Optimum motion control of palletizing robots based on iterative learning

The purpose of this paper is to provide a maximum speed algorithm for serial palletizing robots, which guarantees relatively low system modeling requirements and can be easily implemented in actual applications.

Operation speed is an important index of palletizing robots performance. In order to improve it, features of palletizing motions are analyzed, and a refined iterative learning control algorithm for maximum speed optimization is proposed. The refined algorithm learns to increase local speed when the following error does not exceed a predefined tolerance, unlike conventional applications which make actual output identical to its reference. Furthermore, experiments were developed to illustrate the new algorithm's ability to take full advantage of motor capacity, drive ability and repetitive link couplings to improve palletizing efficiency.

Experiments show that motion time decreases more than 20 percent after optimization.

The new iterative control algorithm can be easily applied to any repetitive handling operations where manipulating efficiency matters.