Intelligent Vehicle Energy Saving Speed Optimization Based On Driving Style Recognition And Driving Cycle Prediction

With the gradual deepening of research on diversified vehicle sensing information and intelligent vehicle control systems,environmental information,vehicle information,and driving operation information have been further integrated into the problem of vehicle energy conservation.Vehicle energy-saving technology has evolved from simply reducing engine fuel consumption to developing new energy sources,and from single-vehicle driving energy-saving to multi-vehicle coordination and human-vehicle-road coordination.In the study of human-vehicle-road collaborative energy-saving,it is crucial to incorporate driving style and driving condition information to deepen the understanding of vehicle energy consumption mechanisms and develop more effective energy-saving measures.This paper aims to investigate the impact of driving style and driving conditions on the energy-saving performance of vehicles in the context of "human-vehicle-road collaborative energy-saving".The main challenge lies in accurately quantifying and classifying these factors.The primary research objectives of this study are as follows:Firstly,a virtual driving scene software was utilized to construct a model that replicated the actual road conditions of Changchun City and the technical standards of highway engineering.The driving scenario encompassed city,highway,and suburban areas,spanning a total of 7.6 kilometers.Simulation driving data was collected using an in-loop driving test platform.Subsequently,driving data of 28 drivers with varying genders,driving ages,and ages were collected,preprocessed,and subjected to feature extraction,training,and inference.A sample data of characteristic parameters representing driving conditions and driving styles was constructed and utilized for training and optimizing the recognition model.Secondly,samples that best represent driving conditions were selected,and a driving condition identifier was established using a particle swarm optimization BP neural network.Next,the principal component analysis method was utilized to optimize the characteristic parameters of driving style.A driving style classification algorithm based on principal component analysis and self-organizing mapping neural network was proposed,and the classification model was trained under various working conditions.Driving styles were categorized into three groups: radical,ordinary,and mild,and the feature data was marked with a driving style evaluation table.Building upon this,the driving style recognition model was established by improving the key coefficients of particle swarm optimization support vector machine.Finally,a weight coefficient rule was established to quantify the influence of working conditions and driving style on speed optimization in the problem description.This weight coefficient primarily affects the size of driving and braking torque through the objective function.An ecological driving speed optimization controller was designed,integrating information on driving style and driving conditions into the optimization problem.The GPM was employed to solve the problem quickly while ensuring economy and safety constraints.Experimental results demonstrate that the proposed control strategy not only ensures driving safety but also reduces excess energy consumption caused by habitual driver behavior,thereby improving the economy and comfort performance of electric vehicles.