Autonomous Obstacle-Avoiding Movement in Unknown Environments

By performing path planning in a known environment, mobile robots can move autonomously without collisions. In an environment without any information, environmental perception and path planning need to go hand in hand. We conducted research on autonomous obstacle-avoiding movement in unknown environments. This paper proposes a framework for an autonomous obstacle-avoiding movement method in an unknown environment, including the LeGO-LOAM algorithm, which adds scan-context-based loop detection for environmental perception; the Rapidly-exploring random tree (RRT) composite frontiers guide points for sampling observation; the A∗ algorithm, which introduces path smoothness for smooth path planning; and path tracking based on the maximum natural estimation method. Additionally, we carried out experiments on a simulation platform to verify the feasibility and stability of the proposed method.