Application of machine learning algorithms in lane-changing model for intelligent vehicles exiting to off-ramp

The primary objective of this study is to evaluate how intelligent vehicles equipped with cooperative adaptive cruise control (CACC) improve freeway efficiency and safety at an off-ramp bottleneck. Applying randomized forest and back-propagation neural network (BPNN) algorithms, lane-changing characteristics are obtained based on ground-truth vehicle trajectory data extracted from the NGSIM dataset. The results show that both CACC penetration rate and length of diverge influence areas exert considerable influence on road capacity and traffic safety. Overall, the capacity will peak after an initial decrease as the CACC penetration rate increases. The maximum capacity obtained in 100% of CACC vehicle scenarios improved by over 60%, compared with 50% CACC penetration rate scenario. The proposed integration system with 100% CACC penetration rate significantly reduced the rear-end collision risks, decreasing time exposed time-to-collision and time integrated time-to-collision by 70.8%–97.5%.