Master of Science

thesisIt can be expected that connected and automated vehicles (CAVs) and human-driven vehicles (HVs) will coexist on the transportation network in a long period. Hence, to support various traffic control tasks, it is critical to develop a reliable model to understand the real-time traffic pattern in mixed CAV andHVenvironment. Taking CAV speed harmonization and control as an application, this research develops an optimization framework that can yield the profile of optimal desired speed to CAVs. In summary, the thesis includes three main research components: 1) basic optimization model which aims to determine the desired speeds of CAVs in each freeway segment; 2) improved traffic flow model that employs Kalman filter for real-time estimation corrections; and 3) extended optimization model that integrates the improved traffic flow model. More specifically, the basic model has an objective function of minimizing total travel time of both HVs and CAVs along the freeways. The optimization model is based on a novel macroscopic traffic flow model that treats CAVs and HVs as separate groups. In addition, a new set of impact factors are introduced to represent the speed change of HVs due to following CAVs in the traffic stream. Recognizing the estimation accuracy of the macroscopic traffic model can significantly affect the effectiveness of the speed optimization framework, this research further introduces an improved model using the extended Kalman filter for estimation corrections. Then grounded on the improved traffic flow model, an extended optimization model is formulated. Through numerical experimental tests, the study indicates that the proposed CAVs speed control model can significantly reduce the total travel time on freeways