New criteria of exciting trajectories for robot identification

Inertial and friction parameters of robots can be estimated using the minimal linear energy model in relation to the parameters and least squares techniques or a Bayesian approach when a priori statistical information (expected value, covariance matrix) on the solution vector are available. These techniques are sensitive to noise measurements and error modeling, and require the use of exciting trajectories. New criteria are presented taking into account a priori information on the solution to generate exciting trajectories. An application for a two degrees of freedom SCARA robot is simulated to show the efficiency of these methods