CAV-generated mobility data modeling mechanism for adaptive signal control

The effectiveness of adaptive traffic signal control highly relies on accurate and accountable identification of dynamic arrival turning movement demand on approaches and other traffic flow parameters measuring traffic states. Emerging connected vehicle (CV) and/or autonomous vehicle (AV)-generated mobility data can be potentially used as a new data source in support of the adaptive signal control. In the long-run, the CV/AV-generated data source could gradually substitute traditional inductive loop data as the maturity levels of the relevant data process techniques are progressively increasing. However, use of the CV/AV-generated data is still yet mature due to lack of the data process mechanism and models to integrate the data into the adaptive traffic signal control system. It is hence an imperative need to develop the mechanism for processing the CV/AV-generated data source in order to facilitate improving the efficiency of the adaptive traffic signal control schemes. This paper presents a developed methodological framework along with associated data models which can be used to configure an intelligent CV/AV data fusion in support of adaptive signal control operations. A proof-of-concept study has been conducted to test the developed models via comparison of the CV/AV-data-driven scenario and the traditional-detection-data-supported scenarios. The paper presents the modeling framework along with performance analysis of the testing study, which demonstrates positive outcomes in terms of reduced queue length and throughput, as well as benefit-cost ratios. • Enabling CAV-generated data to rebuild traffic monitoring data sources. • A framework along with CAV data fusion models to enhance adaptive signal control. • A proof-of-concept study using the realistic infrastructure in the state of Ohio. • Positive outcomes (i.e., reduced queue, throughput) from simulation-based testing.