Model-based vehicle dynamics control system and states estimation for 4WD Formula SAE electric vehicle track performance assessment

Yaw Control is an interesting topic especially when applied to 4WD electric vehicles. Given the high number of degree of freedom, the torque distribution can be performed according to many ways, both with good engineering practices or with numerical optimization strategies. The need of a validated Yaw Controller for Squadra Corse PoliTo has become important since the level of the competition in Formula Student has increased in the last years. This controller is meant to help the driver both during tight turns, by make the car to rotate faster, and high speed cornering, with a stabilizing effect. To make proper advantage of the Yaw Control, vehicle sideslip angle information should be used as a safety measure. An Extended Kalman Filter and a combined estimation strategy was implemented in past works, but was lacking of reliability and some measures had been taken in order to properly validate this system. This work aims to propose a validation method, with a track performance assessment of both the Yaw Controller and the sideslip angle estimation, also presenting state-of-the-art methodologies for Yaw Controllers, estimators and model validation techniques. Moreover, a deeper detail description of SC vehicle control system is performed, giving some hints based on the last years experience of how it can be evolved. In the end some strategies for different estimation strategies blend, between dynamic, kinematic and combined are suggested based on the experience and on the possible working condition of the vehicle. The work has been divided into two main tasks: data acquisition and data processing, which comprehends the sideslip estimation enhancement. Data was acquired in a dedicated test session, comparing passive vehicle and controlled performance. The controller was loaded on a dSpace MicroAutobox2 that is usually employed by the team as VCU. Sideslip angle measurement is performed, via a Kistler SF-Motion sensor, to provide the information to the controller and retrieve data for the sideslip estimation validation. During data processing, relevant KPI analysis is performed to evaluate the behavior of the controller and estimator, while tire temperature information has been used to enhance the performance of the EKF. The results show that the controller is very effective, also in the low grip conditions in which has been tested, with a reduction during a Double Lane Change maneuver in IACA of steering , from 37.15 deg to 14.13 deg, and maximum steering, from 104.3 deg to 35.0 deg with the vehicle in passive mode that was running at an entry speed of 53 km/h, while 54 km/h is kept during TV ON tests. The final laptime reduction is in the order of 5%. The performance of the EKF has been enhanced for all the test maneuvers, and Especially during DLC and slalom maneuvers. The GOF(NRMSE) concerning dynamic estimation values lowered from 1.2 down to 0.44 for a DLC and from 0.9 down to 0.5 for a slalom maneuver.