Research On Evaluation And Control Methods For Expressway Dynamic Traffic Data

Urban highway system is to assume the city long-distance transport, links the city traffic distribution and ensure the city traffic running fast and smooth functions, these functions allow urban highway system carrying the main part of the volume of traffic. With the rapid increase in traffic volume, urban highway system can not function in normal because of frequent traffic jams. Urban highway network basically completed under the premise of, traffic managers use dynamic traffic management strategies, such as traffic-induced measure. However, the necessary condition for effective traffic management measures is seized of accurate and timely data on traffic parameters. Owing to the failure of detection system, the violations of the driver, the badness of detecting environments, the detected traffic parameters data can not reflect the actual operation of the traffic flow or correspond with traffic flow of long-run trends. Lack of accurate and timely traffic parameters data,traffic management strategy can not effective no matter how accurate models and algorithms. Therefore evaluating and controlling of urban highway traffic parameters data, and providing complete, effective, real-time traffic parameters data for various traffic management strategies is a very crucial and urgent issues.In this paper, the types of parameter of urban highway traffic including traffic volume, speed, occupy location traffic parameters and travel time,travel speed range of traffic parameters. As an example to location traffic parameters, analyze these data's change trend at different time intervals, time characteristics, space characteristics, quality problems by historical data. On this basis, put forward the urban highway dynamic traffic parameters data quality evaluation index system, design of the evaluation and control methods of the traffic parameters data quality from the micro and macro point of view.In this paper, time serial is divided horizontal time-serial and longitudinal time-serial two types. By analyzing the time serial of traffic parameters data, their characteristics we reached as follows. The horizontal time serial, the volume of each month, different days per week or non-working days changes of different; the same traffic parameters changes largely within 24 hours of the day; the three parameters are interrelated., the same traffic parameters fluctuate slightly around a range of data values in a certain range in the same time period with the adjacent working day. The longitudinal time-serial is first put forward in this paper. Analysis of traffic parameters is the foundation to determine the types of traffic parameters data quality, establish evaluation index, design evaluation and control methods.After statistical analyzing historical data, the types of urban highway system traffic parameters data is determined as lost data, event data, false data, truth data. Under the guidance of the principle of the data evaluation, design the micro-evaluation index system and macro -evaluation index system,which to adapt location traffic parameters data and range traffic parameter data. The aim of micro-evaluation index system is evaluating and signing the types of each micro-traffic parameter data quality effectively and correctly. Micro-indicators is comprised by existent evaluation index, break and true evaluation index; macro-indicators is comprised by loss rate, incident rate, false rate, true rate. No matter micro-evaluation or macro–evaluation serve for data quality control. Data quality evaluation index system is first put forward by this paper, which overcome the problem of repeated evaluation and missed evaluation. Micro-evaluation and macro-evaluation index system makes data quality evaluation and control more targeted, thus improves the efficiency of evaluation and control.The orders of micro-traffic parameters data quality evaluation are as follows. First of all, utilize existent evaluation method to evaluate lost data. What's more, utilize break evaluation method to evaluate incident data. Finally utilize true evaluation method to evaluate false data and true data. All the evaluated data must note the type of data quality, in order to facilitate the use of micro-data quality control. When there is a continuous problem about data quality, choose moving average method of the longitudinal time-serial to control the data quality. Or else, choose exponential smoothing of horizontal time-serial. Incident data evaluation is first put forward in data quality evaluation, consequently detect incident earlier; the method of incident data evaluation which designed in this paper can detect small and medium-sized incident which can not detected by had methods. False data evaluation had been advance from changed value with the traffic flow to fixed maximum and minimum value. Moving average method of longitudinal time-serial used in data quality control not only shortened the time to control time and reduce the memory, but also increased the number of repaired data.Macro–evaluation serve for data quality estimate and control is first put forward in this paper. The bases of Macro–evaluation are the actual value and theoretical value of traffic parameter data in Specific road network and Specific time. Then calculate each index of Macro–evaluation. Not only analyze what induce data quality problems in this paper, but also analyze what reason bring what problem. When a certain type of the actual value of macro-index exceed the stated actual of macro-index, take appropriate method directly by the reasons induce this type of data quality. Therefore control method is clearer and save lots of manpower and material resources.Use the actual data of Shanghai highway the loop detector to verify evaluation and control methods of micro-traffic parameter data quality. Designed method can effectively evaluate the type of date quality and choose appropriate method of data quality control to repair the data effectually according to the specific circumstance. Error can meet the needs of engineering and the methods have a good value of engineering applications.