A gradient descent-based vehicle optimizator for enhanced state estimation

Accurate vehicle state estimation is crucial for enhancing safety in modern automotive control systems. This paper presents a gradient descent-based algorithm designed to optimise the parameters of a 5-degree-of-freedom (DOFs) vehicle model for improved state estimation and control application. The algorithm is validated through simulations and experimental tests, demonstrating its ability to adjust vehicle parameters effectively and force the model to mimic the behaviour of the reference vehicle, reducing errors in key dynamics such as vehicle velocity and yaw rate.