Estimating Bus Mass Using a Hybrid Approach: Integrating Forgetting Factor Recursive Least Squares with the Extended Kalman Filter

• Published in: Sensors (Basel, Switzerland) Vol. 25; no. 6; p. 1741
• Main Authors: Du, Jingyang, Wang, Qian, Yuan, Xiaolei
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 11.03.2025; MDPI
• Subjects: Accuracy; Algorithms; Buses; Cost control; Deep learning; EKF; Energy consumption; Hybrid vehicles; Local transit; low-cost solution; mass estimation of bus; Methods; Neural networks; OEM; Roads & highways; Robust FFRLS; Sensors; Simulation methods; Traffic accidents & safety; Vehicles; EKF; mass estimation of bus; low-cost solution; Robust FFRLS
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s25061741

The vehicle mass is a crucial state variable for achieving safe and energy-efficient driving, as it directly impacts the vehicle’s power performance, braking efficiency, and handling stability. However, current methods frequently rely on particular operating conditions or supplementary sensors, which limits their ability to provide accurate, stable, and convenient vehicle mass estimation. Moreover, as a form of public transportation, buses are subject to stringent safety standards. The frequent variations in passenger numbers result in substantial fluctuations in vehicle mass, thereby complicating the accuracy of mass estimation. To address these challenges, this paper proposes a hybrid vehicle mass estimation algorithm that integrates Robust Forgetting Factor Recursive Least Squares (Robust FFRLS) and Extended Kalman Filter (EKF). By sequentially employing these two methods, the algorithm conducts dual-stage mass estimation and incorporates a proportional coordination factor to balance the outputs from FFRLS and EKF, thereby improving the accuracy of the estimated mass. Importantly, the proposed method does not necessitate the installation of new sensors, relying instead on data from existing CAN-bus and IMU sensors, thus addressing cost control concerns for mass-produced vehicles. The algorithm was validated through MATLAB(2022b)-TruckSim(2019.0) simulations under three loading conditions: empty, half-load, and full-load. The results demonstrate that the proposed algorithm maintains an error rate below 10% across all conditions, outperforming single-method approaches and meeting the stringent requirements for vehicle mass estimation in safety and stability functions. Future work will focus on conducting real-world tests under various driving conditions to further validate the robustness and applicability of the proposed method.