Artificial Neural Network Based Driver Modeling for Vehicle Systems

Modeling of driver plays an important role in predicting vehicle performance accurately by a forward looking vehicle system models. It is quite difficult to capture driver behavior accurately as different driver behaves differently based on his/her judgment & reflex action. In this paper an Artificial Neural Network (ANN) based driver model is developed & compared with a traditional PID based driver model. The ANN driver model is developed based on a real accelerator pedal by a driver to follow standard drive cycle for a medium duty truck on a chassis dynamometer. The proposed ANN driver model is simulated with a validated vehicle model and comparison shows that the ANN driver model predicts vehicle performance better than PID based driver model. This method of developing driver model would be useful to improve performance prediction & control algorithm development using a vehicle system model.