Abstract
A vole foraging-inspired dynamic path-planning method considering slippery road conditions is proposed for wheeled humanoid robots. Glazed and oily roads create a high risk of slipping for wheeled humanoid robots and hinder the realization of high-speed movement. But in a dynamic environment, road conditions such as material, texture, and attachments vary uncertainly in both space and time, and cannot be processed as quickly and easily as moving obstacles. Inspired by the process of voles searching for food, to address this challenge, a slip-risk-assessment method based on time-space decoupling is designed and integrated into a grid-based environmental model. On this basis, the dynamic path-planning model is constructed by combining the cost functions and constraints based on the slip-risk information. A two-level non-periodic cyclical dynamic planning mechanism is proposed based on conditional triggering. It adaptively and cyclically calls the global planning algorithm and the local re-planning algorithm according to the characteristics of environmental changes to autonomously avoid high-slip-risk areas and moving obstacles in real time. The experimental results show the effectiveness and practicality of the proposed planning method.