Escape Path Planning for Unmanned Surface Vehicle Based on Blind Navigation Rapidly Exploring Random Tree* Fusion...

Area of Science:

• Robotics
• Autonomous Systems
• Navigation

Background:

• Unmanned Surface Vehicles (USVs) require robust navigation strategies for autonomous operation.
• Escaping constrained environments presents significant challenges for USV path planning.
• Existing algorithms may struggle with efficiency and success rates in complex scenarios.

Purpose of the Study:

• To develop an efficient path planning algorithm for USVs to autonomously escape constrained environments.
• To enhance the RRT* (Rapidly Exploring Random Tree*) method for improved navigation in confined spaces.
• To minimize sensor requirements for autonomous escape maneuvers.

Main Methods:

• Proposed the Blind Navigation Rapidly Exploring Random Tree* (BN-RRT*) algorithm, an adaptation of RRT*.
• Integrated GPS positioning and collision sensor data for navigation.
• Combined Artificial Potential Field (APF) for directional guidance and an active collision mechanism for obstacle identification.
• Implemented an obstacle memory mechanism to prevent revisiting erroneous areas.

Main Results:

• The BN-RRT* algorithm demonstrated significant improvements in planning speed compared to RRT, RRT*, and APF-RRT*.
• The proposed method achieved a notable reduction in path length for escape maneuvers.
• Experimental validation in MATLAB confirmed the algorithm's effectiveness.

Conclusions:

• BN-RRT* offers a more efficient and successful approach for USVs navigating out of constrained environments.
• The integration of APF and active collision mechanisms enhances directional control and obstacle avoidance.
• The obstacle memory mechanism improves the reliability of the escape path planning process.