Energy-Efficient Adaptive Cruise Control for EVs in Urban Scenarios with Traffic Lights Negotiation

The rapid urbanization and increasing environmental concerns have driven the demand for efficient and sustainable transportation solutions. Small electric vehicles are becoming a popular choice for urban commuting due to their low emissions and cost-effectiveness. However, optimizing energy consumption remains a critical challenge for enhancing the overall efficiency and practicality of these vehicles in complex urban environments. Nowadays, the Advanced Driver Assistance Systems (ADAS) of electric vehicles are flourishing. With the gradual enhancement of the computational power of onboard chips, more complex algorithms, such as Model Predict Control (MPC), can be applied in real-time to ADAS. With the advancement of communication technology, more information such as Signal Phase and Timing (SPaT) can be obtained through Vehicle-to-Infrastructure (V2I) technology, providing the potential for further enhancing the capabilities of ADAS. The main work of this thesis is developing an advanced vehicle controller based on MPC that seamlessly integrates vehicle following and traffic light information to minimize energy consumption while optimizing driving comfort. The controller utilizes Vehicle-to-Vehicle (V2V) technology or estimation methods to obtain the lead vehicle's trajectory. By dynamically adjusting the headway distance to the preceding vehicle, the controller achieves efficient energy management within the complex and congested urban traffic conditions. A significant aspect of the proposed system is the integration of SPaT information through V2I) communication technology. This allows the vehicle to anticipate upcoming traffic signals, proactively adjust its speed, and thereby reduce unnecessary acceleration and braking. In addition, this approach also decreases overall resistance, realise enhances energy efficiency. The controller no longer switches between speed tracking and car-following tasks as two separate modes; instead, it integrates them into a single algorithm that balances both tasks in real-time. The CasADi toolbox coded in MATLAB is used for controller implementation. Firstly, design the no linear programming architecture then transform it into quadratic programming to realize real-time performance. The algorithm is compiled with MinGW64 to generate C code and implemented in Simulink with the vehicle model for Model-in-the-Loop (MiL) testing, The driving cycle used is based on a human driver driving in Turin We use Monte Carlo simulation-based method tune the controller and the simulation results demonstrated that the proposed controller performs well in different scenario, significantly reducing the energy consumption of electric vehicle, improving travel efficiency, and providing a safe and reliable driving experience. Energy consumption can be reduced by 2.9% to 10.64%, depending on different scenario and working logic. The inclusion of V2V information is shown to markedly improve performance in terms of energy savings and driving comfort, and it is particularly beneficial for multiple vehicles engaged in cooperative adaptive cruise control. This study serves as a valuable reference for the future development of intelligent transportation systems. It underscores the potential of integrating advanced communication technologies and predictive control strategies to achieve sustainable and efficient urban mobility solutions for electric vehicles.