Autonomous Driving Framework for Bus Transit Systems Towards Operation Safety and Robustness

Autonomous driving buses for public transportation have significantly different technical specifications from autonomous driving cars in terms of size, kinematics, dynamics, and measures of safety. This paper presents an autonomous driving framework for public bus transit systems with the novelties of optimal sensor deployment, interactive and feasible motion planning, and fault-tolerant robust control. Optimization methods are developed for distributed sensor installation and sensor fusion. Both path planning and system control take into account the kinematic and dynamic constraints of buses as well as the intention and behavior of other road participants. The framework emphasizes the measures of safety at multiple levels. A 6-meter medium-size autonomous driving bus is under development using the proposed methods. Both simulation and experiment results have initially verified the effectiveness and advantages of the proposed framework.