MPC para Rastreamento de Referência com Evitamento de Obstáculos Usando Funções de Barreira de Controle

Abstract

In this work, a finite horizon optimal control approach is presented to address the tracking problem while adding avoidance characteristics to the closed-loop system. Inspired by the Model Predictive Control (MPC) framework for tracking, we explore the idea of introducing artificial variables into the control problem. This allows us to integrate avoidance features into MPC without compromising the properties of feasibility in the face of changing references. Additionally, we consider these artificial variables together with control barrier functions to formulate avoidance constraints. Finally, a numerical example is presented to demonstrate the controller's behavior, ensuring avoidance of multiple obstacles within the admissible space while tracking the reference.