Model Predictive Control with Variable Prediction Horizon for a System including Time-varying Delay

Abstract

This paper proposes a model-predictive control (MPC) based method to compensate systems with time-varying delay. The proposed method uses a prediction model to estimate the future state of a controlled object. In case the system includes time-varying delay, the model based predictor on the local side is compelled to assume maximum delay for sending packets. However most of the packets will arrive within the maximum delay. In this paper, we propose to utilize the early-arriving packets to generate a reference trajectory for model predictive control on the remote side. The proposal of this paper is a method to deal with varying horizon for model predictive control. The performance of the proposal was evaluated by comparing the proposed method with a conventional method by conducting numerical simulations and experiments with a human operator.