Stochastic estimation of human forearm and wrist impedance using robots

Abstract

This thesis presents a practical methodology for the estimation of dynamic impedance in human forearm and wrist in order to provide an objective and quantitative measure of abnormal muscle tone and following synergy pattern for the stroke rehabilitation; to provide a quantitative foundation for the robotic therapy of forearm and wrist rotations; to observe the degree of influence of pronation-supination (PS) of the forearm to the global wrist motion.

The proposed method includes the extended usage of the stochastic estimation using the internal model based impedance control (IMBIC) and a gravity compensation scheme, which was developed and used for the experiments with a spatial 3 degree of freedom (DOF) PUMA robot. Before the proposed method was applied to the human forearm and wrist impedance estimation, the stochastic estimation using the IMBIC was validated through the experiments on the spring array and the human arm with a 2 DOF Selective Compliant Assembly Robot Arm (SCARA) robot. Also, the accuracy and reliability of the proposed method were verified by realistic simulation work. After the proposed method with a PUMA robot was validated by conducting preliminary experiments on a mechanical system with a known impedance TFM, the method was finally applied to the estimation of human forearm and wrist impedance. The simulation and preliminary experiment results demonstrated that the proposed method yields accurate and reliable estimations even under substantial frictions: the multiple coherence functions exceeded 0.95 throughout the frequency range investigated. The proposed method has also demonstrated effectiveness in the estimation of human forearm and wrist impedance using conventional robots and the estimated impedance TFM demonstrated reasonable agreement with the results of previous research.

The effect of couplings between PS and FE or RUD on step-tracking wrist rotations was also investigated with the estimated impedance in this thesis. The cou...