확률적 위험도 평가가 고려된 자율 주행 차량에서의 지역 경로 계획 전략

Abstract

There are numerous latest technologies in autonomous driving. Among them, path planning studies have contributed greatly to avoiding collisions with static objects and finally reaching the target point. However, for this path planning to be applied to vehicles and more intelligent robots, both the effects of dynamic objects (E.g. surrounding vehicles) as well as static, convenience, and safety must be considered. In this study, the risk level of the ego and adjacent lanes is analyzed based on the surrounding vehicle information measured by sensors, and this is optimized and reflected as one of the factors that determine the local path of the vehicle. The main contribution of this study is as follows. First, it is possible to prevent selection even if a path candidate occurs in an area that is not the actual by reflecting the lane information from the risk assessment of the lane to the local path candidate. Second, a path planning strategy can be established considering the situation of lateral rearward regions where local path candidates are not created. As a result, it is expected that vehicles equipped with this path planning strategy will be able to plan safer paths than before.