We utilized vehicle trajectory data obtained from video observation prior to and after the occurrence of traffic congestion in the passing lane of the outbound Yamato sag section of the Tomei Expressway. We then analyzed vehicle behavior and interaction within the congested and preceding uncongested vehicle platoons. In a car-following model that incorporates a relative acceleration term, the reaction delay time is estimated to be longer yet still within a reasonable range, and the reproducibility is higher compared to a model that substitutes the same term with the relative velocity term. We calibrated a parameter set for this model, aiming for the smallest deceleration wave amplification factor of less than 1.0, and assumed it to represent an Adaptive Cruise Control (ACC) vehicle. Subsequently, we conducted 10,000 simulations, replacing the observed vehicles with the model. The results showed that when the vehicle mixing rate exceeded 30%, the traffic congestion avoidance rates exceeded 50%. The increase in the vehicle mixing rate may have equalized the distribution of vehicle density in the vehicle platoons at the sag section, thereby potentially suppressing the peak value of the traffic flow rate downstream of the sag section and improving the stability of the traffic flow.