A Novel AGV Path Planning Approach for Narrow Channels Based on the Bi-RRT Algorithm with a Failure Rate Threshold

Area of Science:

• Robotics
• Artificial Intelligence
• Path Planning Algorithms

Background:

• Rapidly Exploring Random Tree (RRT) algorithms struggle with efficiency in narrow-passage environments, leading to slow convergence and high path costs.
• Existing algorithms often fail to adapt effectively to complex, constricted spaces, limiting their practical application in robotics.

Purpose of the Study:

• To introduce a novel path-finding algorithm, Narrow-Channel Path-Finding Algorithm (NCB-RRT), designed to overcome RRT limitations in constricted environments.
• To improve convergence speed, reduce path costs, and enhance the success rate of path planning in narrow channels.

Main Methods:

• Developed NCB-RRT based on Bi-RRT, incorporating a Research Failure Rate Threshold (RFRT) concept.
• Implemented a three-stage search strategy with real-time sampling failure rates and a balance strategy for dual-tree searching.
• Integrated parent node re-selection, path pruning, and segmented quadratic Bezier curves for path optimization.

Main Results:

• NCB-RRT demonstrated superior adaptability to complex narrow channel environments compared to RRT, RRT*, and Bi-RRT.
• Achieved significant improvements in path length, reduced redundant nodes and inflection points, and enhanced trajectory smoothness.
• Reported success rate increases of 2400% (vs. RRT), 1900% (vs. RRT*), and 11.11% (vs. Bi-RRT).

Conclusions:

• NCB-RRT offers a robust solution for efficient path planning in challenging narrow-channel environments.
• The algorithm's novel strategies lead to faster convergence, shorter paths, and smoother trajectories, outperforming existing methods.