Probabilistic Driving Models and Lane Change Prediction

Probabilistic traffic models, providing a statistical representation of the future behavior of traffic participants, are crucial for risk estimation in automotive collision avoidance systems. Current research has focused on large-scale behavior, primarily in the form of lane change prediction. These models are limited in their use for high-fidelity driving propagation. This paper investigates a methodology for dynamic model construction based on a Bayesian statistical framework successfully employed in aviation collision avoidance systems. Machine learning techniques are also used to develop a lane change predictor.