DEVELOPMENT AND VALIDATION OF TRACKED VEHICLE DYNAMIC MODEL FOR MODEL-BASED PATH TRACKING CONTROL

Abstract

One of the most common intelligent path tracking controllers is model-based predictive controller which can only be functional with a good dynamic model that can effectively predict the vehicle behavior within a set prediction horizon. Therefore, this study presents the modelling and validation process to develop a dynamic model of a small-scale tracked vehicle using actual data from a mounted MPU6050 accelerometer. The model will be validated in several situations, namely forward motion and left and right cornering by comparing the predicted vehicle trajectories to an actual one. The validation process employs Root-Mean- Square Error (RMSE) calculation between the mathematical model and experimental data. This rigorous comparison method establishes a reliable foundation for the development of a model-based controller designed for tracked vehicle path tracking control. The utilization of RMSE serves as a quantitative measure, ensuring the accuracy and fidelity of the model to the real-world experimental outcomes. The successful implementation of this model-based controller holds the potential to significantly enhance tracked vehicle performance, safety, and autonomy, making substantial contributions to the advancement of these critical aspects in the field.