Multi-objective trajectory tracking optimization for robots with elastic joints

We consider the trajectory tracking optimization for an elastic joint robot with noisy joint-velocity readings and actuation torque limits. We propose an approach targeting the minimization of the trajectory tracking error, of the velocity noise amplification on the control commands, and of the trajectory execution time. We integrate the use of a feedforward/feedback trajectory tracking control scheme with efficient and accurate simulations of the controlled system dynamics. These are used within a genetic algorithm optimization scheme for optimizing the control gains and trajectory execution time. We verify the effectiveness of our proposed approach using simulations.