Improving road vehicles safety by means of innovative control systems and estimators of vehicle dynamics

In 2023, road crashes in the European Union caused 20,384 fatalities and injured over 1.14 million people, with passenger cars responsible for nearly half of all recorded deaths. Electronic stability control (ESC) systems, mandatory in many countries, represent a critical safety technology. However, most current ESC systems rely heavily on accurate sideslip angle estimation -- a notoriously difficult measurement that remains challenging and expensive in production vehicles, potentially compromising system performance under extreme conditions. This dissertation presents a comprehensive investigation into advanced vehicle dynamics and stability control systems that addresses these fundamental limitations. Through six interconnected research papers organised in three thematic blocks, this work proposes innovative solutions that reduce dependency on sideslip angle measurements while maintaining or improving safety performance. The first research block develops novel phase plane analysis methods for vehicle stability assessment without requiring sideslip angle measurements. The proposed $r-\dot{\beta}$ phase plane and sideslip-free stability region demonstrate that vehicle stability can be effectively evaluated using only easily measured quantities. This block also includes a systematic analysis of vehicle cornering motion, distinguishing three categories: stable-normal turn, unstable-normal turn, and drifting manoeuvres. Drifting is investigating in detail by comparing simulation and experimentla results. The second block focuses on advanced state and parameter estimation techniques. A combined dynamic-kinematic extended Kalman filter is proposed as an alternative to purely dynamic or kinematic approaches for sideslip angle estimation. Additionally, the Simulator-enhanced Estimation Loop algorithm offers an innovative parameter optimisation approach that continuously adjusts vehicle model parameters using only easily measured signals, improving model accuracy and state estimation. The third block presents a complete stability control system based on nonlinear Model Predictive Control that operates without direct sideslip angle regulation. This controller demonstrates comparable -- if not superior -- performance compared to sideslip-informed baselines, while maintaining real-time feasibility. The system has been validated through simulation, driver-in-the-loop experiments, and experimental testing on a full-scale vehicle prototype equipped with brake-by-wire systems. This dissertation proposes a change of perspective on the development of vehicle stability control and estimation techniques, challenging conventional approaches and opening new possibilities for robust, cost-effective safety systems that contribute to the automotive industry's evolution toward enhanced safety.