| Summary: | 碩士 === 國立臺灣科技大學 === 機械工程系 === 97 === The application of differential braking or direct yaw moment control (DYC) for vehicle roll-over prevention receives increasing attentions in recent years. In this research the model predictive control (MPC) technique is employed to design the DYC for rollover prevention. The predictive nature of the MPC resembles human’s vehicle driving characteristics, and is therefore a intuitive choice for the rollover prevention control design. Due to the braking actions resulted from the DYC, the vehicle longitudinal velocity is reduced, and the reduction in vehicle speed causes significant changes in the vehicle dynamics. Therefore, how to incorporate the variation in speed in the MPC design is investigated in this research, and the modified design is evaluated using computer simulations and experiments with a scaled vehicle.
The vehicle model, driver model, and the DYC are implemented as mathematical models in MATLAB/SIMULINK. The performance of the controller is evaluated by several sets of simulation scenarios, with different choices of driver models and vehicle speed settings. The simulations indicate that with a proper driver steering control model in the loop, and the internal model used in MPC updated on-line with varying speed, the DYC generates the most appropriate rollover prevention actions. The peak value of roll angle is reduced by approximately 30%, while maintaining slight improvement in lateral position tracking. The other choices may result in degradation in lateral position tracking, or even worsen the rollover motion.
The DYC design is also evaluated on a scaled vehicle setup, through a human driver controlling the steering actions. The experimental results indicate that the designed DYC can provide reasonable improvement to rollover prevention under double lane change maneuver.
|
|---|
