Dynamic Matrix Control With Feed-Forward for Target Tracking

This paper proposes a dynamic matrix control (DMC) algorithm that uses the proper future target information to improve the tracking performance. A default choice of feed-forward implies no anticipation and could lead to degeneration of performance. By first illustrating the underlying feed-forward structure in the original DMC scheme without constraints, this paper gives a systematic method for selecting a good choice for advanced knowledge and proposes a feed-forward DMC algorithm with constraint handling. A typical case study demonstrates the efficacy of the proposal.