Hazard-constrained global-local path planning for fault-tolerant hexapod robots on unstructured terrain

Abstract

This study proposes an integrated path-planning framework for fault-tolerant hexapod robots navigating unstructured environments, addressing challenges posed by leg joint failures. The framework combines an enhanced A* algorithm with an adaptive Dynamic Window Approach (DWA) to improve navigation robustness. The A* algorithm incorporates a hazard-based model assessing terrain features like slopes, obstacles, and trenches, optimizing global paths by refining heuristic functions and minimizing path complexity to ensure safety and efficiency. The adaptive DWA dynamically adjusts local trajectories, balancing goal alignment, obstacle avoidance, stability, and energy efficiency through fault-specific evaluations, with weights tuned for optimal performance. Simulations and physical experiments demonstrate that the approach outperforms conventional methods, producing smoother, safer paths and enhancing stability across diverse terrains, even under fault conditions. This framework provides innovative solutions for reliable navigation in complex environments, offering significant potential for applications in search and rescue operations and extraterrestrial exploration, where adaptability and fault tolerance are critical for mission success.