Real-time Damper Force Estimation for Automotive Suspension: A Generalized H2/LPV Approach

Thanh-Phong Pham; Gia Quoc Bao  Tran; Olivier  Sename; Thi Thanh Van  Phan; Dung  Hoang; Quoc Dinh  Nguyen

doi:10.3311/PPtr.20076

Authors

Thanh-Phong Pham

Affiliation

Department of Automation, Faculty of Electrical and Electronic Engineering, University of Technology and Education – The University of Danang, 48 Cao Thang street, 550000 Danang, Vietnam
Gia Quoc Bao Tran

Affiliation

GIPSA-lab, CNRS, Grenoble INP, Université Grenoble Alpes, 11 Rue des Mathématiques, 38402 Saint-Martin-d'Hères, France
Olivier Sename

Affiliation

GIPSA-lab, CNRS, Grenoble INP, Université Grenoble Alpes, 11 Rue des Mathématiques, 38402 Saint-Martin-d'Hères, France
Thi Thanh Van Phan

Affiliation

Department of Automation, Faculty of Electrical and Electronic Engineering, University of Technology and Education – The University of Danang, 48 Cao Thang street, 550000 Danang, Vietnam
Dung Hoang

Affiliation

Department of Automation, Faculty of Electrical and Electronic Engineering, University of Technology and Education – The University of Danang, 48 Cao Thang street, 550000 Danang, Vietnam
Quoc Dinh Nguyen

Affiliation

Department of Automation, Faculty of Electrical Engineering, University of Science and Technology – The University of Danang, 54 Nguyen Luong Bang street, 550000 Danang, Vietnam

https://doi.org/10.3311/PPtr.20076

Abstract

The real-time knowledge of the damper force is of paramount importance in controlling and diagnosing automotive suspension systems. This study presents a generalized H₂/LPV observer for damper force estimation of a semi-active electro-rheological (ER) suspension system. First, an extended quarter-car model augmented with the nonlinear and dynamical model of the semi-active suspension system is written into the quasi-LPV formulation. Then, the damper force estimation method is developed through a generalized H₂/LPV observer whose objective is to handle the impact of unknown road disturbances and sensor noise on the estimation errors of the state variables thanks to the H₂ norm. The measured sprung and unsprung mass accelerations of the quarter-car system are used as inputs for the observer. The proposed approach is simulated with validated model of the 1/5-scaled real vehicle testbed of GIPSA-lab. Simulation results show the performance of the estimation method against unknown disturbances, emphasizing the effectiveness of the damper force estimation in real time.