Optimal iterative feedforward parameter tuning for varying trajectory under actuator constraints

This paper presents a feedforward control algorithm that combines the benefits of optimal iterative learning control (OILC) and model-based feedforward control (MFC) using iterative feedforward tuning and input shaping filter (IFT-ISF) for industrial motion systems. OILC effectively compensates for tracking errors in repeating tasks under actuator constraints. However, its performance deteriorates when the trajectory changes. In contrast, MFC can achieve high performance for varying trajectory tracking tasks, but its performance may degrade for constrained systems if the control force exceeds the actuator saturation boundary. The proposed algorithm aims to overcome these limitations to achieve optimal trajectory tracking performance for varying trajectories under actuator constraints. Simulation and experimental results demonstrate that the proposed algorithm achieves optimal tracking performance while complying with the actuator constraints. The algorithm provides a data-driven approach without requiring the tedious process of model identification. By combining the benefits of OILC and IFT-ISF, the proposed algorithm can achieve high-performance trajectory tracking for both repeating and varying tasks under actuator constraints, making it suitable for industrial motion systems.