Experimental Results of a Learning Controller Applied to Tip Tracking of a Flexible Beam

This paper presents the experimental results of a learning controller applied to the tip trajectory tracking of a flexible beam. The overall controller consists of a feedback portion to ensure closed loop stability and a feedforward portion obtained via off-line learning to improve the tracking performance. Neither the model information nor the states of the flexible beam are required for the feedforward learning. Experimental results have verified that the tracking performance of the tip position is satisfied while maintaining the boundedness of the internal states.