Zuduo Zheng

Fundamental issues in calibrating and validating traffic flow models

An integrated platform of traffic flow simulation, communication simulation, and driving simulator

Modelling transportation systems with traditional, connected and autonomous vehicles

Connected vehicles communicate with neighboring vehicles (V2V) and infrastructure (V2I), and automated vehicles drive themselves without the need for human intervention for some of or all the driving tasks. Recent and rapid technology advancements are transporting the concept of CAVs from science fiction to scientific fact. While the valuable information collected and communicated by CAVs provides unprecedented opportunities for optimising traffic flows, the lack of a robust, theory-based operational plan for mixed traffic flow will lead to more chaotic roads. Firstly, much of our modelling of the traffic flow and traffic operations of traditional vehicles will be obsolete at best, and dangerously misleading at worst. Secondly, for connected vehicles, driver’s response and compliance to information received is critical, e.g., the total ignorance of a driver to the information would render connectivity useless; and for automated vehicles, depending on the level of automation, drivers need to frequently switch between two different roles: as a driver to execute driving tasks, and as a supervisor to monitor the driving environment, and when needed, resume vehicular control. Previous studies have reported that automation may lead to overreliance, erratic workload, skill degradation, and reduced situation awareness. And finally, the impact of CAVs on transport systems, while revolutionary, is also evolutionary. For the foreseeable future, traditional vehicles will need to co-exist with CAVs in a mixed traffic flow, which is likely to be more dynamic and volatile, posing serious operational, control, and safety challenges.

This project addresses this knowledge deficit, and develops an analytical tool with the capability of accurately modelling mixed traffic flow. This new knowledge and model are prerequisites to effective operation and control of traffic flow of traditional, connected, and automated vehicles .