INVESTIGATING THE IMPACT OF COMPONENT UNCERTAINTY ON AUTONOMOUS VEHICLE OVERTAKING MANEUVERS

摘要

Autonomous vehicles sense the environment and act accordingly using modern electronic and mechanical components. Although these subsystems and components are delicately designed and manufactured, uncertainties in real-world vehicle operations could cause various undesirable performance deviations or even failure. In this work, we study the impact of component uncertainties on vehicle overtaking maneuver, one of the most common actions in traffic. A model is built to simulate the flow of information and the actions of each component in the control strategy of overtaking. Possible variations of major components are investigated. The impact of uncertainties on the overtaking maneuver are first analyzed. Based on the results, we see that the angle resolution of the distance-detecting device has the biggest impact while the angular velocity-detecting device has relatively small influence. While the control strategy changes, we also observe the impact swifts. Therefore, one could systematically use the existing uncertainty to see how output performance changes by using the proposed model; alternatively, one could also study the robustness of the control strategy under uncertainties.