Research On Traffic Streaming Data Management Method Based On Wavelet Synopses

Wavelet synopses technology is an important technology in the field of streaming data.Traffic sensor data is a typical streaming data.With the rapid development of traffic system,cities have accumulated a large number of traffic data,and the number is still growing.In the face of massive,continuous and real-time traffic data,the management of data such as storage,transmission,query and use has become an urgent problem to solve.An effective solution is to compress data with high quality and reconstruct an approximate data set efficiently from compressed data.Wavelet synopsee technique utilizes the principle of Wavelet Transform to use a small number of energy-intensive wavelet coefficients as the synopses data of the entire data set,so that the stream data can be compressed with high quality,and the approximate data can be efficiently reconstructed/query through the summary data,and it has strong applicability to compression and reconstruction in the context of streaming data.Road link travel speed is a typical traffic streaming data and it is one of the most important parameters for traffic management and induction.The data is usually gathered by estimating the floating car speed on the road.Therefore,based on the data of Section and large-scale floating car of Shenzhen,this thesis firstly obtains the road link travel speed data through the existing road link travel speed estimation model.Later,combining both the descriptive analysis method and visualization method of the data,it analyzes the data characteristics of the road link travel speed.These characteristics show the Haar wavelet class synopses method is suitable for the compression and reconstruction of road link travel speed in the context of streaming data.Finally,combined with the characteristics of road link travel speed,a model of road link travel speed profile data generation is designed.On the premise that the reconstructed data has the maximum error-bound,a Haar wavelet synopses generation algorithm,S-Greedy,is proposed to obtain a smaller number of Haar wavelet synopses.The results show that S-Greedy algorithm can effectively and efficiently generate the synopses of road link travel speed,and the errors between each reconstructed data and the original data have an error-bound.In addition,S-Greedy algorithm performs better than several existing classical algorithms in terms of execution time,number of generated synopses data(compression)and error limit control of individual elements.This approach can help city managers and relevant transportation researchers use traffic flow data more efficiently.