Riders of Powered two-wheeled vehicles (P2WV) are among the most vulnerable drivers on roads. In France, riders of P2WVs are just 1.9% of road traffic, but account for 24% of road user deaths. This high rate of mortality may be explained by several factors. One of them is the lack of specific Intelligent Transportation System (ITS). In fact, because of the complexity of the P2WVs dynamics and the unavailability of some states (such as the steering torque, the roll angle and the lateral forces), it is more difficult to design specific ITS systems for P2WVs. In this context, my thesis aims the estimation of relevant dynamics of P2WVs. Firstly, a Takagi-Sugeno (TS) observer is proposed to estimate the lateral dynamics. The observer completely decouples the steering torque - which is considered as an unknown input (UI) – from the estimation error dynamics. Then, to estimate the lateral dynamics and the steering torque, another TS observer is proposed. The observer in question has a proportional derivative (PD) adaptation law to reconstruct the UI. However, in both observers, nonlinearities with respect to tire forces are neglected. To circumvent this difficulty, lateral forces are considered as UIs and they are algebraically estimated thanks to high order sliding mode observers. Finally, we focus on the longitudinal dynamics and propose an UI observer with a PD adaptation law to estimate the longitudinal forces and the thrust and braking torques. Once the lateral and longitudinal dynamics estimated, we were interested by the control of the braking system. In this context, a robust controller is proposed in order to optimize the braking for P2WVs. The controller is based on sliding mode techniques and allows the tracking of the longitudinal slip to a reference one. The latter is computed through a program inspired by MPPT algorithms. Finally, to validate the several observers developed in my thesis, an observer based controller using TS techniques is proposed in order to stabilize and automate an instrumented scooter and to track a reference roll angle.
展开▼
