Date of this Version



connected vehicle (CV), traffic signal controller, event data, signalized intersection corridor, green probability distribution, prediction, predictability and confidence interval, human-machine interface (HMI), SAE J2735, arrival threshold


The communication between connected vehicles and traffic signal controllers is defined in SAE Surface Vehicle Standard J2735. SAE J2735 defines traffic signal status messages and a series of 16 confidence levels for traffic signal transitions. This paper discusses a statistical method for tabulating traffic signal data by phase and time of day and populating the SAE J2735 messages. Graphical representation of the red-green and green-yellow transitions are presented from six intersections along a 4-mile corridor for five different time of day timing plans. The case study provided illustrates the importance of characterizing the stochastic variation of traffic signals to understand locations, phases, and time of day when traffic indications operate with high predictability, and periods when there are large variations in traffic signal change times. Specific cases, such as low vehicle demand and occasional actuation of pedestrian phases are highlighted as situations that may reduce the predictability of traffic signal change intervals. The results from this study also opens up discussion among transportation professionals on the importance of consistent tabulation of confidence values for both beginning and end of green signal states. We believe this paper will initiate dialog on how to consistently tabulate important data elements transmitted in SAE J2735 and perhaps refine those definitions. The paper concludes by highlighting the importance of traffic engineers and connected vehicle developers to work together to develop shared visions on traffic signal change characteristics so that the in-vehicle use cases and human-machine interface (HMI) meet user expectations.


This document is the preprint version of TRB Paper No. 20-04533, submitted on August 1, 2019 for the 99th Annual Meeting of the Transportation Research Board of the National Academies held on January 12–16, 2020 in Washington, DC. The paper was presented on January 13, 2020. The paper was subsequently revised to respond to review comments in February 2020 and published on June 23, 2020. The citation for the published manuscript in the Transportation Research Record is shown below.

Mathew J., Li, H., & Bullock, D. M. (2020). Using stochastic variation of cyclic green distributions to populate SAE j2735 message confidence values along a signalized corridor. Transportation Research Record, 2674(9), 426–437.