Active Control method for Postural Tremor via Iterative Learning Algorithm

摘要

This paper presents the behavior and the effective way to suppress the human hand postural tremor using the active force control method. The main advantages of proposed method include; its simple control technique, practically applicable in real-time due to the much lesser computational burden and the extra robustness feature the method generates. The actual tremor data from a Parkinson disease patient is validated through an artificial vibration exciter to investigate a tremor suppression technique based on the Mechatronics approach in which a proportional-derivative (PD) with active force control (AFC) strategy is applied to a single degree-of-freedom (DOF) hand model dynamic system incorporating a Voigt muscle model. A number of iterative learning (IL) algorithms were explicitly employed to compute the estimated mass in the AFC loop that is necessary to trigger the control action. The acceleration and displacement dynamic responses of the human hand model were captured and recorded using the light-weight accelerometer and a laser displacement sensor for validation purpose. The results may be considered as raw data that can be used for further analysis of human tremor especially in Parkinson's disease patients, which can later be used to assist in developing strategies in the design and development of a hand tremor suppression device. A piezoelectric actuator is employed as the main active element within the AFC-based system for the compensation of the disturbances. A number of different conditions were also simulated and rigorously tested for the effectiveness of the proposed scheme to counteract the disturbances and establish the system behaviors. Simulation results show that the proposed AFC-based scheme is very robust compared to conventional PD counterpart.