Abstract:
According to statistics, accidents related to heavy vehicles are generally more serious than those involving light vehicles. Safety driving is the key in prevention of risk situation, one of the most deadly risk involved in driving situations is the rollover of a heavy vehicle. Detection and Prevention of this situation requires the knowledge of the roll angle and roll rate which depends on the vehicle dynamic state and other vehicle parameters. Warning systems and roll mitigation control for vehicle rollover can only be developed if state information of roll angle and roll rate is available. Onboard estimation of the attitude of a ground vehicle is motivated by its use in active anti-roll bar design. Traditionally, the ground vehicle attitude estimation problem is a complex one and its solution is computationally very intensive. Secondly, the existing solutions been developed add extra cost making it available for only limited number of users. The aim of this research is to apply algorithms to provide decent estimate for states using low cost sensors and onboard processor adding marginal cost to the vehicle. The setup can be deployed on any commercial vehicle providing true estimate of states thus making it available for large number of users.
The research carried out focuses on algorithms to estimate roll angle. The algorithms investigated include dynamic observers which rely on lateral accelerometer and gyroscope and a sensor fusion algorithm that is based on kinematic relationship helping us to avoid both vehicle and tire models and minimizes the tuning effort required and parameter variation sensitivity. The roll angle estimate compensates the gravity component which is involved in measuring lateral acceleration. The results of simulations and vehicle tests demonstrate accuracy of the proposed methods.
