Improved Bidirectional RRT* Algorithm for Robot Path Planning

Area of Science:

• Robotics
• Artificial Intelligence
• Computer Science

Background:

• Traditional bidirectional RRT* algorithms suffer from high randomness, low search efficiency, and numerous path inflection points.
• These limitations hinder optimal path planning in dynamic and complex environments.

Purpose of the Study:

• To improve the traditional bidirectional RRT* algorithm by addressing its shortcomings.
• To enhance path planning efficiency, reduce randomness, and minimize inflection points in planned paths.

Main Methods:

• Constraining random tree expansion using an improved artificial potential field method to guide growth towards the target.
• Biasing sampling points in the target-biased tree towards those in the start-biased tree.
• Optimizing the planned path by extracting key points and segmenting using the dynamic window method.

Main Results:

• The improved bidirectional RRT* algorithm demonstrated shorter path lengths compared to traditional A*, RRT, and bidirectional RRT*.
• Achieved higher path-planning efficiency and significantly fewer inflection points.
• The fused algorithm produced smoother paths with excellent avoidance of temporary obstacles in complex environments.

Conclusions:

• The proposed enhancements effectively overcome the limitations of the traditional bidirectional RRT* algorithm.
• The improved algorithm offers superior performance in terms of path length, efficiency, and smoothness.
• This approach provides a more robust and reliable path-planning solution for robotic applications.