Global and local path planning of robots combining ACO and dynamic window algorithm

Area of Science:

• Robotics
• Artificial Intelligence
• Computer Science

Background:

• Traditional robot path planning methods struggle with efficiency and adaptability in complex environments with static and dynamic obstacles.
• Effective path planning is crucial for safe and efficient task completion in robotics.

Purpose of the Study:

• To propose a novel global and local path planning method combining improved ant colony and dynamic window algorithms.
• To enhance the optimization ability, convergence efficiency, and obstacle avoidance performance of robot path planning in complex environments.

Main Methods:

• Implemented an improved ant colony algorithm with cone pheromone initialization, adaptive heuristic factor regulation, and division of labor strategies.
• Integrated an improved dynamic window algorithm featuring path direction angle evaluation and dynamic velocity sampling optimization.
• Combined global and local planning approaches for comprehensive path optimization.

Main Results:

• Achieved an average path reduction of 30.18% and an accuracy increase of 98.46% compared to basic ant colony algorithms.
• Demonstrated rapid convergence in grid maps (23rd iteration for 20*20, 81st for 30*30) with reduced path lengths.
• Showcased superior path smoothness (0.94, 0.91, 0.79, 0.65) across four designed environments compared to existing algorithms.
• Effectively avoided dynamic obstacles using the improved dynamic window algorithm.

Conclusions:

• The proposed hybrid path planning method significantly improves efficiency, robustness, and obstacle avoidance capabilities.
• Enhanced autonomous navigation in complex environments, offering adaptive solutions for industrial automation, service, and exploration robots.
• The novel strategies provide a more effective approach to robot path planning challenges.