Assessment of connected vehicle information quality for signalised traffic control

Connected vehicles (CVs) present a great opportunity to smooth and improve traffic flows at intersections thanks to their communication capabilities, which may allow a real-time flow of information with the controllers operating traffic signals. Therefore, it is reasonable to envision that, in the near future, CV data may complement or replace spot detector data that is currently used to operate traffic signals. However, CV data may be affected by errors, such as positioning error, which may depend on the technology that is employed for collecting such information. In this paper, we investigate the performances of different control strategies, namely a strategy that employs only aggregated information, such as queue lengths, and a strategy using disaggregated vehicle-based information, when they are operated with CV data, considering various realistic measurement accuracy settings. Our experiments, conducted via microscopic simulations, show that the disaggregated strategy features better performance and robustness in most of the tested scenarios.