Research On Prediction Of The Arrival Time Of Connected Buses Considering The Influence Of Time Factors

Development of intelligent connected technologies will promote transformation of public transportation industry and change the operation and travel mode of public transportation in the near future.The development of intelligent connected technologies has made the real-time acquisition of bus operation data possible.The instability of operation of bus caused by the complex traffic environment and human interference makes it difficult to predict bus arrival time accurately.With the help of intelligent connected technologies,the bus arrival time can be accurately predicted to improve operation efficiency and provide decision-making basis for bus command and management.The operation of bus in urban road is restricted by many factors and the arrival time of buses will change significantly during different time periods,which increases the difficulty of prediction.Accurately predicting the bus arrival time during different time periods can improve transportation efficiency,reduce environmental pollution and save energy consumption.The paper systematically sorts out domestic and foreign papers and carries out research on improving the accuracy of bus arrival time prediction under intelligent connected vehicle environment so as to provide passengers with better travel services.Due to the uncertain characteristics of traffic data,the traffic data needs to be preprocessed to reduce the impact on prediction results.The paper analyzes the data collection process and analyzes the factors affecting the bus arrival time.Through the preprocessing of the original traffic data,the quality of data improves so as to obtain the original characteristics.Based on the time and space analysis of bus operating information on Route 1 in Zigong,the time of day is divided into peak periods and nonpeak periods,the operating route is divided into the urban part and the suburban part,and the weather conditions are divided into good part and bad part.One-way analysis of variance is used to analyze categorical variables and Pearson correlation analysis is used to analyze continuous variables respectively to keep relevant characteristics and eliminate irrelevant characteristics.Time series method,support vector machine and artificial neural network model are used to predict the bus arrival time under different time periods,space and weather conditions.According to the MAPE value of ten prediction results,average and standard deviation of MAPE are used to analyze the application scope of different methods.Finally,based on the operating data of bus on Route 1 in Zigong,a case study is carried out to verify the research.The accuracy of the prediction model is analyzed based on the running time and station waiting time.