Image: NatalyaBurova/iStock/Getty Images Plus/Getty Images
Of the hundreds of papers researchers presented at the Institute of Navigation’s annual ION GNSS+ conference, which took place virtually Sept. 21–25, the following four focused on autonomous vehicle positioning for automobiles on city streets. The papers are available at www.ion.org/publications/browse.cfm.
Digital Maps with Tethered Positioning
The authors propose a new method for tight integration of digital map and dead-reckoning (DR) system (inertial measurement unit plus wheel odometer) to provide reliable navigation solutions in challenging GNSS environments for extended periods. Integrated DR and GNSS have been widely used as the backbone of any navigation system for the internet of things (IoT) and vehicle navigation applications. Dollar-level micro-electro-mechanical system (MEMS) inertial measurement units (IMUs) aided by vehicle-wheel odometers have been recently used as low-cost DR systems to bridge GNSS gaps in harsh environments, such as urban canyons, tunnels and under bridges.
However, DR drift errors rapidly increase over time and cannot satisfy most IoT and land-vehicle navigation requirements. Plus, the GNSS receiver may fail to provide accurate position or even experience a complete outage for more than 15 minutes, causing the tethered positioning error to reach several hundred meters. Because land vehicles are supposed to travel on roads, feedback from a digital map can be used to constrain their position.
The authors used a fuzzy-logic map-matching algorithm to identify the correct road segment on which the vehicle moves. A feedback filter senses a correct map-matched position as well as the road segment as measurement updates to the Kalman filter (KF) of the tethered positioning system. The proposed tight integration of digital maps and a DR system is evaluated using datasets collected by Profound Positioning Inc. in Calgary, Alberta, Canada. Results show the proposed method has an average of 0.15% of relative horizontal position error for Calgary datasets — a considerable improvement over the tethered-solution-only with 3.3% of relative horizontal position error. The average azimuth error of the proposed system is 1.3 degrees, while the tethered positioning system shows an average azimuth error of 9.7 degrees.
Citation. Yashar Balazadegan Sarvrood, Haiyu Lan, Aboelmagd Noureldin, Naser El-Sheimy and Profound Positioning Inc., Calgary, Alberta, Canada. “Tight Integration of Digital Map and Tethered Positioning and Navigation Solution for IoT applications and Land Vehicles.”
5G Signals for Opportunistic Navigation
This paper presents a navigation framework in which 5G signals are used for navigation purposes in an opportunistic fashion. A carrier-aided code-based software-defined receiver (SDR) produces navigation observables from received downlink 5G signals. The SDR produces navigation observables from 5G signals and a navigation filter in which the observables are processed to estimate the user equipment’s position and velocity.
An experiment was conducted on a ground vehicle to assess the navigation performance of 5G signals. In the experiment, the vehicle-mounted receiver navigated using 5G signals from two 5G base stations (also known as gNodeBs, or gNBs) for 1.02 km in 100 seconds. The proposed 5G navigation framework demonstrated a position root-mean-squared error of 14.93 m, while listening to signals from only two gNBs.
Citation. Ali A. Abdallah, Kimia Shamaei and Zaher M. Kassas, “Assessing Real 5G Signals for Opportunistic Navigation.”
Using Low-Cost Onboard Sensors
For autonomous vehicles, accurate positioning must be ubiquitous — reliably available at all times and in all places in which the vehicle is expected to operate. While GNSS commonly provides the basis for absolute positioning, it suffers from the problem of availability whenever a direct view of enough satellites is not possible. To address this failure mode, additional complementary sensors can be added to the overall navigation solution through a technique known as sensor fusion. Sensors such as inertial measurement units (IMUs), cameras, lidars, radar and more can be selected in such a way that the individual shortcomings of each sensor are mitigated, and the overall robustness and reliability are improved.
Although current autonomous-vehicle applications employ sensor-fusion techniques, they tend to rely on high-performance sensors to meet the accuracy requirements. These high-performance sensors tend to induce a much higher cost burden than would be acceptable for commercial production, and therefore make mass autonomy too expensive.
This paper focuses on using the lower cost sensors already available on most modern vehicles. These include low-resolution odometry and consumer-grade IMUs currently used for dynamic stability control and wheel-slip detection. A novel approach for combining vehicle speed, steering angles, transmission settings and multiple odometry inputs is presented along with achievable results while operating under a GNSS-denied environment. The test trajectory mimics a typical parking structure with many corners and short, straight segments. The only a priori information required for the filter is the wheel track and wheelbase (separation distance of the wheels).
A 90% performance improvement compared to the stand-alone GNSS/INS solution was observed during GNSS outages of up to 30 minutes. Furthermore, up to a 50% improvement was observed when comparing the multi-odometry to the single-odometry outages during the same 30-minute outage condition. Beyond GNSS outage performance, this paper shows how the use of the extra input to the filter can improve the positioning system’s protection levels to allow for more frequent engagement of the autonomous navigation system.
Citation. Ryan Dixon, Michael Bobye, Brett Kruger and Jonathan Jacox, “GNSS/INS Sensor Fusion with On-Board Vehicle Sensors.”
Radar and INS/GNSS
An autonomous vehicle requires a ubiquitous, accurate, precise and reliable localization system. Many sensors can be used for positioning and navigation, each with its strengths and weaknesses. Inertial measurement units (IMU) are usually used to build inertial navigation systems (INS). INS can be accurate for short durations; however, an INS accumulates errors and loses its accuracy quickly, especially when using low-cost MEMS-based sensors. GNSS can provide an absolute position and velocity to update the INS over time. A barometer provides absolute elevation information, and an odometer provides a speed update.
An integrated navigation solution consisting of an IMU, a GNSS-RTK receiver and odometer can perform well in open-sky areas and on highways. This system can achieve lane-level accuracy most of the time based on the condition of the sensors and the quality of the measurements. However, in downtown and urban environments, the degradation, multipath and blockage of the GNSS signal leads to poor performance for such an integrated navigation system, which is challenged to maintain lane-level positioning.
This paper presents a version of AUTO (formerly known as Coursa Drive), a real-time integrated navigation system that provides an accurate, reliable, high-rate and continuous navigation solution for autonomous vehicles by integrating INS, RTK GNSS, odometer and radar sensors with TomTom’s HD Maps. AUTO performs a tight nonlinear integration of the radar data and maps with the INS/GNSS/odometer system.
Results demonstrate that radar measurements and HD Maps can be tightly integrated with INS/GNSS in an effective manner, such that the integrated system can provide a high-rate, accurate, reliable and robust navigation solution. This is a crucial requirement for realizing a fully autonomous vehicle that can operate in urban environments under a wide range of conditions, including adverse weather and lighting conditions, even in downtown areas with degraded or denied GNSS signals.
Citation. Abdelrahman Ali, Billy Chan, Amr Shebl Ahmed, Medhat Omr, Dylan Krupity, Qingli Wang, Amr Al-Hamad, Jacques Georgy and Christopher Goodall, “Tight Coupling Between Radar and INS/GNSS with AUTO Software for Accurate and Reliable Positioning for Autonomous Vehicles.”