Successive constrained optimization and interpolating in non-linear model based predictive control

摘要

Interpolation involving constrained optima offers distinctive advantages for linear and non-linear model based predictive control algorithms. For non-linear systems the computation of minima is non-trivial and sub-optimal solutions can be sought through linearization about particular trajectories. Here we show that successive linearizations and locally constrained optimizations can, at a modest extra computation, lead to significant improvements.