Font Size: a A A

Prediction Of Arrival Distribution Of Departing Passengers In Security Area

Posted on:2020-10-23Degree:MasterType:Thesis
Country:ChinaCandidate:C WangFull Text:PDF
GTID:2392330596494387Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
With the increase in the passenger traffic volume of domestic civil aviation,the current situation of the near-saturated service utilization efficiency of civil aviation passenger security inspection areas has become a bottleneck restricting the quality of airport passenger service.The traditional security inspection area passenger forecasting method is difficult to provide very accurate data support for the operation command of the passenger service resources in the security inspection area,and the complexity of the security check service of the departing passengers has become a serious disaster area for passengers to stay in the queue.Therefore,how to design accurate and effective passenger arrival arrival distribution prediction model has become the research focus of improving the efficiency of security inspection in the terminal building.In order to solve the passenger arrival distribution prediction problem in the airport security inspection area,this paper uses the historical flight data to curve the arrival probability density of the security check of each passenger departing.Through the argumentation analysis of the fitted model,the second-order Gaussian mixture model is determined.Then,the correlation analysis between the structural parameters of the single flight fitting model and the flight planning features is carried out,and the correlation between flight characteristics such as short-term passenger traffic and structural parameters is found.Based on this,a single-flight Gaussian simulation is constructed.Prediction model for structural parameters(RBF-GMM).For the new flight that has not been executed in history,the k-nearest neighbor algorithm is used to select the historical flight that is closest to the predicted flight feature information,so as to predict the structural parameters of the fitted model.Finally,the arrival distribution of each individual flight passenger is reconstructed based on the previous forecast results and flight passenger load factor.Through the translation superposition operation of all flights in the same security inspection area,the arrival distribution prediction model of multi-flight departure passengers in the security inspection area is completed.The forecasting model constructed in this paper can combine the latest departure flight plan to predict the passenger arrival distribution of each security inspection area within one day,and to some extent exclude the interference caused by the change of flight plan.Through simulation verification analysis and prediction of multiple security inspection areas,The accuracy can reach nearly 90%,which provides a new solution for the accurate prediction of passenger arrival distribution in the security checkpoint of the terminal building.
Keywords/Search Tags:passenger transportation, arrival distribution of security area, probability density curve, Gaussian mixture model, RBF neural network, traffic to nearby flights
PDF Full Text Request
Related items