A Dynamic Driver Trust Model for Freeway On-Ramp Merging Considering Driving Style

In freeway on-ramp merging scenarios, hu-man-machine cooperative driving systems face a significant risk of misoperation due to imbalances in driver trust. This paper proposes a personalized dynamic trust modeling framework based on Kalman filtering for on-ramp merging. Simulation results from 72 participants reveal substantial differences in trust levels among aggressive, normal, and conservative drivers. By incorporating driving style parameters, the proposed model accurately captures the dynamic evolution of driver trust, maintaining a root-mean-square estimation error(RMSE) between 2% and 5%, which demonstrates robust predictive performance and adaptability.