Research On Quick Identification Of Passengers In Specific Areas

The wide spreading of smart devices,especially mobile phones,provides an important idea for the security management in public areas,such as coach stations.Specifically,by linking IDs of passengers and Mac addresses of their mobile phones,and the real-time collection of Mac addresses by WiFi probe,passengers can be certified rapidly.The core of this thesis is to achieve associations between passenger IDs and Mac addresses of phones effectively.Two methods are proposed to obtain effective associations between them:1)Associating the passenger ID card information and the Mac address of smart devices based on D-S Evidence Theory.Firstly,a mass function is generated based on the ID capturing times and the Mac address capturing times.Then,the unrecognized part which is selected by minimum probability in each mass function avoids high evidence conflicts.Finally,the multiple mass functions are combined by the Dempster rule to get the fusion probability.Simulation experiments based on real data demonstrate the effectiveness and robustness of the evidence theory method.2)The correlation results of the evidence theory method are consistent with the intuitive experience,but it is easy to misjudge the smart devices of the station staff as the passengers.Therefore,the dynamic time warping algorithm was choosed to calculate the similarity between the passenger's ID card time series and the Mac address time series.Firstly,the correspondence between Mac address time series and ID card time series is discussed,and the shortcomings of dynamic time warping algorithm are pointed out.Then the penalty function is proposed to correct the DTW distance.Finally,the experiment is carried out on the real data.The selection of the threshold and time window is discussed,and the effectiveness of this method is proved.The correct dynamic time warping method can effectively distinguish the station staff,but the Mac address with the shortest DTW distance may have a contradiction with the intuitive experiences.In view of the advantages and disadvantages of the two methods,two methods are merged at the end of this thesis.Compared with the distance threshold value of the dynamic time warping algorithm,the evidence theory is used to constrain the results of the dynamic time warping algorithm in merged method.The Mac address which satisfies the constraint is selected as the optimal result,and the recognition rate and recognition accuracy of the fusion method are improved.