Iterative Learning Control with Optimal Weight Vector for Incompatible Multisensor Systems

For the multiobjective tracking problem in incompatible multisensor systems, there exist conflicts among the tracking performance indexes of different sensors. A commonly used solution is to combine all performance indexes into a single index by weighted summation. However, the selection of weights has always been a crucial issue. In this paper, we extend the recently developed weighted iterative learning control framework and optimize the weight vector within given incompatible multiple tracking tasks, leading to significant performance improvements, e.g. energy reduction. The problem is formulated into an optimization framework with weight constraints and a reference tracking requirement, for which a two-stage design approach is developed. This leads to an algorithmic solution that minimizes input energy based on iterative learning control and gradient minimization. Numerical simulations verify the validity and effectiveness of the proposed approach.