Research On Reliability Improvement Technology Of Bus Arrival Time Prediction

Prediction of bus arrival time is an important technical basis for realizing intelligent public transport.In practical applications,the bus information perception equipment,communication transmission and other links may be affected by the factors such as system failure and environmental change,resulting in abnormal bus operation data,which directly affects the prediction result.At the same time,the operation of the bus is affected by a variety of traffic factors,so bus travel time has certain volatility and uncertainty.However,the existing forecasting methods generally ignore the influence of these factors.Therefore,in this paper,aiming at above problems,in order to improve the reliability of the bus arrival time prediction,the research is carried out.It is of great significance to improve the practical application effect of the bus arrival time prediction system.The paper starts with the analysis of the key factors that affect the reliability of bus arrival time prediction,and then carries out the improvement research on the prediction reliability from the two aspects of data processing and prediction methods.For the problem of data anomaly,the data missing and data error evaluation method are established,and the algorithm of bus abnormal data repair is proposed to provide the available data for the subsequent prediction.For bus travel time volatility problem,an improved Markov chain prediction algorithm is proposed to predict the travel time between bus stations.On this basis,the prediction model of bus arrival time is established.Finally,the validity of the above mentioned reliability enhancement method is proved by the application test.The main contents of the paper include:(1)Analysis and processing of bus data anomaly.This paper generalizes and summarizes the situation of abnormal data,and discusses the impact of data anomaly for prediction reliability.Then,the data integrity evaluation method under data bus arrival time prediction angle is established for data missing,and the data correctness evaluation method based on statistics and physical identification is proposed for the data error.Finally,an abnormal data repair algorithm based on front vehicle data is proposed.And Chongqing bus data is used to test the data processing method proposed,the experimental results show that the proposed method can effectively identify and repair abnormal bus data.(2)The reliability improvement method of the bus arrival time prediction model.Paper considers the volatility of bus travel time,an improved Markov chain prediction algorithm is proposed to predict the travel time between the bus stations,the autocorrelation coefficient is introduced to dynamically adjust the information weight of each vehicle in the prediction model,and the prediction results are corrected in real time by moving error to enhance the robustness of the prediction model.Finally,combined with inter Station travel time and stop parking time,the prediction model of bus arrival time is established,and the reliability evaluation indexes of the predicted results are determined.Experiments show that the prediction method is superior to the existing method and has better prediction result.Finally,the VS2010 development tool is used to design and implement the bus arrival time prediction system,and the application tests of only data processing methods,comprehensive utilization of data processing methods and improved prediction methods are carried out respectively.The result proves that the method proposed in this paper can effectively enhance the reliability of bus arrival time prediction results.