Model-Based Adaptive Control of Transfemoral Prostheses: Theory, Simulation, and Experiments

This paper presents and experimentally implements three different adaptive robust controllers as the first steps toward using model-based for transfemoral prostheses. The goal of this is to translate these control methods robotic domain, from bipedal walking prosthesis walking, including a rigorous stability analysis. human/prosthesis system modeled two-domain hybrid asymmetric system. An optimization problem formulated obtain stable human-like gait. proposed are then developed combined model optimized reference all proven Lyapunov theorem, ensuring convergence desired verified on robot in simulation. They tested treadmill with an able-bodied subject third iteration AMBER Prosthetic (AMPRO3), custom self-contained powered prosthesis. Finally, outdoor tests carried out AMPRO3 test subjects level ground, uphill slopes, downhill slopes at slope angles 3° 8°, demonstrate real-world environments.