Automated Testing of Ultrawideband Positioning for Autonomous Driving

Benjamin Vedder,Jonny Vinter,Bo Joel Svensson,Magnus Jonsson

doi:10.1155/2020/9345360

Benjamin Vedder, Jonny Vinter + Show 2 more

Open Access

https://doi.org/10.1155/2020/9345360

Copy DOI

Abstract

Autonomous vehicles need accurate and dependable positioning, and these systems need to be tested extensively. We have evaluated positioning based on ultrawideband (UWB) ranging with our self-driving model car using a highly automated approach. Random drivable trajectories were generated, while the UWB position was compared against the Real-Time Kinematic Satellite Navigation (RTK-SN) positioning system which our model car also is equipped with. Fault injection was used to study the fault tolerance of the UWB positioning system. Addressed challenges are automatically generating test cases for real-time hardware, restoring the state between tests, and maintaining safety by preventing collisions. We were able to automatically generate and carry out hundreds of experiments on the model car in real time and rerun them consistently with and without fault injection enabled. Thereby, we demonstrate one novel approach to perform automated testing on complex real-time hardware.

Highlights

Accurate positioning is an important technology for autonomous vehicles
E purpose of generating random drivable trajectories is to expose the positioning systems to a wide range of scenarios without having to manually create all the scenarios; instead, some properties on how the scenarios can be created are defined, and tests are generated based on the properties. ese properties include the geometry of the area in which the car is allowed to drive, the driving dynamics of the car, and speed limits. e performance metrics of the tests are how well the uwb- and Real-Time Kinematic Satellite Navigation (RTK-SN)-based positioning systems agree with each other
We have shown how Property-Based Testing (PBT) can be used in combination with Fault Injection (FI) to generate many tests randomly where the golden run can be derived on-the-fly from the model used in the PBT tool [12]. is way functional and nonfunctional requirements can be tested simultaneously using the same test setup, which can reduce the total required testing effort

Summary

Introduction

Accurate positioning is an important technology for autonomous vehicles. A positioning system needs to be both accurate and dependable; there is a need for extensive testing and evaluation. We address this need by automating test case generation in simulations [1] and Hardware-In-the-Loop (HIL) tests [2, 3], and on full-scale hardware To demonstrate this approach, we equip our self-driving model car [4] with an ultrawideband (UWB) positioning system in addition to the Real-Time Kinematic Satellite Navigation (RTK-SN) positioning system it already has and evaluate the performance of the UWB system against the RTK-SN system. Our test method consists of automatically generating random drivable trajectories for our model car, injecting faults into the UWB system and comparing the position outputs of both positioning systems. We have extended the firmware of our model car controller with a position sensor fusion algorithm that merges distance measurements from the UWB module on the car to the anchors with odometry data from the motor controller and heading information from the IMU. Deviations from this assumption degrade the UWB positioning performance, but in our experience, the practical impact on the performance is less than 0.5 m, which is within our requirements

Test-Case Generation

Results

Building

Conclusion