Assessment of gradient-based point-to-point ILC for MIMO systems with varying interaction

This paper examines the performance of a gradient-based point-to-point iterative learning control (ILC) algorithm applied to multivariable input, multivariable output (MIMO) systems. Whilst ILC is concerned with tracking a reference trajectory defined over a finite time duration, the point-to-point formulation addresses application domains where the output is not critical at all points over the task duration. The algorithm therefore enforces tracking of only an arbitrary subset of points, with the advantage that the convergence rate increases and input energy decreases as points are removed from the reference. Experimental results presented using a MIMO test facility which can be configured with variable levels of input-output interaction and exogenous disturbance/noise injection confirm the theoretical findings.