Path Planning Method Using Dyna-Q Algorithm under Complex Urban Environment

Path planning and obstacle avoidance problems are now the focus of robotics research. This paper uses the Dyna-Q reinforcement learning algorithm to implement an obstacle avoidance and a path planning algorithm for unmanned ground vehicle(UGV) under urban environment. Using the reinforcement learning algorithm, we calculate the waypoints of the unmanned vehicle and achieve obstacle avoidance tasks and path planning using a vector field. Finally, we use a PID controller on unmanned aerial vehicle (UAV) to realize the air-ground collaboration task. The algorithms and the agents' modeling in this paper are implemented in the lab's simulation platform.