Abstract

Accurate real-time estimation of vehicle mass is paramount for optimizing the performance and safety of electric vehicle (EV) control systems, including energy management, braking, and stability control. This paper presents a novel method for vehicle mass estimation by integrating a longitudinal vehicle dynamics model with a Joint Extended Kalman Filter (Joint EKF). Leveraging readily available Controller Area Network (CAN) data, the proposed approach processes real - time acceleration and driving force signals to infer the vehicle's mass. To validate the method, we collected 45 sets of driving data from an electric test vehicle under diverse longitudinal acceleration scenarios. The estimator demonstrated exceptional accuracy, achieving an error of less than 3% in 44 of the 45 test runs. Furthermore, a sensitivity analysis was conducted to assess the robustness of the estimator against variations in vehicle design parameters. The results confirm the method's practical applicability and high fidelity for real -time mass estimation using standard in-vehicle sensors.