Dynamic Path Planning for Forklift AGV Based on Smoothing A* and Improved DWA Hybrid Algorithm.

Area of Science:

• Robotics
• Artificial Intelligence
• Intelligent Transportation Systems

Background:

• Automated Guided Vehicles (AGVs) are crucial in modern intelligent storage environments.
• Efficient and safe path planning for AGVs is essential to avoid obstacles and maintain smooth operation.
• Existing algorithms may struggle with dynamic obstacles and maintaining global optimality.

Purpose of the Study:

• To develop a hybrid dynamic path planning algorithm for AGVs in intelligent storage.
• To improve the global path planning capability for AGVs.
• To enhance the local path planning to avoid sudden obstacles while adhering to the global path.

Main Methods:

• Proposed a hybrid algorithm combining an improved A* algorithm for global path planning and an improved Dynamic Window Approach (DWA) for local path planning.
• Improved A* algorithm enhances suitability for AGV path requirements.
• Improved DWA evaluation function and rolling window method enable dynamic obstacle avoidance.

Main Results:

• The hybrid algorithm successfully plans global paths suitable for AGVs.
• The improved DWA effectively ensures local paths align with the global optimal path.
• Simulations demonstrated the algorithm's ability to dynamically avoid obstacles without significant deviation from the global path.

Conclusions:

• The proposed hybrid dynamic path planning algorithm is effective for AGVs in intelligent storage.
• The algorithm achieves dynamic obstacle avoidance while maintaining proximity to the global optimal path.
• This approach enhances the safety and efficiency of AGV operations in complex environments.