A Novel Neurodynamics-Based Approach to Fatigue-Aware Motion Planning in Human–Robot Collaborative Assembly

Abstract

This paper proposes a neurodynamics-based motion planning framework for fatigue-aware human–robot collaborative assembly (HRCA). The proposed framework integrates human fatigue information into robot motion planning to generate fatigue-dependent trajectories with adaptive clearance from human workspaces. In the proposed framework, the vision-derived fatigue-related index is constructed from observed eye and mouth behaviours, which is incorporated into the inhibitory inputs of the bio-inspired neural network (BINN). Instead of relying on a fixed safety margin, the proposed motion planner is able to produce a fatigue-dependent inhibitory field around the human workspace based on the fatigue input and the evolving neural activity landscape without training a navigation policy. In addition, a nonholonomic neural connection strategy is designed to ensure that the generated trajectory satisfies the kinematic constraints of the mobile robot. Simulation studies and real-world experiments show that the robot uses the available narrow passage under low-fatigue conditions and selects a larger-clearance detour under high-fatigue conditions.