Study On The Key Technologies Of Prediction Of Urban Road Congestion

Urban traffic congestion is becoming more and more serious problem all over the world, which affects development and people's daily life. Traffic congestion also occurs in large and medium-sized cities of our country which badly affects development of the country.We can know from the operation and management of urban traffic system that traffic congestion can be avoided or lightened if we can reliably estimate the traffic congestion which will occur in a few minutes or occurred already a few minutes ago, and then take prompt, effective traffic management measures.Therefore, it is of great significance for prediction of traffic congestion to improve the efficiency and security of traffic system.The paper focuses on the study on the key technologies of prediction of urban road congestion.The followings are the main contents in the paper.1. Definition and measurement methods of traffic congestion In this part, author made the research on definition and measurement methods of traffic congestion in order to set the research objective of this paper and plan the whole work. And this research leads the following parts of the whole paper. The contents in this part conclude: introduced the definition of traffic congestion; made analysis of the characteristics of traffic congestion; summarized the main parameters which can describe the traffic status; established the quantitative and qualitative indexes which can measure traffic congestion.2. Research on methods of pretreatment of short-time traffic parameters based on the characteristics analysisThe precise short-time traffic parameters must be provided to the prediction system in order to make sure that the prediction system work well. And the whole contents conclude: presented the method of short-term traffic parameters pretreatment of urban highway based on characteristic analysis. The characteristics of short-term traffic flow data of urban highway were analyzed to make the example. Based on the analysis, the process of pretreatment was designed including identifying and modifying abnormal data, and de-noising and smoothing analysis of normal data based on the analysis; presented the location and cancel method of abnormal data and the de-noise method based on wavelet analysis. Then the methods were tested with traffic flow data. The conclusion showed that the new pretreatment methods were applicable to the real condition and had good result.3. Automatic identification of traffic congestionAutomatic identification of traffic congestion is an important part of the prediction system. It is of great significance to identify the urban congestion well and truly and timely for prediction of traffic congestion. And degree of congestion is an important index in prediction system.In this part, based on the analysis of the conventional methods, author presented a quantitative measurement method for identification of urban road based on real-time traffic data. The numerical test results showed that the model can well work on the automatic identification of urban road.4. Study on the prediction methods of short-time traffic parametersThe degree of congestion is decided by the traffic demand. Therefore, the prediction of short-time traffic parameters is an important part of the prediction of status of traffic congestion. In this part, conventional short-time prediction methods were summarized and introduced at firstly, and then the new short-time prediction methods were presented. It includes: designed an adaptive exponential smoothing model for dynamic forecasting aiming at the deficiency of traditional exponential smoothing model. The new model can forecast the data in series with adaptive exponent. It also has simple formula and can easily be carried out which preserved the strongpoint of traditional exponential smoothing model. It can be used for on-line forecasting; put forward integrated prediction model combine wavelet analysis and BP neural network. This model has the strongpoint of traditional integrated prediction model based on wavelet analysis in non-linear analysis and has improvement in precision and speed; put forward project of shorter-time prediction. We tested the methods with traffic information. The conclusion showed that the new methods were applicable to the real condition and had good result.5. Establishment of decision support system for urban traffic congestion management based on the prediction of traffic congestionBased on the identification and prediction of traffic congestion, we set up decision support system for urban traffic congestion management. The main contents of this part includes: designed the function and structure of the decision support system; analyzed the key implementation technologies and key support technologies.It is not easy to solve the traffic congestion. The solution to this problem depends on advanced technologies and policies. Policy factors are not involved in this paper, we try to solve traffic congestion by using latest technologies. The outcomes of this paper are very helpful in realization of prediction of traffic congestion and establishment of decision support system for urban traffic congestion management.