Abstract:

Due to the complexity and variability of the desert environment, the key to the high-efficient of mobile robot is how to avoid obstacles and plan its path. To solve the problems of poor search efficiency and slow convergence of deep reinforcement learning algorithm in complex environment, an improved deep reinforcement learning path planning algorithm is proposed. The exploration factor is improved and dynamically adjusted according to the convergence degree of the algorithm, so that the exploration factor dynamically decreases with the increase of the understanding degree of the agent to the environment, thus speeding up the convergence speed of the algorithm. To improve the search efficiency, a dynamic reward function is set up, the quadratic function is applied to its settings to obtain different reward values by selecting various actions. Simulation results show that compared with the original algorithm, the improved algorithm reduces the path length, iteration times, and planning time by 11.9%, 32.6%, and 17.4% respectively, more adapting to complex environment.

Key words: path planning, robot, deep reinforcement learning, exploration factor, reward function