Adaptive Leading Cruise Control in Mixed Traffic Considering Human Behavioral Diversity

This paper proposes an adaptive leading cruise control (LCC) strategy for connected and automated vehicles (CAVs) operating in mixed traffic flows. The method explicitly models the behavioral diversity of surrounding human drivers — captured as a joint distribution calibrated from field-collected naturalistic driving data — and integrates it into a reinforcement learning-based control framework. Simulation and hardware-in-the-loop experiments show a 6.41% reduction in energy consumption compared with baseline adaptive cruise control while maintaining safety under heterogeneous human behaviors.