Adaptive model predictive control in the IPA-SQP framework

In this paper, we propose an approach and a specific algorithm to integrate a parameter estimation with the receding horizon model predictive control. We derive this adaptive MPC algorithm based on the integrated perturbation analysis and sequential quadratic programming (IPA-SQP) framework. Previously this approach was exploited for repeated constrained optimization in MPC when the initial conditions change. It is now shown that a similar algorithm can be derived to perform MPC updates when model parameters change. The detailed algorithm derivation is presented, along with discussions on the performance and implementation. An example based on the nonlinear dynamics of an inverted pendulum on a cart is included to demonstrate the effectiveness of the proposed algorithm.