Connectivity between vehicles and infrastructure allows the efficient flow of information in a dynamic traffic environment. This information can be used to provide recommendations to vehicles in order to alleviate traffic congestion, improve mobility with considerable benefits to the environment. The traffic flow environment however is very complex and involves many uncertainties that include inaccurate measurements, missing data, etc. Any approach to manage or control traffic should be able to handle such uncertainties in a robust way. This project focusses on variable speed limit (VSL) control as an approach to reduce congestion at bottlenecks despite the presence of uncertainties. Numerous research efforts have been made over the years in the field of VSL control in order to resolve bottleneck congestion and improve traffic mobility. Nevertheless, few of them have looked into the issue of robustness with respect to measurement or model uncertainties. In this project, a robust VSL controller is designed based on a modified multi-section cell transmission model (CTM) to alleviate freeway traffic congestion and reject uncertainties. The proposed VSL controller computes the speed limit recommendations using measured flows and densities and communicates them to the upstream vehicles. The optimum location where the speed limit recommendation should be communicated to vehicles is another control variable addressed in the project in order to maximize performance and benefits to the environment. The proposed VSL controller is integrated with ramp metering (RM) controllers and lane change (LC) recommendations to maximize performance. The effectiveness of the integrated control scheme is demonstrated using extensive Monte Carlo microscopic simulations under several traffic demand scenarios and different types and levels of uncertainties. The microscopic simulations are carried out using the commercial traffic software VISSIM. Real data are used to validate the traffic simulator. The benefits in terms of mobility, safety and emissions are quantified.
