Traffic demand grows rapidly over the past decades around the world, which leads to severe traffic congestion problems. Congestion has numerous negative effects, such as wasting time of drivers and passengers, increasing delays, decreasing travel time reliability, wasting fuel, and increasing air pollution and greenhouse gas (GHG) emission. In addition, when congestion occurs, the variation in speeds and headways between vehicles might lead to longer queues, longer travel time on the highways, higher accident possibilities and more frustrated drivers. In conclusion, traffic congestion is detrimental to the operational efficiency as well as travelers’ safety. In order to relieve highway congestion, the departments of transportation (DOT) have been seeking new ways to satisfy the increasing demand and make full use of the infrastructure resources. Thus, some ad hoc traffic management strategies have been developed and deployed by the DOTs so that the existing roadway resources can be fully optimized. Among different types of traffic management strategies, active traffic management (ATM) is a scheme that can be used for relieving congestion and improving traffic flow on the highways. Among these ATM strategies, variable speed limit (VSL) control has been implemented around the world (e.g., Germany, the United Kingdom, and the United States). VSL control systems are deployed to relieve freeway congestion, improve safety, and/or reduce the emission of greenhouse gases and fuel consumption under different situations. Moreover, with the development of emerging technologies, various novel methods on the basis of the intelligent transportation systems have been developed in recent years. Connected autonomous vehicle (CAV) belongs to such technology. The CAVs integrate vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) and infrastructure-to-vehicle (I2V) communication into control systems. The existing research efforts proved that enhanced performances could be achieved using CAV technologies.The research intends to systematically develop a VSL control framework in a CAV environment, in which the V2V, V2I, I2V, and platooning technologies are integrated with the VSL control. In addition, mixed traffic flows (including trucks and cars) are taken into account in the developed VSL control models. The policies (such as left-lane truck restriction policy) that are used to reduce the impacts of trucks on cars and CAV technologies (e.g., vehicle platooning) integrated with VSL control are explored. Multi-objective optimization models are formulated. In terms of the discrete speed limit values in the real world, discrete optimization techniques, such as genetic algorithm (GA) and tabu search (TS), are employed to solve the optimization control models. Different scenarios are designed to compare the control results. Sensitivity analyses are presented, and comprehensive characteristics underlying the VSL control are discussed in detail. Summary and conclusions are made, and further research directions are also given.
