MODEL PREDICTIVE CONTROL AND STATE ESTIMATION: A NETWORK EXAMPLE

摘要

Model Predictive Control (MPC) is of interest because it is one of the few control design methods which preserves standard design variables and yet handles constraints. This paper focuses on a stochastic MPC problem with constraints specified in a probabilistic sense. Our aim is to study the incorporation of state estimates into the MPC problem. The original problem can be approximated by a deterministic constrained MPC problem for the conditional mean by absorbing the state estimates' covariances into the constraints. Our idea is explored in a standard discrete time Linear Quadratic Gaussian problem, and is demonstrated with a simple application in network congestion control.