Research On Link Travel Time Extraction Technology Based On Floating Car Data In Cloud Computing

The improvement of the economic level results public daily work and activities required for the vehicles increase rapidly, yet the urban road network resources is limited and it’s growth is so slow; The driving behavior and social responsibility consciousness of the social public civilization in our country are not along with the improvement of the material life level; The traffic data accumulation is also a rising trend in the index level, which has reached the PB level, however we can still not use the PB level of data mining to provide services to the public.So in our country, traffic congestion, traffic safety and environmental problems are becoming more and more serious.Link travel time is an important measure of traffic condition. It has been considered as the most important and can reflect the road traffic parameters of road network. In order to master the traffic operation status of road network and improve the condition of road traffic, cloud computing technique is used to efficiently deal with the massive floating car data to achieve real-time acquisition of link travel time.In this paper, the study on link travel time extraction technology includes the study of travel time estimation and prediction. In terms of link travel time estimation, we improved the instantaneous velocity time integral model to carry out link travel time estimation. The accuracy and timeliness of the algorithm are verified by comparing and analyzing. In the aspect of link travel time prediction, we have improved the particle swarm optimization algorithm to realize the least square support vector machine parameter optimization, Moreover, the Spark memory computing framework is also introduced to carry out efficient iterative. Finally, the performance of the proposed algorithm in terms of accuracy and performance is verified by experiments.In this paper, we built a frame of link travel time prediction system in the cloud environment. And We realize the efficient matching of massive floating car data and link travel time estimation using the Map Reduce model.Through the contrast analysis, the proposed method in this paper can effectively improve the efficiency of data processing, and the prediction of the link travel time based on PSO-LSSVM can meet the needs of practical applications.Finally, a series of functions and services based on real-time and efficient access technology of link travel time are supported to achieve the overall framework.