Real-Time Non-Linear Vehicle Dynamics Prediction Model.

Sport Utility Vehicles (SUV’s) have earned a reputation for being unsafe due to high accident numbers. These vehicles are designed to be used on both smooth roads and rough off-road terrains. These vastly different operating conditions require vehicle and suspension parameters that lie at opposite ends of the design space. Harder suspension is required for adequate handling on smooth roads and softer suspension, combined with large ground clearance, allows for improved ride comfort and off-road capability. This causes a compromise in the suspension setup. As a result of the typically softer suspension, as well higher centre of gravity (CG), compared to passenger vehicles, SUV’s are more prone to rollover.
This paper proposes a real-time non-linear preview model that can predict vehicle parameters up to 300 milliseconds in future, based on the current vehicle state. Predicted parameters include yaw rate, side-slip angle, roll angle, roll rate and lateral acceleration. Using the predicted vehicle parameters, the severity of a future manoeuvre, such as a potential vehicle rollover, can be analysed and the necessary precautionary measures (semi-active suspension settings, active anti-roll bar settings, changing CG height etc.) taken in advance to improve the safety and comfort of the occupants in the vehicle. Simulations are validated against experimental results for a Land Rover Defender 110 vehicle.
The accuracy of the preview information deteriorates as the preview time become larger but good results can be achieved with preview times of up to 300 milliseconds.