Control Strategy for intelligent Walker Human-Robot Shared Control Based on Dynamic Artificial Potential Field and Model Predictive Control

DD MMMM YYYY MMMM YYYY The rising prevalence of Parkinson's disease has drawn significant research attention to patient rehabilitation and safe mobility support. Patients utilizing intelligent walker systems frequently operate in highly dynamic unstructured environments (e. g. , homes, hospitals, shopping malls) containing both static obstacles (furniture, walls) and dynamic obstacles (pedestrians, mobile devices). Conventional navigation algorithms in such complex settings often exhibit sudden curvature changes and inadequate path smoothness, substantially increasing patient mobility risks. To address these challenges, this study proposes a human-robot interaction (HRI) framework integrating user intent recognition with a Dynamic Artificial Potential Field (DAPF) approach, employing Model Predictive Control (MPC) for rolling-horizon optimization to achieve real-time trajectory refinement. To validate the efficacy of the proposed methodology, comprehensive experiments were conducted in virtual simulation. Comparative results obtained on the MATLAB simulation platform demonstrate that compared to baseline algorithms, the proposed method significantly enhances path smoothness during steering maneuvers, achieves a 97. 4% navigation success rate, and the obstacle avoidance success rate approaches 100%. . The system consequently enhances operational safety and convenience while preserving user autonomy through real-time guidance and timely intervention against hazardous maneuvers.