MADS-based fast FPGA implementation of nonlinear model predictive control

In this paper, the derivative-free optimization algorithm MADS (mesh adaptive direct search) is adapted for implementation on field programmable gate array (FPGA) with fixed-point data representation. MADS is then exploited to solve constrained nonlinear optimization problems arising from non-linear model predictive control. The application on two examples taken from the literature shows the advantages of the proposed circuit architecture over the existing work, in terms of latency and resource occupation.