Nonlinear Model Predictive Control for Enhanced Trajectory Tracking with Multi-Obstacle Avoidance

In this study, trajectory planning was conducted employing the Model Predictive Control (MPC) approach. This method predicts the future state variables of the system by utilizing a simplified vehicle model known as the bicycle model. These predicted state variables are used in generating future control signals. Initially, the traditional equations of vehicle dynamics were elucidated, and these equations were rederived. Subsequently, multiple static obstacles were introduced to ensure a collision-free path in the application of the MPC method. The simulation of this method is executed using MATLAB with the assistance of CasADi, referencing a trajectory tracking scenario. The optimization of objective criteria in the model was achieved through the Interior Point Optimization method. Upon incorporating velocity as an input using a nonlinear model, encountered results were explained, and challenges were addressed by employing acceleration as an input. Consequently, by reducing the input signal to acceleration, constraints on all states were implemented, enabling successful trajectory tracking. The outcomes of the proposed method in the simulation environment are presented.