Corridor-Based Shared Autonomy for Teleoperated Driving

Abstract

Abstract Given the on-going development of powerful hardware, software and algorithms for automated driving, the number of tasks that vehicles can solve autonomously steadily increases. However, fully autonomous driving in all situations is highly demanding and currently not feasible yet. It may happen that the vehicle faces situations in which the decision system is overstrained and, hence, falls back into a predefined safe state. In the future, moreover, neither control interfaces like a steering wheel and foot pedals nor a qualified driver may be available to assist the car in a blocking situation. This contribution, hence, presents a concept for teleoperated driving, i.e., a remote operation with distinct human machine interactions that explicitly addresses such highly complex driving tasks. It quickly copes with such undesired blocking situations in order to minimize the need for a both time- and cost-intensive road assistance. The concept focusses on teleoperated driving of road vehicles in urban environments. Based on the methodology of a shared autonomy, a corridor-based planning scheme is derived. The remote operation task takes advantage of a fusion of automated driving functions and human-predefined corridors. Within this specified corridor, a path planning algorithm using dual projected Newton method determines a collision-free path that the vehicle is capable to follow. Simulation results show the effectiveness of the proposed method and highlight the achieved driving safety.