UC Irvine

With connected vehicle technology, eco-driving strategies could bring benefits in decreasing traffic oscillation, reducing fuel consumption along with air pollution, and improving traffic mobility. This thesis proposes an eco-driving strategy to reduce traffic oscillation and smooth trajectories for vehicles at different traffic situations, including highway sections and urban traffic intersections. The eco-driving strategy assumes a connected vehicle environment, in which each vehicle exchanges information in real-time through vehicle-to-infrastructure (V2I) communications and vehicle-to-vehicle (V2V) communications. The proposed strategy devises the advisory speed and the control duration for each vehicle through various control algorithms that are available in different traffic situations. After receiving the information from V2I and V2V communications, each connected vehicle would apply a simple cruise control method during the control duration. The proposed algorithm determines proper advisory speed and control duration through a heuristic solution of the optimization model, which is a tangent line approach. The optimization model, which has the objective function to minimize the speed oscillation of each vehicle is applied at the origin or estimated trajectories under different traffic situations, including when vehicles are approaching a signalized intersection, non-signalized intersection, and following a moving bottleneck at an arterial road. Based on the optimization model, this study proposes a heuristic solution which is a tangent line method between control starting point to the estimated objective trajectories, with the slope of the tangent line serving as the static advisory speed applied for the vehicle. The kinematic wave analysis is applied for deriving or estimating vehicle objective trajectories at each traffic situation. In a highway with a slow-moving vehicle, this study solves a more generalized moving bottleneck problem to derive the traffic flow-rate and density upstream of the moving bottleneck, and get the following trajectory for each vehicle after the moving bottleneck. At traffic intersections, the kinematic wave analysis is first applied to describe vehicle start-up behaviors, and demonstrates the requirements for fully improving traffic efficiency. According to the requirements, the estimated time points (i.e., the passing points) are created for vehicles to pass through the intersection. Accordingly, the origin trajectories are estimated in order to make vehicles enter the intersection at each passing point with the designed optimal speed. For a signalized intersection, the passing points are created based on a specific signal plan of a given phasing. In addition, the eco-driving strategy could also be extended into a Cooperative Vehicle Intersection Control (CVIC) algorithm and applied at a non-signalized intersection. In this situation, the passing points are created and updated according to the previous vehicles passing through the intersection from different directions.

The benefits of this strategy are presented from a set of numerical simulations, along with the sensitivity analysis with the flow-rate taken at the intersection situations. The simulations are conducted for each traffic situation. In the highway sections, four scenarios with different bottleneck movements are applied, where the moving bottleneck has a constant speed, accelerates, decelerates, and stops-and-goes. The results show that both speed variance and fuel consumption are reduced with the algorithm in each of the scenarios. At the signalized intersections, all vehicles applying the algorithm would never stop before the intersection and would pass through the intersection with the free-flow speed. The number of vehicles that could pass through the intersection also increases. Also, traffic oscillation is reduced when approaching the intersection. With the CVIC algorithm applying at the intersection, all vehicles could pass through the intersection with free-flow speed, and no traffic conflictions happens between vehicles. Furthermore, in an illustrative study, the algorithm reduces the total intersection delay by 34.95 % on average, the fuel consumption by 65.79 % on average, and the speed variance by 5.75 % on average. The simulation result shows the algorithm could be beneficial in decreasing traffic oscillation, reducing environmental impact, and improving traffic efficiency as well when applied at the intersections. The basic idea behind this study is to make a predictable time and speed for the vehicle to eliminate the moving bottleneck effects and to pass through the intersection. This information provides a theoretical basis for further optimizing the traffic flows through the V2I, V2V communications, and CAV technologies, which could be applied to various situations.