Design and implementation of a vehicle observer for sideslip angle computation without vehicle and tire models

Sie sind hier: Homepage > Suche

Design and implementation of a vehicle observer for sideslip angle computation without vehicle and tire models

Reim, Andreas / Klier, Willy / Hillenbrand, Stefan / Otterbein, Stefan

This paper describes a design and implementation approach to estimate the longitudinal and lateral velocities as well as the sideslip angle of a passenger car based on a commercial Domain Control Unit (DCU) with integrated inertial sensors for all 6 degrees of freedom of the vehicle chassis. Since the considered approach does not include any vehicle or tire model equations, it provides a convenient robustness with respect to model validity ranges, parameters and uncertainties. Computer simulations and driving tests show a high accuracy for the estimated roll and pitch angles, lateral and longitudinal velocities and the sideslip angle even during high-dynamic maneuvers on road surfaces with low friction coefficient. The achieved results on roads with a low friction coefficient were obtained with a real-time function in a prototype vehicle. This is enabled by an improvement of the longitudinal vehicle velocity estimation on roads with a low friction coefficient and by the developed adaptive covariance matrix computation. An important criterion for the design of the Vehicle Observer was to keep the interface to other systems and electronic control units as small as possible and to base the functionality of the Kalman Filter upon a minimum set of known states or signals without affecting estimation quality and performance. Moreover, the results in a control loop show that the accuracy of the estimated signals is suitable for vehicle dynamics control systems even on roads with low friction coefficient. The authors will also focus on current prototypical applications like vehicle stabilization by means of steering interventions.