Journal of Bionic Engineering ›› 2025, Vol. 22 ›› Issue (6): 2980-2998.doi: 10.1007/s42235-025-00769-7
Time-dependent Reservoir-based Gait Planning for Enhanced Lower-limb Prosthetic Control and Perturbation Adaptability
Chang Lu1, Yang Lv1, Wen Zhang1, Hao Sun1, Qidi Wu1, Shuai Wang1, Xiaoxu Zhang1,2,3, Jian Xu1
- 1 College of Intelligent Robotics and AdvancedManufacturing, Fudan University, Shanghai 200433, China 2 MOE Frontiers Center for Brain Science, Fudan University,Shanghai 20043, China 3 MOE Engineering Research Center of AI & Robotics, FudanUniversity, Shanghai 200433, China
Time-dependent Reservoir-based Gait Planning for Enhanced Lower-limb Prosthetic Control and Perturbation Adaptability
Chang Lu1, Yang Lv1, Wen Zhang1, Hao Sun1, Qidi Wu1, Shuai Wang1, Xiaoxu Zhang1,2,3, Jian Xu1
- 1 College of Intelligent Robotics and AdvancedManufacturing, Fudan University, Shanghai 200433, China 2 MOE Frontiers Center for Brain Science, Fudan University,Shanghai 20043, China 3 MOE Engineering Research Center of AI & Robotics, FudanUniversity, Shanghai 200433, China
摘要： Gait coordination in lower limbs plays a critical role in maintaining stability of the human body during walking. For transfemoral amputees, the absence of limbs disrupts this coordination, reducing prosthesis control accuracy. Hip-knee coordination mapping offers a feasible solution for lower-limb prosthesis control, involving the generation of a reference trajectory for the knee joint by leveraging information from the hip. However, current reference trajectories are usually derived from static models, which cannot generate reference trajectories robustly when dealing with perturbations. Therefore, this paper introduces a time-dependent model based on the Delayed Feedback Reservoir (DFR) for hip-knee coordination in lower-limb prosthetic control. Experimental results show that DFR outperforms classical gait planning approaches when facing perturbations, achieving a 20% lower Root Mean Square Error (RMSE) and reducing residuals by up to 18.14 degrees. This research contributes to understanding gait mapping approaches and emphasizes the potential of time-dependent models for robust and strong lower-limb prosthetic control. The discovery provides a novel way to enhance the perturbation adaptability of prosthetic control.