Modeling Vehicle Lateral Dynamics by Gaussian Processes

A method of modeling a vehicle’s lateral dynamics is investigated. Gaussian processes (GP) are used to estimate the side slip and yaw rate of the vehicle. A method of clustering the inputs to the GPs is also investigated and the effects of a key parameter, the clustering threshold, are examined. Sensitivity of this parameter is considered and a compromise is found between model accuracy and computation time. A discussion of how these methods can be extended to online adaptation of the estimators is included.