SE PhD Prospectus Defense of Arian Houshmand

TITLE: ECO-ROUTING AND SCHEDULING OF CONNECTED AND AUTONOMOUS VEHICLES

ABSTRACT This prospectus studies the problem of eco-routing connected and autonomous vehicles (CAV) to minimize the overall energy consumption costs with an emphasis on Plug-In Hybrid Electric Vehicles (PHEVs). Eco-routing refers to the procedure of finding the optimal route for a vehicle to travel between two points, which utilizes the least amount of energy costs. First, the optimal eco-route for a single vehicle is found assuming known traffic composition in the network. Unlike thetraditional Charge Depleting First (CDF) approaches in the literature where the power-train control strategy is fixed, in thiswork a Combined Routing and Power-Train Control (CRPTC) algorithm is proposed which can simultaneously calculatethe optimal energy route as well as the optimal power-train control strategy. It has been shown that the CRPTC algorithmoutperforms the traditional CDF approach in terms of energy savings.

Second, a socially optimal solution is proposed to minimize the total energy consumption costs in a traffic network considering 100% CAV penetration rate. When a vehicle enters the network at an origin given its destination, the algorithm gives it the desired socially optimal route in terms of a sequence of links to follow, as well as the optimal power-train control strategy on each link. The social optimal solution is compared with the selfish routing results and the price of anarchy is calculated under different scenarios.

Finally, as an ongoing study, the eco-routing problem is being investigated in the presence of mixed traffic (CAV andnon-CAV). Moreover, to address the high dimensionality of a routing problem, a network decomposition method is underinvestigation. Throughout this prospectus, to validate the proposed methods, eco-routing algorithms have been applied tosubnetworks of the Eastern Massachusetts transportation network using actual traffic data provided by the Boston RegionMetropolitan Planning Organization. As an alternative benchmark, the traffic behavior of the network has been simulated inSUMO using the extracted flow data from the aforementioned traffic dataset.

COMMTITEE: ADVISOR/CHAIR Christos Cassandras, SE/ECE; Yannis Paschalidis, SE/ECE/BME; Pirooz Vakili, SE/ME; Sean Andersson,SE/ME

When 12:00 pm to 2:00 pm on Friday, February 1, 2019
Location 15 Saint Mary's Street, Rm 105