Abstract
This paper proposes the force control using the instantaneous state observer (ISOB). In a two-inertia system, a conventional disturbance observer (DOB) is widely used to achieve the robust motion control. In general, the DOB is used for the force sensor-less force control. However, the DOB exhibits an estimation delay attributed to pole allocation. The estimation delay has influence on the performance of the controller. The ISOB using an acceleration sensor is proposed for the estimation load-side torque in this paper. The ISOB does not exhibit the estimation delay because the acceleration signal is used for the input signal of the observer. Therefore, the performance of the force controller is improved by using the ISOB. On the other hand, the measured acceleration signal is susceptible to the effect to noise. To address the noise problem, the variable noise-covariance (VNC) Kalman filter is used to the estimation of the load-side acceleration. Hence, the noise is suppressed by using the VNC Kalman filter. Thus, the high performance load-side torque response is achieved by using the VNC Kalman-filter-based ISOB. The effectiveness of the proposed method is confirmed by performing the numerical simulation using an industrial robot.
