Learning feedforward control

Some new results in learning feedforward control are presented in this paper. The proposed procedures are applicable to processes with or without feedback control. Under the ideal situation, this discrete-time learning control (LC) scheme can perfect the task with one repeat. Convergence analyses are also included for noisy and imprecise learning. New algorithms ore proposed to address these noisy learning issues. Interesting connections between the proposed approach and iterative image deblurring, LQ control and Kalman filtering are identified. With all the required process knowledge readily measurable, the proposed scheme is relatively simple to implement and can be used as an on-site feedforward tuning tool. Encouraging simulation results are included.