Research on visual sensor occlusion test tools for autonomous vehicles

Abstract In the research on the safety of intelligent driving vehicles, it is essential to test and verify the safety redundancy characteristics of the vehicle’s autonomous driving system. The safety redundancy capability of autonomous vehicles comes from the redundancy at the vehicle control level on the one hand, and the redundancy of environmental recognition sensors on the other. Among them, the camera sensors of the vehicle are generally distributed at multiple points inside and outside the vehicle to identify the external environment of the vehicle and the status of the driver in the vehicle. When some camera sensors on the outside of the vehicle are blocked while the vehicle is moving, the autonomous driving system should have a strategy to deal with it to avoid danger. For the verification of the above test scenarios, it is necessary to use the visual sensor occlusion tool to complete. In this paper, how the designed exterior vision sensor occlusion tool works normally under emergency braking conditions is studied.