Abstract
In this paper, we present our recently developed stochastic driver-behavior model, which is based on a Gaussian mixture model (GMM) framework, and employed it to anticipate the car-following behavior of drivers in terms of pedal control operations in response to observed driving signals, such as the vehicle velocity and following distance to a lead vehicle. In addition, the proposed driver modeling method allows adaptation of the model to better represent any particular driving characteristic of interest (e.g., a driver’s individual driving style, a particular car-following event). Validation and comparison of the proposed driver-behavior models using realistic car-following data from more than sixty drivers showed promising results in predicting driver behavior. Furthermore, the adapted driver models showed consistent improvement over the unadapted models in both short-term and long-term predictions.
Keywords: Driver-Behavior Model, Probabilistic Model, Model Adaptation, Car Following, Realistic Driving Data