Vehicle State Estimation Based on Unscented Kalman State Estimation

Abstract. Accurate real-time information of wheel slip angle is essential for various active stability control systems. A number of techniques have been proposed to enhance quality of GPS based estimation. This paper exhibits a novel cost-effective strategy of individual wheel slip angle estimation for a rear-wheel-drive (RWD) vehicle. At any slip condition, the slip angle can be estimated using only measurement of steering angle, front wheel rolling speeds, yaw rate, longitudinal and lateral accelerations, without requiring GPS data. On the basis of zero longitudinal slip at the both front tires, the closed-form solutions for direct computation of wheel slip angles were derived via kinematic analysis of a planar four-wheel vehicle, and then primarily verified by computational simulation with prescribed functions of radius of curvature, vehicle speed, sideslip and steering angle. Neither integration nor tire friction model is required for this estimation methodology. In terms of implementation, a 1:10th scaled RWD vehicle was modified so that the steering angle, the front wheel rolling speeds, the vehicle yaw rate and the linear accelerations could be measured. Preliminary experiment was done on extremely random sideslip maneuvers beneath the global positioning using four recording cameras. By comparing with the vision-based reference, the individual wheel slip angles could be well estimated despite extreme tire slip. Other vehicle state variables-i.e., radius of curvature, vehicle sideslip and speed-might be obtained from kinematic relations. This proposed estimation methodology could then be alternatively applied for the full range slip angle estimation in advanced chassis control applications.

show abstract

Vehicle State Estimation Based on Unscented Kalman State Estimation

Cited by 3 publications

References 6 publications

Neural Network with Cubature Kalman Filter for Earthquake Prediction in Java

Neural Network with Cubature Kalman Filter for Earthquake Prediction in Java

Development of 4WS/4WD Experimental Vehicle: platform for research and education in mechatronics

Kinematics-Based Analytical Solution for Wheel Slip Angle Estimation of a RWD Vehicle with Drift

Contact Info

Product

Resources

About