Global risk assessment in an autonomous driving context: Impact on both the car and the driver

摘要

Highly automated driving (HAD) is a part of the future ways for the "intelligent" road mobility. In this framework, some studies have shown that drivers' situational awareness decreases when using HAD. In this HAD context, drivers can engage non driving tasks as reading or sleeping. These non-driving tasks lead to increased reaction time in case of hazardous situations or risky events (hardware, sensor, actuator failures, or front obstacle or crashes, or dense traffic congestion, or adverse conditions). In this paper, a global risk indicator is proposed using local information coming from surrounding vehicles or infrastructures (V2X communication). This paper shows firstly the gain of such a global risk indicator comparatively to the local one, and secondly its impact on the behaviour of both the autonomous car and the driver.