A recursive propagator-based subspace method for vehicle handling dynamic system model identification. (February 2019)

• Record Type: Journal Article
• Authors: Yang, Weimiao; Guan, Xiqiang; Zhang, Jianwu
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 233:Number 3(2019:Mar.)
• Page Start: 495
• Page End: 516
• Publication Date: 2019-02
• Subjects: Linear time-varying model -- propagator method -- recursive subspace identification -- standard road test -- VARX model -- vehicle handling dynamic model; Mechanical engineering -- Congresses; Transportation engineering -- Congresses; 629.2
• DOI: 10.1177/0954407017746840
• Languages: English
• ISSNs: 0954-4070

Abstract

A high-precision vehicle handling dynamic model is of great importance in both the analysis of system performance and development of stability controller. This study introduces a new modeling algorithm based on subspace identification. Different from traditional batchwise subspace identification method, the new modeling method is proposed in the framework of online realization and unbiased estimation, which is ensured by a recursive propagator method (PM) combined with a vector autoregressive with exogenous (VARX) input model. Also, an observable and controllable 4-order linear time-varying (LTV) vehicle handling dynamic model, including rolling motion, is presented. To validate the identification algorithm, real vehicle standard road tests, including the step input test and impulse input test, are conducted, and sampling data are collected to apply into the identification procedure. The feasibility and accuracy of this online identification method are demonstrated. Furthermore, the stability of identification method is validated by an identification process based on road test data with measurement noise interference, and the superiority of the proposed identifiable LTV model is proved compared to the linear time-invariant (LTI) model.