IA-MPC supervised by fuzzy logic: an application to a mechatronics system

Depending strongly on the parameter tuning, the performances of model predictive control can be improved using the online process identification. Such an approach can't always ensure optimal results when the process parameters change with time. In this paper, an indirect adaptive model predictive control supervised by fuzzy logic is proposed to solve this issue. The predictive controller parameters are computed on line regarding measurable performance criteria and inequality constraints on control and output signals. A performance comparison in terms of overshoot, stability and rise time is carried out to highlight the relevance of this approach.