Adaptive Compliance Shaping with Human Impedance Estimation

Human impedance parameters play a key part in the stability of strength amplification exoskeletons. While many methods exist to estimate the stiffness of human muscles offline, online estimation has the potential to radically improve the performance of strength amplification controllers by reducing conservatism in the controller tuning. We propose an amplification controller with online-adapted exoskeleton compliance that takes advantage of a novel, online human stiffness estimator based on surface electromyography (sEMG) sensors and stretch sensors connected to the forearm and upper arm of the human. These sensor signals and exoskeleton position and velocity are fed into a random forest regression model that we train to predict human stiffness, with a training set that involves both movement and intentional muscle co-contraction. Ground truth stiffness is based on system identification in essentially perturburator-style experiments. Our estimator's accuracy is verified both by the offline validation results and by the stability of the controller even as stiffness changes (a scenario where the ground truth stiffness is not available). Online estimation of stiffness is shown to improve the bandwidth of strength amplification while remaining robustly stable.