Multiobjective path optimization of an indoor AGV based on an improved ACO-DWA.

Area of Science:

• Robotics
• Artificial Intelligence
• Operations Research

Background:

• Automated guided vehicles (AGVs) are increasingly used in industrial and service sectors for efficiency.
• Path-planning remains a significant challenge for AGVs, especially in large, complex environments.
• Existing algorithms struggle with high-quality path generation and efficiency in complex scenarios.

Purpose of the Study:

• To propose an efficient path-planning algorithm for indoor AGVs in large-scale, complex environments.
• To achieve multiobjective path optimization, focusing on shorter paths, fewer turns, safety, and smoothness.
• To validate the algorithm's performance and practicality using simulations and a real-world AGV platform.

Main Methods:

• Developed an improved ant colony algorithm (IACO) for global path planning, prioritizing shorter paths and fewer turns.
• Enhanced the dynamic window approach (IDWA) for local optimization between key nodes, improving path security and smoothness.
• Integrated IACO and IDWA into a hybrid algorithm (IACO-DWA) for comprehensive path optimization.

Main Results:

• The IACO-DWA algorithm demonstrated significant improvements in simulation experiments on two different-scale maps.
• Path length was reduced by 9.9% and 14.1%.
• The number of turns was reduced by 60.0% and 54.8%, respectively, while maintaining path safety and smoothness.

Conclusions:

• The proposed IACO-DWA algorithm effectively addresses the AGV path-planning problem in complex indoor environments.
• The fusion algorithm offers a superior solution for multiobjective path optimization compared to single-algorithm approaches.
• Practical verification on the QBot2e platform confirms the algorithm's real-world applicability and advantages.