Study On Intelligent Driving Strategy Of Electric Vehicle Considering Multi-Influences In Urban Road Network

In March 2021,the State Grid Corporation of China officially released the action plan of "Carbon Peak,Carbon Neutral",which emphasizes the role of the power grid as a hub and the maximum utilization of renewable energy.In recent years,China has vigorously developed clean energy,among which electric vehicles have gained unprecedented opportunities for development with their advantages of energy saving and environmental protection.However,the promotion and use of EVs has been limited to some extent because the driving range is limited by the battery capacity.In order to alleviate drivers’ range anxiety and consider the interaction between electric vehicles and the road network,this paper predicts the energy consumption level of electric vehicles and plans their travel paths.The main contents of this paper are as follows:1.The relationship between the energy consumption level of electric vehicles and the driving environment is deeply explored,and the influence of weather factors,social factors and road network characteristics on driving speed is comprehensively analyzed,so as to build a forecast model for the energy consumption of electric vehicles based on the average driving speed.2.Analyze the travel rules of different types of electric vehicles,and generate the initial path set of electric vehicles randomly according to the probability distribution.Each path in the path set contains an O-D(Origin-Destination)chain.The improved LSTM(Long Short-Term Memory)model is used to predict the average driving speed in each O-D chain in real time,and the energy consumption and time are calculated.3.Traditional neural networks are mostly used to deal with sequence problems and can only carry out short-term memory.In order to improve the prediction accuracy of the algorithm,this paper proposed a long and short term memory neural network(S-LSTM)considering the sample Similarity to measure the error credibility between model output and actual output according to the correlation between training set and test set,so as to improve the adaptability of test set and model.Part 4,example,real-time monitoring platform for hangzhou city traffic congestion index data,using S-LSTM neural network for electric cars under different driving environment to predict the average speed,combined with the energy consumption of air conditioning unit mileage electric power consumption,numerical example verifies the S-LSTM neural network compared to the neural network prediction accuracy is higher.Taking a regional road network in Hangzhou as an example,according to the energy consumption prediction results,each O-D chain in the path set is analyzed under the planning constraints,so as to provide the optimal driving path for drivers to meet the planning objectives.5.Based on Qt platform,the research content of this paper is visualized,and the "intelligent travel system" is developed by calling Baidu Map API,and the design process and complete functions of the system are shown.