Interactive path planning for teleoperated road vehicles in urban environments

Remote controlling of road vehicles in complex urban scenarios causes a high workload for human operators which may lead to stop-and-go driving behavior. This paper introduces a novel assistance system to improve the autonomy level of teleoperated driving, keeping human as the main decision maker in all driving tasks. The proposed system detects the hazardous driving commands of the operator which may lead to collisions, generates collision free paths using a LIDAR occupancy grid and returns them as feedback to the operator. For this purpose, a new hybrid path planning algorithm has been developed, which in the first phase searches all possible paths in the environment using a modified Rapidly-Exploring Random Tree (RRT) and in the second phase, optimizes the found and clustered paths using numerical approaches. Simulative and real driving tests confirm the capability of the proposed assistance system to support the human operator in complex driving tasks.