Estimation of Wheel Pressure for Hydraulic Braking System by Interacting Multiple Model Filter

The precise wheel pressure estimation is a necessary input for sensor cross-checking and diagnosis, and is the basis for subsequent vehicle dynamics control. Pressure estimation algorithms based on hydraulic models have been widely used in dynamics control, while stability and convergence are not well guaranteed. To mitigate the hydraulic model error, the Kalman Filter is adopted to fuse the hydraulic model and the additional signal from the vehicle. However, the change of inputs will lead to the change of pressure circuit dynamics which makes the Kalman Filter not well suited for all pressure estimation conditions. This article presents two kinds of hydraulic models depending on the valve dynamics, and an interacting multiple model (IMM) filter is presented to integrate different hydraulic models to suit various control inputs. The filter developed in this article is verified in typical driving conditions. Experimental results show that the proposed filter presents better pressure estimation compared to Kalman Filter based on a single model.