Predictive Steering Angle Generation Algorithm for High Efficiency Vehicle’s Path Optimization

Predictive performance simulation of a high-efficiency lightweight vehicle is performed through development of a multi-physics MATLAB Simulink model including advanced vehicle dynamics. The vehicle is put into a three-dimensional representation of the racetrack, including its dimensions, slope, banking, and adhesion coefficient along the model space, elaborated from the track GPS data points. The vehicle's reference trajectory is not priorly provided to the model at the simulation start as, during run-time, a predictive Steering Angle Generation (SAG) algorithm based on Nonlinear Model Predictive Control (NMPC) computes the optimal steering angle input needed to drive the vehicle on the track within its limits. Computation is based on fast predictive simulations of a simplified version of dynamics modelling of the vehicle. Each single simulation exploits a different possible steering angle to be applied by the virtual driver, starting from the initial conditions given by the actual simulated state of the system. The results of the various steering angle simulations are collected and used by a cost function minimization algorithm. The performance target of the path optimization is described by the tunable parameters inserted in the algorithm's cost function, allowing to prioritize speed or fuel consumption. The model is being tested and validated, with good accuracy (the error on the lap time is below 0.1%), on the vehicle track data obtained during 2023 and 2024 racing events and can be used as a basis for developing an automated race strategy algorithm for vehicle performance enhancement.