The Research On Traffic State Forecasting For The Mobile Crowdsensing Network

With the growing popularity of the smart phone,mobile crowdsensing network collects all kinds of sensory data using intelligent terminal with a variety of sensors,and transfers to data center through the mobile Internet.These data can be easily and efficiently used in traffic,social and other fields.Compared with traditional intelligent transportation network use induction coil,GPS,floating car to collect data,mobile crowdsensing network has great advantage at deployment and maintenance costs.So,mobile crowdsensing network has been more and more used in intelligent transportation,and is becoming a new research hotspot.Real-time traffic state forecasting can be used for traffic flow control and guidance,so it can effectively alleviate traffic congestion and improve the utilization efficiency of road.Based on the vehicle speed data obtained from the mobile crowdsensing network and the accurately traffic state forecasting,we can reduce the cost of intelligent transportation infrastructure construction and improve the efficiency of the intelligent traffic management.Therefore,the research on the traffic state forecasting based on the mobile crowdsensing network is of traffic movement perception environment has important sense.To deal with the lack of foresight and having obvious hysteresis in traditional traffic control and guidance based on real-time data this article discuss a new method to forecast the road traffic condition.Based on the vehicle speed data from mobile crowdsensing network,support vector regression algorithm,and a seasonal operator,this artical present the CSA-SSVR forecasting model,which use cuckoo search algorithm to determine the parameter in seasonal support vector regression algorithm.Experiments show the CSA-SSVR forecasting model is more efficient and has higher accuracy at traffic state forecasting in mobile crowdsensing network.Due to the user data collected by intelligent terminal has some problem in data quality such as missing,inaccurate,and even wrong data,this article present a method to control data quality in mobile crowdsensing network.This method subdivide the data quality control for individual control and data center control.Intelligent terminal has ability to calculate data.This ability can reduce the pressure of data center,and effectively control the data quality.Experiments show that this method can further improve the precision of traffic state forecasting algorithm.