Fault-Tolerant Sensor Fusion for Autonomous Vehicles 2020-01-0694
Operation of an autonomous vehicle (AV) carries risk if it acts on inaccurate information about itself or the environment. The perception system is responsible for interpreting the world and providing the results to the path planning and other decision systems. The perception system performance is a result of the operating state of the sensors, e.g. is a sensor in fault or being adversely affected by the weather, and approach to sensor measurement interpretation. We propose a trailing horizon switched system observer that minimizes the difference between sensor measurements and the weighted combination of different sensor observation model outputs; the sensor observations models are associated with different sensor operating conditions including faults. The outputs of the observation models are determined using the best estimate of the target dynamics after fusing different sensor measurements. The preferred observer target is a stationary landmark so as to remove noise resulting from tracking of moving targets. Results show the observer identifies the appropriate sensor model in different test scenarios no more than a few sample intervals after the model changes.
Mark Omwansa, Rick Meyer, Zachary Asher, Nick Goberville