LEARNING OF A CONTROLLER FOR NON-RECURRING FAST MOVEMENTS

In this paper a learning method is described which enables a conventio nal industrial robot to accurately execute the teach-in path in the pr esence of dynamical effects and high speed. After training the system is capable of generating positional commands that in combination with the standard robot controller lead the robot along the desired traject ory. The mean path deviations are reduced to a factor of 20 for our te st configuration. For low speed motion the learned controllers' accura cy is in the range of the resolution of the positional encoders. The l earned controller does not depend on specific trajectories. It acts as a general controller that can be used for non-recurring tasks as well as for sensor-based planned paths. For repetitive control tasks accur acy can be even increased. Such improvements are caused by a three lev el structure estimating a simple process model, optimal a posteriori c ommands, and a suitable feedforward controller, the latter including n eural networks for the representation of nonlinear behaviour. The lear ning system is demonstrated in experiments with a Manutec R2 industria l robot. After training with only two sample trajectories the learned control system is applied to other totally different paths which are e xecuted with high precision as well.