Navigating Mobile Robots in 2D Environments using Model Predictive Control

This paper describes a model of predictive designing and implementing a control system for a mobile robot in MATLAB. The controller is designed to move the robot around in a 2D environment with various test scenarios. To set up the MPC problem, both the translation and rotation motions of this mobile robot are modelled. Such points include the definition of states and inputs, prediction horizon, formulation of cost function et cetera. In this instance, an algorithm was developed using MATLAB which optimizes control inputs based on system dynamics and given objectives. The performance of the controller will be tested through various tests like trajectory tracking and goal-reaching tasks among others. Different situations will be considered during simulations showing how well the MPC works through using as a navigation tool for different environments with varying challenges. This study contributes towards improving self-driving methods used in mobile robots employed in industries ranging from transport to warehousing or even security services, especially during search operations, where human life may be at stake.