UC Berkeley

It is estimated that there were 1.3 billion vehicles in the world at the end of 2016, which is almost two times more than the number of vehicles 20 years before. With the ever-growing population of vehicles, transportation has caused many problems. It was the sector with the most significant contribution to greenhouse gas emissions in the United States in 2019. About 90% of the fuel burned in transportation is based on petroleum, a non-renewable energy source. The transportation system is also unsafe. For instance, with 1.10 fatalities per 100 million vehicle miles traveled in the United States in 2019.

To resolve these problems, significant innovations have been made in vehicle technology via electrification, automation, and communication. For instance, electric vehicles (EVs) can vastly reduce greenhouse gas emissions and utilize sustainable energy, such as solar-generated electricity. Also, Connected and Automated Vehicles (CAVs) promise to improve road safety, enhance traffic network performance, and increase fuel efficiency by safely driving with smaller headways and smaller air drag.

As more vehicles adopt these advanced technologies, the urban system will find new opportunities and unintended consequences for the related infrastructure, such as the charging facilities, the electrical energy grid, and the traffic network. In this dissertation, we identify some of these issues and enable vehicle electrification and automation technologies to enhance the systematic performance of urban infrastructure through connectivity.

First, we investigate the problem of optimally planning an EV charging infrastructure, subject to the electrical grid pricing and the random charging demand. The facility planner faces a trade-off problem, where the operator has to pay a high price to achieve a high quality of service in charging EVs or pay a low cost but provide a low quality of service. We propose a Pareto-optimal planning solution that uses demand management strategies and achieves a higher quality of service at a lower cost.

Second, we study the relationship between the electrical energy grid and the charging schedule of an EV fleet on a large, regional scale. We identify the opportunities in shifting the EV charging schedule both in time and space to benefit the grid operation in terms of the cost, the renewable energy mix, and the greenhouse gas emissions. We evaluate the maximum potential gain to the grid based on real data from the individual vehicles and the energy grid operation. We recommend an optimal model to schedule the EV charging sessions.

Third, we analyze the balance between the operation of Connected Automated Vehicles (CAVs) and the traffic network performance. We find that in mixed traffic of CAVs and human-driven vehicles, naive operation strategies of CAVs can induce more lane changes and create unnecessary congestion. We propose and validate an operation strategy for CAVs that ensures the maximal performance of the traffic network while allowing the CAVs to enjoy the benefits of automation and connection.