Linear parameter-varying observer design for vehicle yaw rate sensor bias estimation and signal reconstruction

The yaw rate signal is of great significance for vehicle stability and motion control systems. However, contamination of sensor signal may sometimes be involved for various reasons. Although the sensor is well tuned before using, loss of ground may introduce a bias to the sensor signal during vehicle operation. To reconstruct the yaw rate signal and estimate the yaw rate sensor bias, a linear parameter-varying (LPV) approach for an electric vehicle is proposed in this paper, which requires signals of vehicle longitudinal velocity, contaminated yaw rate, and steering angle as the inputs. The H∞ performance from the modeling error to the estimation errors is taken as the design objective in the development of this LPV observer. The time-varying parameter, which is the vehicle longitudinal velocity here, is tackled by the LPV approach. Vehicle experimental investigations are conducted to show the effectiveness of the observer in estimating the sensor bias and reconstructing the yaw rate signal.