Research On Evaluation And Control Methods For Multi-source Traffic Information Data

Urban traffic is an important part of human living space, which is closely relatedto people’s travel activities, and plays an important role in the stable and sustainabledevelopment of the city and the improvement of people’s living standards. With thedevelopment of national economy, the expansion of urban population scale and thegrowth of motor vehicle population, road network scale have been difficult to satisfythe ever increasing traffic demand and the phenomenon of traffic congestion hasbecome more serious. In order to improve the use efficiency of existing road network,many big cities have built the intelligent transportation information platform, whichreal-time reflects the state of road traffic flow and accurately provides tourismreference for travelers and decision support for traffic managers. The precondition forintelligent transportation information platform effectively playing its function is tohave traffic information data with high quality. However, due to the impact of variousdetection system fault factors, traffic information data that acquired in the actualprocess of traffic information collection has quality problems, which presentingoperation condition of traffic flow occur deviation with the actual road network orinconsistent with the long-term trend of traffic flow operation, then leads to themisjudgment of traffic state. Therefore, before the traffic information data input thedatabase, the quality evaluation and control for the original data acquiring fromdifferent ways of traffic information collection, which has the vital significance foreffectively playing its function of the intelligent transportation information platform.According to the above problems, relying on subject of the National HighTechnology Research and Development (863Program)“the multi-source andmulti-dimension urban traffic state perception and interaction processing”, this paperanalyzed the problems of domestic and international research on the evaluation andcontrol methods for the quality of multi-source traffic information data and combinedwith application requirements of intelligent transportation information platform fortraffic data, and further researched the evaluation and control methods for the qualityof the multi-source traffic information data. The details are as follows: Firstly, on the basis of the project relied on and the selected topic background,this paper specifies the object, purpose and significance of the thesis and identifiesresearch ideas and chapters arrangements by analyzing related research results in thefield of data evaluation methods, and summarizing related research present situationof evaluation and control methods for quality of traffic data. Secondly, this paperanalyzes three kinds of distribution characteristics for traffic data from the angle oftime sequence and defines traffic data quality concept. Meanwhile, on the basis ofvarious traffic data and use data evaluation and control theoretical framework ofdomestic and foreign research as reference, this paper designs a evaluation andcontrol system of covering comprehensive traffic information data, which includesbasic principles, overall structure and how to build key parts by using the idea ofAnalytic Hierarchy Process (AHP) of a combination of subjective and objective indecision theory and calculate the methods of evaluation result based on evaluation ofexpected value. Thirdly, the paper makes data quality evaluation index to reflect theirrespective data sources and selects corresponding evaluation and control methodsthrough in-depth analyzing test data which from induction coil and GPS Floating carquality problems, including affected factors and performance characteristics of thedata, and selected the traffic data acquired by these two kinds of detection system inurban road network to effectively verify the evaluation and control methods andsystems. Finally, this paper summarizes research results, and analyzes the existingdeficiencies and future research directions.