Robust Vehicle Roll Dynamics Identification based on Roll Rate Measurements

The modern vehicle development process is required to provide short product update cycles and high cost efficiency. Nonetheless, the demands on vehicle safety and handling quality remain high as ever. Driving characteristics evaluation relies on extensive subjective assessment as well as objective methods. Model-based objective assessment saves time and costs as most of the evaluation work can be done offline, i.e. during or after a test run the parameters of a vehicle dynamics model are identified and used for later simulation work. Referring to vehicle dynamics assessment the roll movement of the vehicle is of great importance (aside from longitudinal and lateral vehicle dynamics).This paper presents two methods for the online parameter identification of a simple roll dynamics model. These parameters can then be used for computation of the roll angle. The first method employs a recursive least-squares algorithm whereas the second method consists of a sliding-mode observer combined with an uncertainty reconstruction technique. Compared to other existing methods for vehicle roll dynamics identification both estimation techniques only require the cost-efficiently measurable roll rate and lateral acceleration signals.