An Integrated Observer for Real-Time Estimation of Vehicle Center of Gravity Height

• Published in: IEEE transactions on intelligent transportation systems 2021-09, Vol.22 (9), p.5660-5671
• Main Authors: Park, Giseo, Choi, Seibum B.
• Format: Article
• Language: English
• Subjects: Adaptation; adaptation law; Algorithms; Center of gravity; Center of gravity height; Lateral stability; linearized recursive least square; Lyapunov stability; Motion stability; Observers; Real time; real-time estimation; Real-time systems; Rollover; Sensors; Stability analysis; Stiffness; Synthesis; Tires; Vehicle safety; Wheels
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2020.2988508

This paper introduces a new integrated observer for estimating the vehicle center of gravity (CG) height in real time. It can assist the vehicle in monitoring the real-time rollover risk and improving the performance of vehicle safety control systems. The proposed integrated observer consists of three parts: a linearized recursive least square (LRLS) algorithm on vehicle longitudinal motion, an adaptation law on vehicle lateral motion, and observer synthesis. First, the LRLS algorithm performs estimation of vehicle mass and CG height during longitudinal braking and exploits the characteristic that normalized longitudinal tire stiffness is the same in the front and rear axles. Second, the adaptation law, accompanied by a roll angle observer, estimates CG height on the vehicle lateral motion based on Lyapunov stability analysis. It includes the following contributions: verification of robustness to the vehicle mass estimation error and prevention of integration drift. Finally, in the observer synthesis, the final estimation of CG height combining the above two results is derived. The overall estimation algorithm has high practicality due to the following features. 1) CG height can be obtained on both longitudinal and lateral motions of the vehicle. This point leads to a fast convergence rate in CG height estimation. 2) The fact that it does not cause any computational burden issues in real-time implementation is also a great advantage in terms of practicality. 3) It utilizes only readily available sensors. An experimental study with various driving scenarios evaluates the effectiveness of the proposed algorithm in real-car application.