Connected Vehicle Data for Freeway Congestion Analysis: Necessity versus Luxury for Transportation Engineers

This study evaluates the utility of disaggregated connected vehicle probe (DCVP) data in augmenting existing incident detection workflows in freeway congestion analysis by comparing it with Bluetooth and aggregated probe (AP) data. Using a year-long dataset from a 60 km segment of I-20/59 in Alabama, more than 80 verified congestion events were analyzed to assess differences in congestion onset detection, queue-length estimation, and shockwave behavior. Shockwave analysis revealed that DCVP data captured queue dissipation more dynamically, indicating greater sensitivity during recovery phases. While AP data produced smoother estimates of shockwave speeds, they often failed to capture short-duration or spatially limited congestion events, especially those that resolved quickly or did not propagate far upstream. Bluetooth data, although more stable, tended to smooth out severe speed drops during congestion, offering a less sensitive view of traffic variability. In validating regional traffic management center (RTMC) incidents logs, DCVP data detected congestion earlier in 77% of events, with an average lead time of approximately 5 min. They also showed better spatial alignment with reported incident locations, with a mean deviation of 0.9 km compared to 1.3 km for AP data. These findings demonstrate that DCVP data can serve as a valuable resource for enabling RTMC operators to detect disruptions more quickly and accurately, reducing reliance on external alerts. Their ability to deliver high-frequency, real-time insights also positions them as a potential enabler for future automation in incident detection and response.