ABSTRACT
ABSTRACT: Predicting rollover is usually performed using rollover indices, where rollover is anticipated when the indices reach certain threshold values. If knowledge about the vehicle driving path is available, rollover can be detected and prevented earlier. In this work, the rollover-prediction and rollover-prevention abilities for simple vehicle models are evaluated and compared against a high-fidelity vehicle model. The analysis is performed by using the models in critical and rolloverprone maneuvers, generated with optimal control methods. The main conclusion is that a simple point-mass model would be sufficient in a velocity based rollover-prevention controller.
