Measuring wheel rotational speed using inertial sensor modalities

Abstract

Vehicle motion state estimation is a critical component of automotive control systems, providing the foundation for effective vehicle motion management. Estimating wheel rotational states like rotational angle, speed and acceleration of the wheel and improving the roll and pitch estimation of the road is a major part of this process. In this study, wheel speed estimation is achieved using inertial sensors mounted on the wheel of a truck. The accelerometer data from the inertial measurement units (IMUs) generated in high fidelity simulation environments is processed to estimate the desired states. Rigid body kinematics and axis transformations are used to relate sensor measurements and our states of interest. The model is verified and tested in various driving and road conditions using simulation environments like IPG TruckMaker and Modelon Impact. Later an Extended Kalman Filter is used for estimation. A wheel slip controller is also developed in Simulink and embedded it into Truck- Maker, and estimated desired states using the Extended Kalman Filter. Additionally, the developed estimator is embedded into TruckMaker for live, closed-loop, simulation testing. Additionally, using a quaternion representation of roll and pitch angle of the wheel, a non-rotating IMU on the wheel hub and acceleration of truck w.r.to ground frame; roll and pitch angles of the road in different scenarios were also estimated using an Extended Kalman Filter. The estimators performed well across different scenarios and vehicle speeds, both high and low. The estimated states were in close agreement with their true values.