Nowadays the control units used vehicle industry reached a level of complexity, that manual examination are not effective enough for testing and validating their performance, hence the application of automated simulation system is inevitable. In order to make the simulation more suitable with the real vehicle system, it is indispensable to create a model of vehicle driver. The inference of the behaviour of driver is based on the data –for instance, pedal positions, state of gearbox, and velocity of vehicle – which was collected by the measurement units integrated in the vehicles control system.
In the first section of the thesis, the historical background of modelling driver, the most remarkable characteristics of human driver is introduced, and then We can learn the basic idea of fuzzy inference system, the generation of its rule base, and how to tune its accuracy using ANFIS.
The last section of the thesis focuses on the new vehicle driver model, principally its generation, and its integration to an existing model. Then the old and new models are compared based on simulation results, using both measurement data and the model of a vehicle.