Path planning for mobile robots by fusing ant colony optimization and dynamic window approach

Area of Science:

• Robotics
• Artificial Intelligence
• Optimization Algorithms

Background:

• Mobile robot path planning in dynamic environments faces challenges with global optimization efficiency and local obstacle avoidance safety.
• Existing methods often struggle with real-time adaptation to complex and changing surroundings.

Purpose of the Study:

• To develop an improved path planning algorithm for mobile robots that enhances both global optimization and local obstacle avoidance.
• To address limitations in convergence speed, path length, and real-time responsiveness to dynamic obstacles.

Main Methods:

• Proposed a fused algorithm, ACO-DWA-DPP, combining Ant Colony Optimization (ACO) with Dynamic Window Approach (DWA).
• Introduced a potential field force-based heuristic function and pheromone strategy for ACO.
• Integrated a dynamic collision risk coefficient into DWA for improved local obstacle avoidance.

Main Results:

• Reduced longest path length by 41.26%-48.28% and shortest path by 10.68%-12.64%.
• Decreased iterations by 83.37%-89.51% and turning points by 66.94%-81.37%.
• Demonstrated real-time avoidance of unknown obstacles in simulations.

Conclusions:

• The ACO-DWA-DPP algorithm effectively combines global path optimization with local dynamic obstacle avoidance.
• This fused approach provides a feasible and efficient solution for mobile robot path planning in complex scenarios.
• Achieved significant improvements in path efficiency, convergence speed, and safety.