|Ph.D Student||Shraiber Arie|
|Subject||Development and Implementation of Robust Roll-Over Avoidance|
Control Algorithms for a Four-wheeled Laboratory
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Emeritus Per-Olof Gutman|
|Professor Tal Shima|
|Full Thesis text|
Vehicle roll-over accidents are typically very dangerous. Research by the National Highway Traffic Safety Administration in the United States shows that roll-over accidents are the second most dangerous form of traffic accidents in the United States, after head-on collisions. Vehicle roll-over accidents may be grouped into two categories, called tripped and un-tripped rollovers. Tripped rollovers are caused by the vehicle coming into contact with external obstacles. Un-tripped rollovers are caused by extreme driving maneuvers, in which the forces at the tire-road contact point are sufficient to cause the vehicle to roll-over.
In order to avoid vehicle roll-over, one can adjust its design parameters which is a very complicated problem from a mechanical point of view, or control the dynamic modes of the vehicle. The dynamic modes control can be applied by two fundamental methods: or controlling the mode itself, without changing the driver induced maneuver e.g. adjusting the pitch or roll modes by active suspensions; or controlling the maneuver itself, by regulating steering and velocity. The method regulating the maneuver is much more suitable for large control actions.
In this research work a Multiple Input Multiple Output (MIMO) robust roll-over avoidance control algorithm that prevents a laboratory vehicle from rolling over due to disturbances and extreme driving maneuvers is presented. A 9 degrees-of-freedom, non-linear model of the vehicle in a roll-over situation is extended to include wheels lift-off mode and the novel roll-over criterion-NRC. The control signals are an auxiliary steering angle and an auxiliary vehicle velocity reference which are added to the steering angle and vehicle velocity reference commanded by the driver or the driving algorithm. The robust linear control law is designed by Quantitative Feedback Theory (QFT), utilizing an uncertain linear MIMO vehicle model identified from a measurement based non-linear vehicle model through Fourier Integral identification method. The designed control scheme is evaluated via simulation and tested in the Cooperative Autonomous SYstems (CASY) laboratory for different roll-over scenarios. The suggested control algorithm successfully imposes steering and velocity correction signals that prevent the vehicle from rolling over, and simultaneously keep the intended path with small deviations.