Ecological Control and Coordination of Connected and Automated PHEVs at Roundabouts under Uncertainty

Abstract

During the last decade, comprehensive research efforts were concentrated on autonomous driving. Annually, many car accidents happen as a result of human faults. Extreme traffic congestion prolongs commute time, increase air pollution and cause other transportation inefficiencies. Consequently, using advanced technologies to make vehicles less dependent on human drivers enable more efficient use of time for passengers and decrease car accidents. Connectivity between vehicles and automation provides a spectacular opportunity to improve traffic flow, safety, and efficiency. There are different main active research subjects under the broad domain of autonomous driving, one of them is intersection control for connected and automated vehicles (CAVs) which can be categorized into centralized and decentralized approaches. The environmental and strict regulatory demands require automotive companies to reduce Carbon Dioxide emissions by investing more in Electric Vehicles (EVs) and Plug-in Hybrid Electric vehicles (PHEVs). A PHEV equipped with connectivity and automation looks more interesting to automobile consumers since they can have advantages of both fewer emissions and enhanced abilities. Since the powertrain of PHEVs consists of different sources of power, advanced control techniques such as Model Predictive Control (MPC) is needed. Coordination of vehicles at roundabouts is a demanding problem especially by knowing that the chance of both lateral and longitudinal collision exists. To this end, first, we proposed a centralized nonlinear MPC-based controller to adhere to calculated priorities for connected and automated PHEVs (CA-PHEVs). We further continued this research by proposing an approach for solving nonlinear multi-objective optimal control problem of decentralized coordination of CA-PHEVs at roundabouts with consideration of fuel economy. It was found that the proposed controller can calculate priority based on a navigation function and provide a safe gap between vehicles. A novel priority calculation logic based on optimal control is proposed as well and its performance is compared with the navigation function approach. In addition to the decentralized control approach, we considered a more realistic robust tube-based nonlinear MPC decentralized approach to solve this problem in the presence of uncertainties. We used simulations to test the controller and a Toyota Prius PHEV high-fidelity model is used in this thesis for simulations. Simulation results show that the addition of robustness, and energy economy to performance index can improve the fuel consumption of the vehicle. One of the major concerns in designing a controller for automotive applications is real-time implementation. The results of hardware-in-the-loop experiments show the real-time implementation of the controllers.