Intelligent Decision and Path Planning Algorithm of AGV Vehicle based on Deep Learning

Abstract

With the widespread application of intelligent unmanned vehicles, intelligent navigation, path planning, and obstacle avoidance technologies have become important research topics. This paper proposes an AGV (Automated Guided Vehicle) intelligent decision-making and path planning algorithm based on deep learning. This algorithm uses environmental information to trace a route to the target point while avoiding both static and dynamic obstacles, making it adaptable to different environments. By employing a combination of global planning and local obstacle avoidance decision-making, this method addresses the path planning problem with better globality and robustness, and the obstacle avoidance problem with improved dynamism and generalizability, while also reducing iteration time. During the network training phase, traditional algorithms such as PID and A* are integrated to enhance the convergence speed and stability of the proposed method. Finally, various experimental scenarios for navigation and obstacle avoidance were designed in the Robot Operating System (ROS) and simulation program Gazebo. The simulation results verified that the proposed method, which takes into account both global and dynamic aspects, is reliable and that the generated paths and time efficiency have been optimized.