Sensor Fusion Based on a Dual Kalman Filter for Estimation of Road Irregularities and Vehicle Mass Under Static and Dynamic Conditions

• Published in: IEEE/ASME transactions on mechatronics Vol. 24; no. 3; pp. 1075 - 1086
• Main Authors: Boada, Beatriz L., Boada, Maria Jesus L., Zhang, Hui
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.06.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accelerometers; Algorithms; Computer simulation; Driving conditions; Dual Kalman filter; Heavy vehicles; Irregularities; Kalman filters; Mathematical model; Mathematical models; multisensor systems; Observers; Parameters; road profile estimation; Roads; Sensors; Suspensions (mechanical systems); Trip estimation; Vehicle dynamics; vehicle mass estimation; Vehicles
• ISSN: 1083-4435; 1941-014X
• DOI: 10.1109/TMECH.2019.2909977

Mass is an important parameter in vehicle dynamics because it affects not only safety but also comfort. The mass influences the three movements corresponding to vehicle dynamics: longitudinal, lateral, and vertical. Therefore, having an accurate value of mass is essential for having a suitable model which will lead to proper controller and observer operation. Additionally, unlike other vehicle parameters, the mass can vary during a trip due to loading and unloading items and passengers onto the vehicle, greatly influencing its dynamics. This is critical in heavy vehicles where the mass can vary by around 400%. Therefore, the mass must be updated online. The novelty of this paper is the construction of a state-parameter observer which allows the mass under static and dynamic driving conditions to be estimated using measurements from sensors that can be mounted easily on vehicles. In this study, a vertical complete model is considered based on the dual Kalman filter for mass and road irregularities estimation using the data obtained from suspension deflection sensors and a vertical accelerometer. Both simulation and experimental results are carried out to prove the effectiveness of the proposed algorithm.