A new driver behavior model to create realistic urban traffic environment

This paper presents a new driver behavior model, which emulates various driving styles (i.e. behaviors) for different categories of drivers. The model was highly parametric and was developed based on a two-layer Hierarchical Concurrent State Machines (HCSM) programming framework. Our study has been specifically oriented to create a realistic urban traffic environment with hazardous situations typical of real life in a driving simulator, and let the novice drivers to practice in a safe environment. In our study, we used TRAFIKENT driving simulator as a test-bed. Experiments and evaluations demonstrated satisfactory results in terms of behavioral validity of our model.

► Using driving simulators for training is becoming increasingly popular.
► Parametric driver behavior model is developed for autonomous vehicle agents.
► It emulates various driving styles (i.e. behaviors) for different categories of drivers.
► It is based on two-layer Hierarchical Concurrent State Machines programming framework.
► Experiments have demonstrated satisfactory results in terms of behavioral validity.