摘要：AbstractThis paper focuses on trackingspatially repeatabletasks. In addition, these tasks arenotnecessarilytemporally repeatablein the sense that thefinitelength of the corresponding time interval maychangewith each repetition. Because of that, the standard Iterative Learning Control (ILC) framework is not directly applicable. Namely, the standing assumption that thefinitelength of the time interval isfixedwith each repetition, is violated. Motivated byhuman motor learning, this paper proposes a Spatial ILC (SILC) framework which leverages the spatial repeatability. In particular, the concept ofspatial projection, closely related totemporal rescaling, is proposed. This allows tospatiallyrelate the relevant information from the past repetition to the present repetition. To demonstrate the proposed framework, a class of nonlinear time-varying systems with relative degree zero is selected. In particular, using contraction mapping technique, it is shown that under appropriate assumptions, the corresponding tracking error converges under the proposed SILC control law. Finally, simulation results support the obtained result.
