Optimization Method for Disturbance Compensation in Bilateral Control

抄録

Recently, real-world haptics which deals with real-world tactile sense has been actively researched. In real-world haptics, tactile information is extracted by a disturbance observer. In other words, tactile information is dealt with by an actuator as a disturbance force. Therefore, the disturbance of an actuator has to be minimized for the extraction of clearer tactile information and for the improvement of operationality. The identification and compensation of actuator's disturbance are necessary for improvement of operationality. However, in general, it is difficult to identify the disturbance because the disturbance has nonlinearity such as friction. In addition, the low accuracy of identification induces an overcompensation to the actuator.
In this paper, the method of improving operationality using optimization problem is proposed. A constraint optimization problem is constructed for the disturbance identification and compensation without overcompensation. The optimization problem is solved by a conjugate-direction method which involves an algorithm of nonlinear optimization. In addition, the proposed method can determine the parameter values of disturbance using only 10s examination and automatic off-line processing. Therefore, considerable time and effort are not needed in the proposed method. The effectiveness of the proposed method is confirmed by experimental results.