Emergency braking control for commercial electric vehicles with electric-pneumatic composite system based on federated adhesion estimation

The performance of composite braking in emergencies is critical to the safety of commercial electric vehicles (CEVs). This study tackles challenges in enhancing braking performance, including road adhesion variations, brake response differences, and real-time computational demands. Firstly, an event-triggered federated extended Kalman filter (ET-FEKF) is proposed to estimate relative adhesion and essential vehicle state parameters. The event-triggered mechanism primarily selects more appropriate estimation models based on vehicle states, rather than reducing computation frequency. Meanwhile, the federated structure integrates estimates from each sensor. Secondly, a nonlinear model incorporating the hysteresis effect of the electric-pneumatic composite braking system is developed. Based on this model, a torque control method using the corrective regulation periodic predictive controller (CRPPC) is introduced to improve emergency braking performance. This approach reduces computational load and enhances energy regenerative efficiency, thereby mitigating brake wear. Finally, the proposed method is validated through hardware-in-the-loop (HIL) and vehicle-in-the-loop (VIL) testing across various scenarios. Results demonstrate that, compared to baseline methods, the proposed strategy decreases computational burden, increases mean fully developed deceleration (MFDD) by 6.36%, enhances torque reduction control actions by 53.6%, and improves energy regenerative efficiency by 77.4%.