A new control system design based on driving characteristics for hybrid electric vehicles

A correctly and effectively designed energy management system is crucial in reducing fuel consumption and harmful emissions in dual-power hybrid electric vehicles, which are an intermediate stage between internal combustion engine vehicles and fully electric vehicles. There are many studies on various control systems that will provide the most appropriate energy management in hybrid electric vehicles. One of the main focuses on control systems lately is the optimization of these systems. One of the innovative ideas put forward to ensure optimal energy use is to optimize the control system for different road, vehicle, and driver types. In light of such an innovative idea, a data-driven method has been developed in this thesis to create a stochastic energy management system by considering realistic uncertainties. It has been shown that, depending on the road type and driving style, a hybrid electric vehicle can be used more efficiently with an appropriate energy management strategy. 38 thousand km of real driving data was collected and analyzed with nine different drivers. The necessary features were determined from these data and their values were calculated, and these values were passed through a comprehensive data preprocessing stage. Then, using data mining methods, the driving style of each drive was determined. With the developed classification algorithm, the driving style for three different roads was estimated with an average success rate of 97%. Much better fuel and emission values have been achieved with the fuzzy logic-based energy management system designed according to the driving style determined by the classification algorithm. According to the fuzzy controller embedded in the fuzzy controller simulation environment, developed with a data-driven approach, fuel improvements of up to 7% on the highway, 9% on the urban road, and up to 16% in the residential area, based on real driving data results. Although there is a trade-off between fuel and pollutant emissions, the proposed fuzzy logic-based control system has also provided significant improvements in harmful emissions. The results obtained in this study can be used as an inspiration and guide in the studies of improving fuel and emissions in (hybrid) electric vehicles.