Model-based Estimation of Vehicle Roll State for Detection of Impending Vehicle Rollover

Abstract

This paper describes a new methodology for designing model-based estimators for detection of impending vehicle rollover. Vehicle roll motions are induced by maneuvering and road disturbances. An estimator is designed based on a three-degrees-of freedom vehicle maneuvering model and a four-degrees-of freedom half-car suspension model to obtain good estimates of the vehicle roll angle and roll rate in driving situations in which both maneuvering and road disturbances affect the vehicle roll motions. The estimator uses already existing sensors, such as steering wheel angle sensor, lateral acceleration sensor, and yaw rate sensor, on a vehicle equipped with an electronic stability control (ESC) system. Since road disturbance is unknown or very expensive to measure, disturbance-decoupled-observer design technique is used in the design of the estimator. The performance of the estimator is evaluated through computer simulations using a validated vehicle simulator. It is shown that, only with already available sensor measurements, good estimates of the roll angle and roll rate can be obtained using the proposed estimator in driving situations in which both maneuvering and road disturbances affect the vehicle roll motions. A rollover index that indicates rollover danger has been computed using the measured lateral acceleration and yaw rate, and estimated roll angle and roll rate. The rollover index computed using the estimated states is shown to be a good measure for the danger of vehicle rollover.