Control method of robot manipulators by inertia identifier without using acceleration signal

Summary form only given as follows. A robust and high speed response position control strategy based on acceleration control in task space has been proposed, in which a disturbance observer in task space is used to control acceleration. Inverse kinematics including an inertia matrix in task space and a transposed Jacobian matrix has been tested. The value of the inertia matrix is set at a nominal value in order to reduce the computation time of the inverse kinematics because the influence of the inertia variation is compensated by the disturbance observer. The nominal value of the inertia matrix in task space is renewed by an estimated inertia from an inertia identifier in order to adapt to wide inertia variation. A new inertia identification method which utilizes a characteristic of the control system to the inertia variation is proposed, The identification method doesn´t need an acceleration test and is not influenced by disturbance. This algorithm can realize high performance control with a short control period. Experimental results show the effectiveness of the proposed strategy.