Real-time Non-linear Vehicle Preview Model

Abstract

Sport Utility 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, compared to passenger vehicles, SUVs are more prone to rollover. This motivates researchers to find methods of improving the handling of Sport Utility Vehicles, which in turn would decrease the number of rollover accidents involving these vehicles. The proposed methods include, amongst others, the use of active anti-roll bars, slow-active, semi-active and active suspension. The control strategies of most of these methods are based on the current vehicle state, giving them the same downfall, which is a delay in switching. To eliminate this delay, some type of preview is required. A non-linear vehicle preview model that solves in real-time was developed and implemented on the Land Rover Defender 110. The vehicle preview model is capable of predicting vehicle states up to (limited by the current processor) with good accuracy. The predicted states can then be used as an input to a control system or the model can be used as a state estimator. Even though there are numerous possible applications of the vehicle preview model, it was only implemented in one existing suspension control system, known as the Running Root Mean Square strategy. This strategy compares the measured lateral and vertical accelerations of the vehicle to decide on the suitable suspension setting. This strategy has a delay of about ms. When the predicted lateral acceleration was used as an input to the existing suspension control strategy, the delay in switching was reduced and improvements in vehicle handling of up to was achieved over a variety of tests.