ESDBO: A Multi-Strategy Enhanced Dung Beetle Optimization Algorithm for Urban Path Planning of UGV

Area of Science:

• Robotics and Artificial Intelligence
• Computational Intelligence and Optimization

Background:

• Unmanned Ground Vehicles (UGVs) face complex path planning challenges in urban environments.
• The Dung Beetle Optimization (DBO) algorithm is popular for UGV path planning but suffers from poor convergence accuracy and local optima.
• Existing DBO methods require enhancement for robust performance in dynamic urban settings.

Purpose of the Study:

• To propose a multi-strategy enhanced DBO algorithm (ESDBO) to overcome the limitations of the standard DBO.
• To improve convergence accuracy, global search ability, and local search capabilities for UGV path planning.
• To validate the effectiveness of ESDBO in generating optimal and safe paths for autonomous navigation.

Main Methods:

• Introduced sine mapping in population initialization to enhance solution diversity.
• Developed an adaptive information volatilization mutation strategy for dynamic convergence-global search balancing.
• Designed a multi-mechanism co-evolution strategy to improve local search ability and stability.
• Conducted ablation experiments and benchmark tests (CEC2017) for algorithm validation.
• Performed path planning experiments on Random MAPF benchmark maps.

Main Results:

• ESDBO demonstrated superior optimization ability and convergence accuracy compared to standard DBO.
• The enhanced algorithm showed improved stability and reduced susceptibility to local optima.
• Path planning experiments confirmed ESDBO's ability to generate globally optimal paths with short lengths, minimal turns, and high safety margins.
• Effective performance was observed across varying obstacle densities and map scales.

Conclusions:

• The proposed ESDBO algorithm offers a significant improvement over the traditional DBO for complex urban path planning.
• ESDBO provides an efficient and reliable solution for autonomous navigation of UGVs in challenging environments.
• The multi-strategy enhancements effectively address the convergence accuracy and local optimum issues of the original DBO algorithm.