| Summary: | 碩士 === 國立臺灣科技大學 === 機械工程系 === 102 === In this research, by using Nonlinear Model Predictive Control (NMPC), a driver model is proposed, which is capable of controlling CarSim® vehicle model by applying steering wheel angle, engine throttle, and brake in simulation. In order to maintain success in different simulations, the models of both vehicle and the parts inside, such as engine, tire, and bake system … etc., should be included in prediction model. For describing different considerations in driver’s thought, multiple objectives and constraints are included in NMPC. In result, by different arrangements in weightings and constraints, the NMPC driver model behaves differently in driving style. In further research, by using the ability of NMPC which can solve nonlinear optimization problems, we discussed driver’s tendency of ignoring small lateral error from center of lane and the nonlinear behavior in tire. In order to describe the tendency of ignoring small lateral error, we used a nonlinear objective function. Furthermore, during the simulation of high speed cornering, we also verified that the prediction model with different accuracy will affect the control capability of the driver model.
|
|---|
