Learning and Ha, Control of an Overhead Crane for Obstacle Avoidlance and Disturbance Rejection

A control strategy is developed for a 3-dimensional overhead crane. Using the differential flatness of the crane and a parameterization method, we first calculate the optimal trajectory of the payload that results in minimum transfer time when there exist obstacles in the direct moving path. A learning algorithm is then used to generate the desired feed forward input signal that can drive the system output to track the optimal trajectory. A key feature of this strategy is that it is model free and thus is robust to uncertainties in modeling and parameters. When there exist external disturbances, an H,, optimal control method is used to reject the disturbances. Simulation results are given to verify the strategy and compare some performances.