A Local Dynamic Path Planning Approach for WMRs Based on Fuzzy Dual CHOMP

Abstract

Wheeled mobile robots (WMRs) are widely utilized in exploration tasks on different planets. The path planning technique is required to guide a WMR to a target point or region. A new path planning method in a dynamic environment is proposed. The method improves the original dual Covariant Hamiltonian Optimization for Motion Planning (CHOMP) method by a fuzzy reasoning module and Nonmonotone Gradient Projection Algorithm (NGPA). Thus, this method can achieve low cost paths while maintaining the accessibility of each path nodes. The method to establish local dynamic cost map based on a new velocity-based artificial potential field (VAPF) is proposed as the environment for path planning. Thus, the low cost paths can guide a WMR away from obstacles. Results of the simulations show that the path planning method can dynamically generate a smooth path for a WMR to avoid obstacles.