Quantitative Tracking Error Analysis and Feedforward Compensation Under Different Model-Based Feedforward Controllers in Different Control Architectures

摘要

This article presents quantitative analysis on the tracking error under different model-based feedforward controllers, such as zero-phase-error tracking control, zero-magnitude-error tracking control, nonminimum-phase-zero-ignore tracking control, in different two-degree-of-freedom control architectures. Analysis shows that, even there is no model uncertainty, extra feedforward terms like velocity, acceleration, jerk, and snap feedforward are required for further compensation, and that noncausal and causal implementation of these model-based feedforward strategies will lead to different tracking errors and require different extra compensation. Based on the analysis, an algorithm to tune the extra feedforward terms is proposed. Simulation and experiment on an ultraprecision motion system both well validate the theoretical analysis and the proposed algorithm.