Research On Energy-saving Driving Behavior And Speed Optimization Method In Vehicle Networking Environment

In recent years,with the rapid growth of global car ownership,automobile exhaust emissions have brought tremendous environmental impact and energy crisis.At present,China has 24,123,000 vehicles,petrochemical energy consumption accounts for 19% of the world’s total,and the average annual emission of automobile pollutants is 43,597,000 tons.Therefore,China is facing enormous challenges in energy supply and environmental protection.How to achieve energy saving and emission reduction of automobiles through policies,traffic management and emerging technologies has been a long-term concern of traffic scholars.Aiming at the problem of energy saving and emission reduction of automobiles,this paper conducts in-depth research from two aspects: fuel consumption prediction and energy saving driving behavior optimization.Firstly,the energy consumption of trucks is analyzed by using the microscopic driving data of heavy trucks obtained from the vehicle network,and the effects of different driving behaviors on fuel consumption are compared.The correlation between truck running status and fuel consumption is verified by correlation analysis.A real-time automatic partition algorithm of vehicle running state based on vehicle speed is proposed,which can automatically divide vehicle running state into three modes: acceleration,deceleration and cruise.On this basis,based on the law of conservation of energy,a hybrid model of fuel consumption prediction driven by physics and data-Energy Consumption Index(ECI)model is proposed.The model uses different calculation methods for fuel consumption under different operation modes,and then weights the calculation results to get the total fuel consumption.Compared with three classical fuel consumption prediction models,it is found that the model has better prediction ability and generality.Secondly,in order to improve the accuracy of fuel consumption prediction,this paper proposes a vehicle fuel consumption prediction algorithm based on LSTM depth network.This algorithm optimizes the traditional LSTM algorithm from the aspect of network parameter selection,transforms the time series data into supervised learning sequence data,and improves the prediction accuracy.At the same time,in order to clarify which vehicle operating parameters play a leading role in fuel consumption prediction,this paper makes sensitivity analysis on the influence of different input parameters on the accuracy of fuel consumption prediction.The experimental results show that when the input is speed and acceleration,the prediction effect is the best.Finally,aiming at the problem of increasing fuel consumption caused by "Stop and go" in the course of vehicle driving,an intelligent network driving strategy based on optimal cruise speed(OESC)is proposed to enable vehicles to pass traffic lights smoothly without stopping.In order to verify the fuel economy of OESC ecological driving strategy,a discrete speed trajectory optimization(DSTO)model is proposed in this paper.The model takes the minimum fuel consumption as the objective function of the system,and formulates relevant constraints,which proves that the objective function has an optimal solution.Through the comparative analysis of the above two methods,it is found that the fuel-saving performance of the two methods is almost the same,but OESC ecological driving strategy is far superior to DSTO model in terms of time efficiency.Therefore,OESC ecological driving strategy can not only effectively reduce fuel consumption,but also have higher engineering application value.Aiming at the problem of vehicle emission,two kinds of vehicle fuel consumption prediction models are established,which can accurately predict vehicle fuel consumption and provide basis for the optimization of ecological driving behavior.On this basis,this paper proposes an ecological driving strategy with optimal cruise speed,which effectively solves the "Stop and go" phenomenon in the process of vehicle driving and reduces fuel consumption in the process of vehicle driving.The research results of this paper have certain social and economic value.