based augmentation network and equipment by improving the performance of satellite-based augmentation combined with existing ground network.

The basic rationale of introducing satellite technologies in the rail sector is to generate economic benefits and spread them among the users by simplifying the infrastructure and equipment needed. As indicated by the economic analyses, minimizing the need and/or the impact of additional “hard” elements, as compared to alternative solutions, is of utmost importance to exploit the benefits (given by the virtual balise concept) and make this solution ground-breaking. To this aim, the RHINOS (Railway High Integrity Navigation Overlay System) project coordinated by Radiolabs and involving GPS experts at Stanford University, DLR, University of Nottingam and University of Pardubice is studying a way to define the protection level to comply with the challenging requirements of the ERTMS safety standards. For more details, see the Additional Resources section. The ultimate goal is to re-use existing and future GNSS augmentation systems world-wide in a seamless fashion as is the case in the aviation sector.

Operational Aspects
The ERTMS system guarantees a safe train front end by using its odometry to measure the distance from the previous valid balise (last relevant balise group). The error accumulated by the odometry (confidence interval) is reset once a new balise group is encountered. Looking to Figure 7, the confidence interval that must be guaranteed within a THR of 0.5.E-9/hour, is bounded by the virtual balise location accuracy plus the over/under reading amount of the odometer that is a function of the travelled distance.

The magnitude of the confidence error is taken into account to ensure a safe distance with the preceding train. However, the most stringent train positioning requirement is in the order of three meters to dis-criminate the track where the train is located. The positioning of virtual balises is a matter of trade-off between the surrounding environment to limit the multipath effects, and operational constraints. The multipath represents the most important threat for rail application and to this respect the characterization of possible operational scenarios is an important issue under investigation in the STARS project. With respect to physical balise whose error is one meter, the virtual balise error is probabilistic (Figure 8) and must be computed, assessed and validated in the rail environment.

From an operational point of view, the introduction of virtual balises and the use of public telecom networks represents a step change innovation for ERTMS. In fact, these technologies belong to other entities not strictly under the control of the signalling suppliers and infrastructure managers. Nevertheless, a safety case must be produced including these technologies, and specific interfaces are mandatory to define a Service Level Agreement (SLA) with the providers of augmentation services and telecom services. Figure 9 shows the relationship among different players.

These providers are not yet in the field as is the case in the aviation sector where the Air Traffic Control agencies have created European Satellite Service Provider (ESSP) as a service provider for EGNOS.

In the United States, where the Federal Railroad Administration continues to support railroads in implementing Positive Train Control (PTC) — a processor-based/communication-based train control system designed to prevent train accidents — telecom services are being provided by specific companies to federate the demand and reach economy of scale, while in Australia, the Australian Rail Track Corporation (ARTC) has involved major telecom operator Telstra, which uses both cellular and satellite technologies, to provide telecommunications to the national rail freight network. ERTMS has a fundamental requirement of interoperability, and until now the GSM-R standard has guaranteed this requirement, leaving competition at the supplier level. With the new technologies a shift from capex to service is inevitable to capture the economy of scale. Service providers will likely be part of the signalling integrator supply chain and must guarantee interoperability, cost effectiveness and technology innovation in order to overcome obsolescence of their products.

Important contributions to innovation on ERTMS are part of Shift2Rail — the first European rail joint technology initiative to seek focused research and innovation (R&I) and market-driven solutions by accelerating the integration of new and advanced technologies into innovative rail product solutions.

Conclusion
Exciting new capabilities and business models are making GNSS use a key contributor to deploying ERTMS on local and regional lines now and in the coming years. ERTMS has been successful in becoming the European train control system and in ensuring interoperability, competitivity, and export on international markets. GNSS use for train localization and public telecom networks are expected to produce significant benefits for the operational costs that are key to making ERTMS economically sustainable for both local and regional lines that represent the bulk of lines. Most of these lines are old and manually operated, and require huge investments in the next years.

A new service-based business model is emerging to fully exploit the benefits of these innovations and to allow the rail industry to provide turn-key systems. The results of the projects launched in 2012 are contributing to consolidate the roadmap and to create a consensus among the rail community. A remark-able synergy potential is expected by exploiting Galileo as a complement to GPS and other constellations, that together with ERTMS represent the two pilasters of the European industrial policy. As we look ahead, with new GNSS features coming on the field and increasing ERTMS needs, GNSS and ERTMS are becoming more tightly intertwined, amplifying their combined business models.

Additional Resources
1. 3InSat Project
2. ERSAT (ERtms on SATellite)
3. First RHINOS Workshop at University of Stanford marked bold commitment on high integrity GNSS for ERTMS
4. Marais, J., IFSTTAR “Etat des lieux des initiatives GNSS – Rail, JOURNÉE INNOVATION GNSS RAILTOULOUSE 24 MARS 2016
5. Next Generation Train Control (NGTC)
6. Railway Stakeholders Commit to Disciplined ERTMS Deployment
7. Shift2Rail Joint Undertaking
8. World’s Rail Market Study, Forecast 2016 – 2021, UNIFE
9. STARS – Satellite Technology For Advanced Railway Signalling

The post GNSS for ERTMS Train Localization appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

A new reports says that the GNSS market landscape is growing because of rapid growth of satellite technology-enabled wearables, unmanned aerial vehicles (UAVs), new innovation opportunities, and low-cost precision receivers.

ABI Research’s new GNSS integrated circuit (IC) vendor competitive analysis says that Broadcom and Qualcomm remain the two top companies for the fourth straight year.

A new reports says that the GNSS market landscape is growing because of rapid growth of satellite technology-enabled wearables, unmanned aerial vehicles (UAVs), new innovation opportunities, and low-cost precision receivers.

ABI Research’s new GNSS integrated circuit (IC) vendor competitive analysis says that Broadcom and Qualcomm remain the two top companies for the fourth straight year.

Within the past year, Broadcom grabbed more headlines with the company’s wearables success and initial work on L1/L5 dual-frequency receivers, the report said. However, the report says Qualcomm continues to lead in total GNSS shipments, as well as such innovative new technologies as LED/visible light communication and [long-term evolution] LTE Direct. Qualcomm’s partnership with Baidu on its IZat platform is also notable and represents the beginning of the era of "always on, ubiquitous location technologies," the report said.

The report lists MediaTek and u-blox in the third and fourth top market positions, respectively. "MediaTek and u-blox once again swapped places," says Patrick Connolly, ABI Research principal analyst. "U-blox had another stellar year financially and, along with Skytraq, led the way on low-cost precision GNSS with its NEO-M8P module. MediaTek, which showed significant success in wearables and smartphones, transitioned back to third place primarily due to growing market share."

The report says that CEC Huada and Samsung, both showing market innovation over the past year, are poised to instill great change in the marketplace. "CEC Huada developed single-frequency real-time kinematic (RTK) GPS, as well as [BeiDou Navigation Satellite System] receivers and INS/MEMS receivers, which the company released to select customers in 2016," Connolly said. "And it is now developing a dual-frequency BDS receiver and a receiver for [India Regional Navigation Satellite System]. Samsung, meanwhile, launched its first embedded GNSS solution, the Exynos CPU chipset. Given its presence across so many GPS-enabled consumer electronic devices, the company looks set to be a major disruptor in the coming years."

ABI said its report includes researched innovation and implementation parameters to determine the companies best positioned for success–and the companies that are in danger of losing out. In addition, the report features emerging competitive threats and technologies, ABI said.

This year, the GPS/GNSS IC market evolved into such new markets as automotive, wearables, and the internet of things (IoT), with huge potential to grow into sizeable GNSS and ubiquitous-location markets where a specific, optimized IC design will be a major plus, according to the report. "Precision GNSS techniques also came to the forefront as companies explore opportunities around vehicle-to-vehicle (V2V), advanced driver assistance (ADAS), and driverless cars. This is now the most significant design trend in the industry, with consumer GNSS IC vendors facing stiff competition for precision GNSS incumbents like Trimble and Novatel, as well as a number of new interesting start-ups," the report said. "Finally, new entrants like Intel and CEC Huada continue to expand their offering, while Samsung’s entry could significantly change the market share dynamic."

Companies featured in the study include Broadcom, CEC Huada, Galileo Satellite Navigation, Intel, MediaTek, Qualcomm, Samsung, SkyTraq Technology, STMicroelectronics, and u-blox.

The post New Report Says Broadcom and Qualcomm Top GNSS IC Vendors appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

Real-time position accuracy, achievable in dense urban areas using low-cost equipment, is currently limited to tens of meters. If this could be improved to five meters or better, a host of potential applications would benefit. These include situation awareness of emergency, security and military personnel and vehicles; emergency caller location, mobile mapping, tracking vulnerable people and valuable assets, intelligent mobility, location-based services and charging, augmented reality; and enforcement of curfews, restraining orders and other court orders.

Real-time position accuracy, achievable in dense urban areas using low-cost equipment, is currently limited to tens of meters. If this could be improved to five meters or better, a host of potential applications would benefit. These include situation awareness of emergency, security and military personnel and vehicles; emergency caller location, mobile mapping, tracking vulnerable people and valuable assets, intelligent mobility, location-based services and charging, augmented reality; and enforcement of curfews, restraining orders and other court orders.

A further accuracy improvement to around two meters would enable navigation for the visually impaired, lane-level road positioning for intelligent transportation systems, aerial surveillance for law enforcement, emergency management, building management, newsgathering, and advanced rail signaling.

This article explores how 3D mapping can be used to achieve a “step change” in real-time GNSS performance in dense population centers, including urban canyons.

The Problem
Buildings and other structures block many GNSS signals in urban environments. When obstructions block satellite signals that cross the street at right angles, allowing only along-street signals to reach a receiver, signal geometry and positioning accuracy decline. In extreme cases, where the ratio of the building heights to the street width is very high, the number of directly received signals may be insufficient to compute a position solution.

Another well-known signal propagation effect occurs when buildings, walls, vehicles, and the ground reflect GNSS signals. Metal, metallized glass, and wet surfaces are particularly strong reflectors. Where the direct line-of-sight (LOS) signal is blocked and only a reflected, or non-line-of-sight (NLOS), signal is received, a ranging error occurs that is equal to the additional path taken by the NLOS signal.

Strong NLOS signals can be difficult to distinguish from direct LOS signals, particularly using a smartphone, which has a linearly polarized antenna. Reflected signals can interfere with reception of the direct LOS signals, a phenomenon known as multipath interference or more simply, multipath, because the signal is received via multiple paths. Figure 1 (see inset photo, above right, for all figures) illustrates both phenomena.

GNSS user equipment can minimize ranging errors due to multipath interference by exploiting receiver signal processing technology at the expense of increased power consumption and hardware cost. However, this approach does not mitigate NLOS reception errors at all.

Multi-constellation receivers have significantly improved the availability of GNSS positioning in dense urban areas. However, accuracy remains a problem. In hostile signal propagation environments such as dense urban areas, the stand-alone, single-epoch positioning accuracy of GNSS is degraded to an average of about 25 meters. In the worst cases, position errors can exceed 100 meters.

Filtered or carrier-smoothed positioning algorithms can perform better, but these require an accurate position solution for initialization. They then need a sufficient number of good-quality GNSS signals to maintain solution accuracy and, hence, the ability to reject poor measurements using innovation filtering. In practice, this can be difficult to achieve in challenging urban environments.

Outdoor Wi-Fi positioning is accurate to around 25 meters, while other radio positioning techniques either offer poorer accuracy or require expensive dedicated infrastructure. Dead-reckoning techniques, including wheel-rotation sensing and pedestrian dead-reckoning using step detection, require accurate initialization, and their position accuracy then decreases as distance measurement errors accumulate.

If we are to achieve low-cost, real-time accuracy of five meters or better in challenging urban environments, a whole new approach is needed. Currently, 3D mapping–aided GNSS presents a great opportunity.

The Opportunity
To implement 3D mapping–aided GNSS, we need three things: measurements, mapping, and algorithms. In dense urban areas, the more satellites that are available for use, the better. Therefore, the return of GLONASS to full operational capability in 2011, followed by the wide availability of GLONASS-capable receivers presented a step-change in signal availability. In 2016, with the Galileo satellite deployment well under way and the extension of BeiDou from a regional to a global system, we are in the midst of a second step-change.

To implement advanced positioning algorithms, we need access to the “raw” measurement data, namely, the pseudorange and signal-to-noise ratio (SNR) or carrier-power-to-noise-density ratio (C/N₀). Pseudorange rate (Doppler) and carrier phase can also be useful. Survey receivers have always provided this information, but obtaining these data from consumer receivers has historically been problematic. Today, however, some low-cost GNSS receivers generate pseudorange and SNR measurements from all GNSS constellations, providing access to this data through the application programming interface (API) on smartphones and tablets running the Android Nougat operating system that have a compatible GNSS chipset.

The second ingredient, 3D mapping, can prove expensive when it comes to highly detailed 3D city models. However, simple block models, known as level-of-detail (LOD) 1, are sufficient for most 3D mapping–aided GNSS implementations. (Figure 2 shows an example.). The free, editable OpenStreetMap provides building maps for major cities worldwide, much of it three-dimensional (This can be viewed online using, for example, the F4 Map demo, noting that a default height is assumed for 2D buildings.)

Data is available from national mapping agencies with various terms and conditions, and private commercial companies such as Google and Apple also hold large amounts of 3D mapping data. Although such digital map coverage is far from universal, it tends to be available in dense urban areas where it is most needed. Conventional stand-alone GNSS positioning works well enough in low-density areas.

The final ingredient is advanced positioning algorithms that take advantage of 3D mapping. Many algorithms have been developed. Some use pseudoranges while others use SNR measurements with varying trade-offs between performance and processing load. Over the past five years, results have been published from more than 10 different research groups spread across Europe, North America, Japan, and the Middle East.

What Can Be Done?
3D mapping can be used to provide three types of information for aiding GNSS positioning. The first, and simplest, is the terrain height. Next are the predictions of which satellites will be directly visible at a given position and time, i.e., which lines of sight are blocked by buildings and which are not.

Finally, 3D mapping can be used to predict signal reflections, including the path delay of the reflected signal with respect to the direct signal and, potentially, its relative amplitude. However, amplitude is difficult to predict because 3D mapping data doesn’t include information on building reflectivity at GNSS frequencies, which can be quite different from optical reflectivity.

Terrain height-aiding is useful because, by constraining the position solution to a surface, it effectively removes a dimension from the positioning problem. In open areas, terrain height-aiding only improves the vertical position solution (as one might expect). However, in dense urban areas where the signal geometry is poor, it can improve the horizontal accuracy by almost a factor of two.

We can use satellite visibility predictions to aid both conventional ranging-based GNSS positioning and a new class of SNR-based positioning algorithms known as shadow matching. Each GNSS signal is predicted to be directly visible in some areas and blocked (shadowed) in other areas.

Shadow matching assumes that the user is in one of the directly visible areas if the received SNR is high and in one of the shadowed areas if the SNR is low or the signal is not received at all. Figure 3 illustrates the general principle. Repeating this calculation for each GNSS signal enables one to reduce the area within which the user may be found.

In practice, the SNR distributions of direct LOS and NLOS signals can overlap, particularly when using a smartphone antenna. Furthermore, real urban environments and signal propagation behavior are more complex than can be represented using 3D mapping. Therefore, a practical shadow-matching algorithm works by scoring a grid of candidate positions according to the degree of correspondence between the satellite visibility predictions and the SNR measurements. This enables inaccuracies in the process to be treated as noise so that a correct position is still obtained provided there is sufficient “signal.”

Due to the building geometry, shadow matching is normally more accurate in the cross-street direction than the along-street direction, which complements ranging-based GNSS positioning in dense urban areas. Shadow matching enables users to determine which side of the street they are on in environments where other positioning technologies do not.

Satellite visibility predictions can be used to aid ranging-based positioning in a number of different ways. Where the position is already known to within a few meters, we can predict which signals are NLOS with reasonable accuracy and simply exclude them from the position solution (assuming there are sufficient direct LOS signals). Otherwise, which signals are directly visible depends on the actual position of the user. A simple approach is to determine the proportion of candidate positions at which each signal is predicted to be directly visible. This information is used to weight each measurement within the position solution and to aid consistency checking. This approach typically improves the positioning accuracy by 20–25 percent and can handle initialization errors of about 100 meters.

To make the best use of satellite visibility prediction, a conventional least-squares (or extended Kalman filter) positioning algorithm should be replaced by an algorithm that scores candidate position hypotheses, according to the difference between the measured and predicted pseudoranges (assuming LOS propagation). Different assumptions about the error distribution can then be made at various candidate positions — according to which signals are predicted to be LOS or NLOS at each position. Thus a symmetric error distribution can be assumed for LOS signals and an asymmetric distribution for NLOS signals, with the scoring adjusted accordingly.

The candidate positions may be distributed in a regular grid or semi-randomly (as in a particle filter). Terrain height-aiding is used to associate a height with each horizontal coordinate, enabling the search space to be limited to two dimensions. University College London (UCL) has combined this hypothesis-scoring ranging algorithm with shadow matching, which has reduced the root mean square (RMS) horizontal positioning error in dense urban areas from 26 meters to 3.9 meters using data from a consumer-grade GNSS receiver — a sixfold improvement. Figure 4 shows the across-street and along-street positioning errors obtained at 18 sites in central London, depicted in Figure 5.

3D mapping can also be used to predict reflected signals. For shadow matching, this potentially provides more information from which to predict the SNR measurements. For ranging, this can be used to predict which direct LOS signals are subject to multipath interference, enabling us to adjust their weighting or assumed error distribution within the positioning algorithm.

Where the path delay of the reflected signals is predicted, NLOS reception errors may be corrected, enabling NLOS signals to contribute to an accurate position solution. However, accurate correction of NLOS errors requires an accurate position solution, a “chicken and egg” problem. If we already know the position to within a few meters, alternate computation of the position solution and NLOS corrections may be iterated until they converge. For larger uncertainties, we will need multiple starting positions to ensure convergence. Another approach, known as the “urban trench” method, incorporates the reflecting surfaces within the positioning algorithm. However, this only works where we can determine which surface reflects which signals.

A more powerful approach adds NLOS error prediction to positioning by scoring candidate position hypotheses. NLOS corrections are computed for each candidate position. Both grid-based and particle-based methods have demonstrated positioning accuracies within two meters. However, they are computationally intensive.

These different approaches should not be treated as competitors. SNR-based and pseudorange-based algorithms are clearly complementary. Better performance can be achieved by using both and combining the results. Similarly, a computationally efficient algorithm, operating over a wide search area, can be used to initialize a higher-resolution, computationally intensive algorithm operating over a smaller area.

How Practical Is It?
For 3D mapping–aided GNSS to be practical, the processing load and data storage or transmission load must match the intended applications. The algorithms for determining position will run relatively efficiently on a PC, tablet, or smartphone. However, the algorithms for predicting GNSS signal propagation using 3D mapping are much more computationally intensive. Performing ray-tracing of the of a signal’s path for 20 satellite positions and 1,000 user positions can take several minutes of central processing unit (CPU) time.

This problem has two solutions. The first is pre-computation. For predicting satellite visibility at street level, 3D building boundaries — representing the surfaces and edges of a structure — can be computed for a one-meter grid of candidate positions. Each building boundary comprises the elevation threshold below which satellite signals are blocked for each azimuth. Satellite visibility can therefore be predicted very quickly simply by comparing the satellite elevation with the building boundary elevation at the appropriate azimuth. The main drawback to this approach is that the building boundary data can take up more space than the original 3D mapping.

The building boundary approach can also potentially be extended to predicting whether or not a reflection is received. The path delays of reflected signals could be pre-computed, but this would create a very large amount of data; so, it may not be practical.

The second approach is to use a graphics processing unit (GPU). A GPU is designed for parallel processing, which enables us to consider multiple candidate user positions simultaneously. Using projection, the visibility of 20 satellites over a 100 • 100-meter position grid may be computed in about a second.

To compute the path delays of reflected signals, ray tracing can be run on a GPU, enabling computation of paths from 20 satellites to 100 candidate positions within a second. Operating the GPU on a mobile device increases power consumption because GPUs typically consume about twice the power of CPUs. However, the power consumption of the GNSS receiver chip is also significant. Because of their operating design, GPUs may also speed up the generation of building boundary data.

Distribution of 3D mapping data is another important factor to consider. Three options for distributing these data include pre-loading, streaming, and server-based positioning, as shown in Figure 6. In all cases, a binary data format should be used to minimize the capacity required.

The terrain height data are easiest to handle. A five-meter grid spacing is sufficient, corresponding to 40,000 points-per- square kilometer. Twelve bits is sufficient to describe the relative height of a point within a tile. Four bytes are needed for the height of each tile’s origin with respect to the datum. Thus, about 60 kilobytes are needed to represent a square kilometer of mapped urban terrain; so, one gigabyte of storage could accommodate map data representing about 17,000 square kilometers, Even more can be stored using data compression.

For 3D modeling of a city’s buildings, about 500 bytes might be used to describe a building to LOD 1, and a square kilometer in a dense urban area might typically contain about 1,000 buildings. Thus, about 100 kilobytes per square kilometer is needed; so, one gigabyte of storage could accommodate about 2,000 square kilometers of data, or enough to create 3D maps for about two cities — again, more with compression.

Building boundaries require a lot more data. To achieve a one degree precision, requires about 300 bytes per building boundary. Assuming that about half the space in a city is outdoors (building boundaries are not required for indoor locations), a 100 • 100–meter tile would require 1.5 megabytes of data without compression. So, one gigabyte of storage would only accommodate about 7 square kilometers of data, perhaps 70 square kilometers with compression. Thus, pre-loading may be practical if the 3D mapping is used directly, but is unlikely to be if building boundaries are used. Pre-loading of pre-computed path delays would not be practical.

Moving on to streaming, if 3D mapping is used directly, buildings within a ~300-meter radius of the predicted user position should be downloaded as they could potentially affect signal reception. This download would require about 150 kilobytes of data. For building boundary data, only the search area is needed, which should be no larger than 100 meters by 100 meters, considering only outdoor locations.

Only azimuths corresponding to the current set of GNSS satellites are needed, which reduces the amount of data required to 90 kilobytes without compression. The pre-computed path delays would comprise one value per satellite, per candidate location, requiring about 125 kilobytes for 10-bit values. Third-generation (3G) mobile download speeds are more than 500 kB/s (4 Mbit/s). Therefore, streaming of mapping data is practical, and substantial data buffering could be accommodated to bridge any gaps in communications coverage.

The final option is to calculate the position solution on a remote server. This requires uploading of the GNSS pseudorange and SNR measurements from the mobile device to the server, and then downloading the position solution. This imposes a minimal communications load and could employ current assisted-GNSS (AGNSS) protocols; so, it would be compatible with all current mobile devices. However, it needs continuous communications coverage. A server would almost certainly use pre-computed building boundaries and possibly pre-computed path delays as well, depending on the size of the user base.

UCL’s combined shadow-matching and GNSS-ranging system, which is accurate to about four meters, uses predictions of terrain height and satellite visibility based on the 3D mapping, but not path delay predictions. Consequently, it can operate in real-time with pre-computed building boundary data, or potentially GPU-based projection from a 3D building model. Thus, 3D-mapping-aided GNSS positioning is a practical proposition.

Other research groups have demonstrated higher accuracy (around two meters) using methods that predict NLOS path delay using 3D mapping. These are too computationally intensive to operate in real-time over a large search area on a mobile device. However, they could be used as the final step in a multi-stage positioning process, following on from the UCL algorithms. With further research, further viable approaches are likely to emerge.

The Way Forward
The GNSS research community has demonstrated that 3D mapping–aided GNSS can vastly improve positioning accuracy in dense urban areas and is practical to implement. Current algorithms could be deployed in consumer location-based services right now. This could operate as an enhancement to server-side AGNSS using current interfaces and protocols. A service provider would have to invest in additional hardware to run the positioning algorithms for multiple users simultaneously and store the 3D mapping.

3D mapping–aided GNSS algorithms could also run on a mobile device, potentially within an app, provided the device has a compatible API and GNSS receiver chip. In this case, a service provider would be required to store the 3D mapping data and download it to users on demand. Clearly, either architecture needs a viable business model to fund it. One possible driver is enhancing a mobile navigation app’s user experience to attract more customers.

Many specialist applications — such as navigation for the visually impaired, situation awareness of emergency, security and military personnel and vehicles; emergency caller location; and tracking vulnerable people and valuable assets — could also make use of current 3D mapping–aided GNSS algorithms. With a much smaller user base than consumer applications, much less server infrastructure would be required to get them operational.

The scientific community is another potential early adopter of 3D mapping–aided GNSS. Many research groups use GNSS for tracking experimental subjects or for building maps of things they wish to study, such as pollution or wheelchair accessibility.

As these users post-process their data, they are much easier to support. All they need is some positioning software that they can download plus a set of instructions. They can acquire GNSS receivers, which can log pseudoranges and SNR data, or use smartphones that can access these through the API. They can also use existing OpenStreetMap 3D mapping to select test sites where the data are available.

A lot more can be done to increase the performance of 3D mapping–aided GNSS. This is still a new field and further research will improve the accuracy, reliability, and processing efficiency of these techniques. An integrity framework could also be developed. Therefore, 3D mapping–aided GNSS — potentially integrated with other navigation technologies — could meet the needs of more demanding applications, such as lane-level road positioning and aerial surveillance, as well as better serving consumer, specialist, and scientific users.

3D mapping–aided GNSS could potentially revolutionize positioning in dense urban areas. It’s time to start implementing it.

Acknowledgement
Figures 4 and 5 are provided by Dr. Mounir Adjrad of UCL, whose research is funded by the Engineering and Physical Sciences Research Council (EPSRC) project EP/L018446/1, “Intelligent Positioning in Cities using GNSS and Enhanced 3D Mapping.”

The photos on the opening pages of this article were taken by Dr. Adjrad and former UCL students Nachuan Li and Christelle Mekemlong Lando.

Additional Resources
[1] Adjrad, M., and P. D. Groves, “Intelligent Urban Positioning using Shadow Matching and GNSS Ranging Aided by 3D Mapping,” ION GNSS+ 2016, Portland, Oregon USA, September 2016
[2] Adjrad, M., and P. D. Groves, “Enhancing Conventional GNSS Positioning with 3D Mapping without Accurate Prior Knowledge”, ION GNSS+ 2015, Tampa, Florida USA, September 2015
[3] Bétaille, D., and F. Peyret, M. Ortiz, S. Miquel, and L. Fontenay, “A New Modeling Based on Urban Trenches to Improve GNSS Positioning Quality of Service in Cities,” IEEE Intelligent Transportation Systems Magazine, 5(3), 59-70, 2013
[4] Bourdeau, A., and M. Sahmoudi, “Tight Integration of GNSS and a 3D City Model for Robust Positioning in Urban Canyons,” ION GNSS 2012, Nashville, Tennessee USA, September 2012
[5] Groves, P. D., and Z. Jiang, L. Wang, and M. Ziebart, “Intelligent Urban Positioning using Multi-Constellation GNSS with 3D Mapping and NLOS Signal Detection,” ION GNSS 2012, Nashville, Tennessee USA, September 2012
[6] Groves, P. D., and L. Wang, M. Adjrad and C. Ellul, “GNSS Shadow Matching: The Challenges Ahead,” ION GNSS+ 2015, Tampa, Florida USA, September 2012
[7] Kumar, R., and M. G. Petovello, “Sensitivity Analysis of 3D Building Model-assisted Snapshot Positioning,” ION GNSS+ 2016, Portland, Oregon USA, September 2016
[8] Hsu, L.-T., and Y. Gu and S. Kamijo, “3D Building Model-Based Pedestrian Positioning Method Using GPS/GLOANSS/QZSS and Its Reliability Calculation,” doi 10.1007/s10291-015-0451-7, GPS Solutions, 2015
[9] Isaacs, J. T., and A. T. Irish, F. Quitin, U. Madhow, and J. P. Hespanha, “Bayesian localization and mapping using GNSS SNR measurements,” IEEE/ION PLANS 2014, Monterey, California USA, May 2014
[10] Ng, Y., and G. X. Gao, “Direct Positioning Utilizing Non Line of Sight (NLOS) GPS Signals,” ION GNSS+ 2016, Portland, Oregon USA, September 2016
[11] Obst, M., and S. Bauer and G. Wanielik, “Urban Multipath Detection and mitigation with Dynamic 3D Maps for Reliable Land Vehicle Localization,” IEEE/ION PLANS 2012, Monterey, California USA, May 2012
[12] Suzuki, T., “Integration of GNSS Positioning and 3D Map using Particle Filter,” ION GNSS+ 2016, Portland, Oregon USA, September 2016
[13] Suzuki, T., and N. Kubo, “Correcting GNSS Multipath Errors Using a 3D Surface Model and Particle Filter,” ION GNSS+ 2013, Nashville, Tennessee USA, September 2013
[14] Suzuki, T., and N. Kubo, “GNSS Positioning with Multipath Simulation using 3D Surface Model in Urban Canyon,” ION GNSS 2012, Nashville, Tennessee USA, September 2012
[15] Wang, L., “Investigation of Shadow Matching for GNSS Positioning in Urban Canyons,” Ph.D. Thesis, University College London, 2015. Available here
[16] Wang, L., and P. D. Groves, and M. K. Ziebart, “Smartphone Shadow Matching for Better Cross-street GNSS Positioning in Urban Environments”. 68(3), 411-433, Journal of Navigation, 2015
[17] Wang, L., and P. D. Groves, and M. K. Ziebart, “Urban Positioning on a Smartphone: Real-time Shadow Matching Using GNSS and 3D City Models,” Inside GNSS, Vol. 8, No. 6, pp. 44-56, November/December 2013
[18] Wang, L., and P. D. Groves, and M. K. Ziebart, “GNSS Shadow Matching: Improving Urban Positioning Accuracy Using a 3D City Model with Optimized Visibility Prediction Scoring,” NAVIGATION, 60(3), 195-207, 2013
[19] Yozevitch, R., and B. Ben-Moshe, “A Robust Shadow Matching Algorithm for GNSS Positioning,” NAVIGATION, 62(2), 2015

The post It’s Time for 3D Mapping–Aided GNSS appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

The 2016 Trimble Dimensions user conference and exhibition will take place at the Venetian Hotel in Las Vegas on November 7, 8 and 9.

The annual event gathers users of Trimble’s products including positioning technology for unmanned systems as well as mapping, GIS, surveying, photgrammetry and remote sensing and other technologies of interest to readers of Inside GNSS.

Four hundred and fifty technical sessions and networking events give attendees an opportunity to network widely within and among industry groups.

The 2016 Trimble Dimensions user conference and exhibition will take place at the Venetian Hotel in Las Vegas on November 7, 8 and 9.

The annual event gathers users of Trimble’s products including positioning technology for unmanned systems as well as mapping, GIS, surveying, photgrammetry and remote sensing and other technologies of interest to readers of Inside GNSS.

Four hundred and fifty technical sessions and networking events give attendees an opportunity to network widely within and among industry groups.

If you are an expert in a field covered by the conference and are an experienced presenter in front of large audiences, Trimble will welcome your proposal for speaking at the event.The organizers are accepting abstracts until 20 here.

Early bird pricing ends on July 31.

The post Trimble Dimensions 2016 appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

The German Institute of Navigation’s (DGON) 2016 symposium on inertial sensors and systems, ISS, and gyro technology will take place in Tulla Hall at the Karlsruhe Institute of Technology (KIT) on September 20 and 21.

As modern systems for navigation, localization and guidance are increasingly making use of supporting data from non-inertial sensors, the conference particularly appreciates papers on hybrid systems, those that fuse inertial with GNSS, visual, infrared, radar or other sensors.

The German Institute of Navigation’s (DGON) 2016 symposium on inertial sensors and systems, ISS, and gyro technology will take place in Tulla Hall at the Karlsruhe Institute of Technology (KIT) on September 20 and 21.

As modern systems for navigation, localization and guidance are increasingly making use of supporting data from non-inertial sensors, the conference particularly appreciates papers on hybrid systems, those that fuse inertial with GNSS, visual, infrared, radar or other sensors.

The event is sponsored by DGON, the Institute for Systems Optimization/Karlsruhe Institute of Technology (ITE), the Royal Institute of Navigation (RIN) and IEEE.

The post DGON Inertial Sensors and Systems 2016 appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

But some of the limelight needs to fall a little further from center stage — out there where the Indian Regional Navigation Satellite System (IRNSS) and Japan’s Quasi-Zenith Satellite System (QZSS) are not waiting idly in the wings.

When China joined the GNSS club in 2007, it turned a satnav triumvirate into a quartet.

But some of the limelight needs to fall a little further from center stage — out there where the Indian Regional Navigation Satellite System (IRNSS) and Japan’s Quasi-Zenith Satellite System (QZSS) are not waiting idly in the wings.

If all goes as scheduled — and both India and Japan are moving ahead with increased confidence — the regional systems will substantially enrich the world’s positioning, navigation, and timing (PNT) resources. IRNSS should be completed this year, with seven satellites on orbit. QZSS will have four satellites up by 2018 and seven by the end of 2023.

It’s all too easy to conflate regional systems with GNSS augmentation systems, such as WAAS, EGNOS, or Russia’s SDCM — systems that largely replicate and improve the quality of the underlying GNSS service features. (Inside GNSS probably contributes to this lack of differentiation between the two types of systems by having a single “Augmentation/Region” button on its website.)

But India and Japan have their own augmentation systems in GAGAN (GPS-Aided Geo-Augmented Navigation) and MSAS (MTSAT Satellite-based Augmentation System), respectively. QZSS and IRNSS, in contrast, are essentially autonomous systems that literally operate outside the orbits and frequencies of the world’s four GNSS systems.

These are not trivial considerations.

As last year’s disappearance of Flight MH370 reminded us, the Indian Ocean region in which IRNSS operates is large and largely unchartered. So, too, the Western Pacific surveyed by QZSS. Having expanded GNSS resources available there will bring benefits to nations well beyond those providing the regional services.

Broadly speaking, regional systems do augment GNSS services, it’s true. But they also move beyond them, bringing their own distinctive characteristics and technological opportunities.

QZSS’s L-band Experimental (LEX) signal — to be renamed L6 when QZSS becomes fully operational — is designed to transmit correction messages that enable positioning, navigation, and timing applications that require centimeter-level accuracy.

The L6 signal (at 1278.75 MHz) will incorporate Geo++ State Space Representation (SSR) technology in which relevant ionospheric effects are estimated in real time using reference network observations, making these suitable for single-frequency GNSS receivers.

Similarly, IRNSS transmits a unique hybrid data structure at L5 (1164.45–1188.45 MHz) that provides grid-based ionospheric corrections for single-frequency users, as well as a ranging signal on S-band (2483.5–2500 MHz) — a relative novelty for GNSS systems.

The QZSS program also attempts to address another challenge: the need for indoor PNT capabilities. The ground-based IMES or the Indoor Messaging System uses pseudorandom noise codes (PRNs), operates in the L1 frequency, and has message formats compatible with QZSS and GPS.

Beyond their technical distinctiveness, regional systems ensure a relatively greater control over strategic national assets and the critical infrastructures and applications that rely upon them. This, in turn, strengthens the sovereignty of nations that possess and operate them.

Both Japan and India are full members of the International Committee on GNSS Providers Forum, along with the “Big Four”: the United States, China, Russia, and the European Union.

They have earned their places at the table.

The post Let Us Now Praise appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.

One of the first feature articles I wrote as a newly minted GNSS magazine editor 26 years ago was about an advanced rail traffic management system based on GPS that Burlington Northern, with the help of Rockwell Collins, had designed and implemented.

Headed up by a couple of former NAVSTAR GPS Joint Program Office leaders — Don Henderson and Ed Butt — BN’s Advanced Railroad Electronics System (ARES) demonstrated its effectiveness on 250 miles of BN track in the Mesabi Iron Range from 1987 to 1992. ARES tracked and controlled seven locomotives and three maintenance vehicles, from a control center in Minneapolis, Minnesota.

I titled the article, which ran in the May/June 1990 issue of GPS World, “On Track with GPS.”

Everyone who came to Minnesota to watch ARES in action — including Federal Railway Administration (FRA) and National Transportation Safety Board (NTSB) officials, congressional staffers, shippers, Draper Lab analysts, and railroad executives — agreed that ARES was a fine example of positive train control (PTC). Using a GPS constellation that was only half-built and the much less robust computers and wireless communications of that era, PTC could still help avoid collisions, control train speed, and improve rail traffic efficiency for the nation’s railroads.

Fast-forward 26 years to a 50-mph–rated railroad curve outside of Philadelphia, Pennsylvania, where on May 12 an Amtrak passenger train left the tracks at 106 mph, killed 8 people and injured 200. Meanwhile, every day thousands of railcars — a 4,000 percent increase since 2008 — carrying highly explosive shale oil are being hauled through American cities and along the nation’s waterways and through other sensitive environments.

Positive Train Control — why is this still not a thing 26 years later? Mostly because of strong resistance from the rail industry and weak oversight by federal regulators.

NTSB has included PTC on its “Most Wanted List” every year from the inception of the list in 1990, but then the board doesn’t regulate U.S. railroads. The Rail Safety Improvement Act of 2008 (RSIA) mandated that PTC be implemented on so-called Class I rail tracks by the end of this year. That legislation came about after the collision of a California Metrolink commuter train and a Union Pacific freight train resulted in 25 deaths and 102 injuries.

In the wake of the latest Amtrak accident, FRA officials say they will issue an emergency order to begin implementing a train control system that notifies an engineer when a train exceeds the speed limit and automatically applies the brakes — that is, PTC. After a long series of oil car accidents, U.S. Secretary of Transportation Anthony Foxx ordered railroads to use stronger-walled railcars to transport oil and implement an automatic braking system to control speeds.

Will all this kerfuffle actually get PTC back on track? Well, as NTSB Chairman Christopher Hart told a U.S. House Committee on Transportation and Infrastructure subcommittee in April, not really.

“We know that several rail carriers have stated that they will not meet the 2015 deadline,” Hart said. “This is disappointing.”

At the same subcommittee hearing, Acting FRA Administrator Sarah Feinberg admitted, “Although the railroads subject to the mandate are working diligently towards implementation of PTC systems, FRA is concerned that the vast majority of these railroads will not be able to meet the deadline.”

Currently, PTC systems are in use on Amtrak lines only on the Northeast Corridor in the United States and on the Michigan line between Chicago, Illinois, and Detroit, Michigan.

Oh, I should mention that on May 14, 2012, the FRA issued a final rule that exempted about 10,000 miles — out of about 140,000 miles — of U.S. track from the RSIA’s PTC mandate.

As Mayor Tom Weisner of Aurora, Illinois, where more than 40 oil trains roll through town each week, told NPR news: the new rules are full of holes and do little to protect those who live near the rails.

“I don’t think our federal regulators did the job that they needed to do here,” he says. “I think they, uh . . . wimped out, as it were.”

The post Still Not a Thing, Part 2 appeared first on Inside GNSS - Global Navigation Satellite Systems Engineering, Policy, and Design.