Application of the Improved Rapidly Exploring Random Tree Algorithm to an Insect-like Mobile Robot in a Narrow...

Area of Science:

• Robotics
• Artificial Intelligence
• Path Planning

Background:

• Intelligent mobile robots face challenges in complex, narrow environments, including inefficient search, redundant nodes, non-smooth paths, and high operational costs.
• Existing path planning algorithms like Rapidly-exploring Random Tree (RRT) have limitations in efficiency and path quality for real-world robotic applications.

Purpose of the Study:

• To develop an improved path planning algorithm for mobile robots based on the RRT approach.
• To enhance search efficiency, reduce path length and cost, and ensure path smoothness while adhering to robot constraints.

Main Methods:

• Implemented target bias sampling to eliminate redundant nodes and an adaptive step size strategy to improve RRT's limitations.
• Modeled and analyzed the mobile robot to ensure path adherence to angle and collision constraints.
• Utilized cubic B-spline curves for pruning and smoothing the generated paths.

Main Results:

• The improved RRT algorithm reduced path length by up to 26.42% and planning time by up to 64.96% compared to RRT, RRT-connect, and RRT* algorithms.
• Demonstrated significant improvements in sampling time, node count, and path length, leading to reduced costs and increased speed.
• Successfully validated through simulations and physical testing with an insect-like robot in narrow environments like elevator entrances.

Conclusions:

• The proposed improved RRT algorithm offers a notable advancement in mobile robot path planning for complex and narrow environments.
• The algorithm effectively addresses efficiency, path quality, and cost reduction, outperforming traditional RRT-based methods.
• Physical verification confirms the algorithm's robustness and applicability in real-world scenarios.