Dynamic Viewpoint-Based Obstacle Avoidance of Robotic Arms in Complex Environments

To address the issues of extensive calculation and low efficiency in path planning for the robotic arm of a mobile exploration robot operating in complex environments, a local environment construction strategy for path planning is proposed. In this strategy, the center of the robot base is used as the viewpoint of the robot at the current working position, and the viewpoint varies depending on the robot's designated working positions and target objects. The local environment was constructed using the robot viewpoint and the robotic arm's maximum range of motion. During the robot's task, the viewpoint and its corresponding local environment will change as the robot's working position changes. Path planning for the robotic arm will be performed in the current local environment rather than in the global environment. This method can significantly reduce the construction of unnecessary models during the obstacle avoidance process, thereby improving the efficiency of collision-free path planning. In this paper, we use the axis-aligned boundary box (AABB) based on octree partitioning for collision detection analysis and conduct simulation experiments on collision detection for a robotic arm. The experimental results show that, compared with traditional path planning based on the global environment, the local environment strategy based on dynamic viewpoints can significantly improve processing efficiency.