MIMO Closed Loop Identification of an Industrial Robot

Abstract

This paper proposes a practical multi-input multi-output (MIMO) closed loop parameters identification procedure for robot manipulators. It is based on the weighted least squares (WLS) method, coupled with particular solutions to facilitate the estimation, reducing the noise effect. More precisely, a two steps procedure to reduce the condition number of the input data matrix with optimal trajectory planning, and a method to estimate the variances matrix to be used as a weight matrix for the WLS method are illustrated. Moreover, the identification problem is solved with reference to an MIMO coupled system. A closed loop identification is needed because the system is open loop unstable, and because the robot has to track an optimal reference input so as to correctly execute the identification procedure. Some solutions are also presented to overtake common identification problems, such as the bias of the estimated parameters, the presence of outliers, the necessity of balancing the kinematics of the third link, and the reduction of the sensitivity to noise of the estimate. The presented procedure has been successfully experimentally tested on a COMAU SMART3-S2 industrial manipulator used in a planar configuration.

Robust motion control of a robot manipulator via Integral Suboptimal Second Order Sliding Modes

“…Note that B(q) and n need to be identified on the basis of experimental tests. In our work, we assume that the identified B(q) coincides with the actual one (or it is a quite accurate replica), which, on the basis of our experience, is often true in practice, whilen is an estimate of n, which does not necessarily coincide with n. In the following we make reference to the experimentally identified B(q) andn in [28]. By applying the feedback linearization to system (27), (28), one obtains…”

“…The inverse dynamics of a rigid robot manipulator can be written in the joint space as a non linear relationship between the plant inputs and the plant outputs, relying on (27) and (28), so that the control law results in being…”

A Supervisory Sliding Mode Control Approach for Cooperative Robotic System of Systems

“…Note that B(q) andn need to be identified on the basis of experimental tests. In our work, we assume that the identified B(q) coincides with the actual one (it is a quite accurate replica), which, on the basis of our experience, is often true in practice, whilen is an estimate of n, which does not necessarily coincide with n. In the following we make reference to the experimentally identified B(q) andn in [21]. By applying the feedback linearization to the system (2)-(3), one obtains…”

“…where y is an auxiliary control variable obtained as in (21). Note that B(q) andn need to be identified on the basis of experimental tests.…”

Nonlinear Friction and Dynamical Identification for a Robot Manipulator with Improved Cuckoo Search Algorithm

“…M(q) represents × inertial matrix, C(q,q ) is a -dimensional vector including Coriolis and centrifugal forces, and G(q) is -dimensional gravity vector. Equation (1) except for friction torques can be rewritten as a linear form if using the modified Newton-Euler parameters [13] or the barycentric parameters [11]:…”