CN106803137A  Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time  Google Patents
Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time Download PDFInfo
 Publication number
 CN106803137A CN106803137A CN201710055877.4A CN201710055877A CN106803137A CN 106803137 A CN106803137 A CN 106803137A CN 201710055877 A CN201710055877 A CN 201710055877A CN 106803137 A CN106803137 A CN 106803137A
 Authority
 CN
 China
 Prior art keywords
 station
 flow
 enters
 passengers
 volume
 Prior art date
 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
 Pending
Links
 230000000739 chaotic Effects 0.000 claims abstract description 8
 238000001514 detection method Methods 0.000 claims abstract description 6
 238000005457 optimization Methods 0.000 claims abstract description 4
 238000006243 chemical reaction Methods 0.000 claims description 7
 230000000875 corresponding Effects 0.000 claims description 3
 238000003672 processing method Methods 0.000 claims description 3
 280000012370 Confidential companies 0.000 description 11
 238000000034 methods Methods 0.000 description 4
 230000002159 abnormal effects Effects 0.000 description 3
 230000000694 effects Effects 0.000 description 3
 238000010586 diagrams Methods 0.000 description 2
 238000005516 engineering processes Methods 0.000 description 2
 238000007689 inspection Methods 0.000 description 2
 230000004048 modification Effects 0.000 description 2
 238000006011 modification reactions Methods 0.000 description 2
 280000867207 Lambda companies 0.000 description 1
 238000004166 bioassay Methods 0.000 description 1
 230000005540 biological transmission Effects 0.000 description 1
 239000012141 concentrates Substances 0.000 description 1
 238000010276 construction Methods 0.000 description 1
 230000002708 enhancing Effects 0.000 description 1
 230000003014 reinforcing Effects 0.000 description 1
 230000001932 seasonal Effects 0.000 description 1
Classifications

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06Q—DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
 G06Q10/00—Administration; Management
 G06Q10/04—Forecasting or optimisation, e.g. linear programming, "travelling salesman problem" or "cutting stock problem"

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06Q—DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
 G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
 G06Q50/30—Transportation; Communications
Abstract
The present invention provides a kind of urban track traffic AFC system and enters the station in real time volume of the flow of passengers method for detecting abnormality,The Lyapunov indexes of the passenger flow data time series that entered the station using improved small data sets arithmetic calculating history first,Verify the chaotic characteristic of the sequence,The time delay and smallest embedding dimension number of chaos time sequence are determined followed by C_C methods,Phase space reconfiguration is carried out to former time series,Determine model training test sample collection,And using trellis search method Optimized model parameter on a large scale,Recycle the volume of the flow of passengers that enters the station of the model prediction day part after optimization,Then using the distribution character of the day part amount of entering the station prediction residual sequence in training sample,Finally determine that under a certain confidence level day part enters the station the confidential interval of passenger flow forecast residual error,So that it is determined that the threshold values upper limit and lower threshold of the volume of the flow of passengers abnormality detection that enters the station of future time period.The abnormality detection scope of the volume of the flow of passengers that enters the station effectively has been shunk in this discovery, reduces the rate of false alarm of data exception detection.
Description
Technical field
Entered the station in real time volume of the flow of passengers method for detecting abnormality the present invention relates to a kind of urban track traffic AFC system, belong to city
Intelligent tracktraffic technology.
Background technology
The raising of realtime passenger flow data information in City Rail Transit System to subway system service ability is most important.
However, due to the diversity and the reason such as complexity of realtime Data Transmission process of the equipment supplier in AFC system so that
The realtime passenger flow data obtained from AFC system can not completely reflect operation actual conditions, and part station is in some periods
The volume of the flow of passengers that enters the station in real time differs greatly with the volume of the flow of passengers that actually enters the station, in order to ensure to obtain the quality of passenger flow data in real time, it is necessary to right
The passenger flow data for obtaining in real time carries out abnormality detection and correction process.Set rational by each station, the day part volume of the flow of passengers
Threshold values, whether the volume of the flow of passengers that enters the station that can effectively judge realtime acquisition is abnormal, so during the threshold values upper limit and lower threshold
Reasonable set it is the most key.
According to same station, meet normal distribution with period Trip distribution the characteristics of, utilize " averagetriple standard difference method " really
It is a kind of simple and easy to do method to determine passenger flow threshold values bound, but there is exceptional value and part station in itself due to sample data
The seasonal passenger traffic fluctuation reason such as larger, cause the threshold range for drawing excessive, it is impossible to effectively to be entered the station to realtime acquisition
Passenger flow data carries out abnormality detection.
The content of the invention
Goal of the invention：In order to solve the existing counted valve of passenger flow data method for detecting abnormality that enters the station in real time based on threshold values
Value scope carries out the not good problem of Detection results of data exception detection, and the present invention provides a kind of urban track traffic AFC system
Enter the station volume of the flow of passengers method for detecting abnormality in real time, and the method determines model training test specimens by verifying the chaotic characteristic of the sequence
This collection, then using the distribution character of the day part amount of entering the station prediction residual sequence in training sample, so that it is determined that future time period
The threshold values upper limit and lower threshold of the volume of the flow of passengers that enters the station abnormality detection.
Technical scheme：To achieve the above object, the technical solution adopted by the present invention is：
A kind of urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time, comprises the following steps：
(1) phase space reconfiguration of chaos time sequence：Parttime is intercepted from the former time series of the passenger flow data that enters the station
Sequence X={ x_{i} i=1,2 ..., K }, the time delay τ and Embedded dimensions m of the parttime sequence are tried to achieve using C_C methods；
If the parttime sequence not chaos, parttime sequence is intercepted again；If the parttime sequence chaos, based on the portion
Sequence pair original time series carries out phase space reconfiguration between timesharing；x_{i}It is ith volume of the flow of passengers data that enter the station of sample, K is the part
The number of samples that time series is included.
(2) sequence chaotic characteristic judges：According to time delay τ and Embedded dimensions m that step (1) is tried to achieve, using improved
Small data sets arithmetic calculates the Lyapunov indexes of the parttime sequence：If Lyapunov indexes are for just, then it represents that during the part
Between sequence chaos；Otherwise, the parttime sequence not chaos is represented；
(3) enter the station Passenger flow forecast model in real time：Training sample set is intercepted out from the former time series after phase space reconfiguration
With checking sample set, each column data to each sample set importation carries out standardized normal distribution conversion；By the training after conversion
Sample set is trained in being brought into Support vector regression model, while (being used using grid search optimization method on a large scale
The method can Support Vector Machines Optimized regression model parameter, the prediction effect of lift scheme) determine Support vector regression mould
Type penalty coefficient C, insensitive coefficient ε and index Radial basis kernel function parameter lambda；
(4) residual distribution of predicted value and actual value：Checking sample set after conversion is brought into Support vector regression
It is trained in model, list is carried out to the checking sample period volume of the flow of passengers that enters the station using the Support vector regression model after training
Step prediction, due to similar date (working day and nonworkdays), the volume of the flow of passengers Changing Pattern that enters the station with station generally day part
It is similar, the enter the station volume of the flow of passengers and the reality of the day part of the Support vector regression model prediction after training are understood by KS assays
The residual sequence distribution that border is entered the station between the volume of the flow of passengers meets normal distribution；With y_{j}I () represents actually entering the station for the ith period of jth day
The volume of the flow of passengers, withRepresent that the prediction of jth day the ith period is entered the station the volume of the flow of passengers,When representing jth day ith
The predicated error of section, e (i) represents the prediction residual sequence of period on similar date ith, then e_{j}I () meets normal distribution, i.e.,：
E (i)~N (μ, σ^{2})
Wherein, μ is the population mean of the ith period of similar date prediction residual, and σ is the period on similar date ith to predict residual
Poor population standard deviation, N (μ, σ^{2}) expression average be μ, variance be σ^{2}Normal distribution sequence；
Checking sample set is predicted using the Support vector regression model after training, the similar date ith for obtaining
The sample mean of period prediction residualIt is respectively with sample standard deviation s (i)：
Wherein, N is the sample size of the ith period of similar date prediction residual；
(5) the passenger flow confidential interval that enters the station is determined：Understand that carrying out following classification considers according to mathematical statistics relevant knowledge：
1. as N ＞ 50,σ≈s(i)；Under given confidence level 1 α, the following period on similar date ith enters the station
Volume of the flow of passengers confidential interval is：
Wherein,For the following period on similar date ith enters the station passenger flow forecast value, Z_{α/2}For on standardized normal distribution
'sQuantile；
2. when N≤50,σ≠s(i)；Under given confidence level 1 α, the following period on similar date ith enters the station
Volume of the flow of passengers confidential interval is：
Wherein, t_{α/2}(N1) in t (N1) distributions (free degree is distributed for the t of N1)Quantile；
Due to entering the station passenger flow numerical quantity for integer in real time, thus need to the lower bound of confidential interval 1. and being 2. calculated to
On round, the upper bound rounds downwards, the confidential interval left end point as lower threshold after rounding, right endpoint is the threshold values upper limit；
(6) enter the station volume of the flow of passengers abnormality detection and processing method in real time：If certain station obtained in real time from AFC system,
The volume of the flow of passengers that enters the station of a certain period accepts and believe the value in the range of corresponding confidence interval threshold, then；Otherwise, the passenger flow that enters the station is judged
Amount is abnormal, and the volume of the flow of passengers that enters the station is predicted using the Support vector regression model after training, accepts and believe the prediction after rounding
Value.
Specifically, in the step (5), as N ＞ 50,σ ≈ s (i), i.e., nowThen constructed variableThen in given confidence
Under degree 1 α the following period on similar date ith enter the station volume of the flow of passengers confidential interval byCalculate；When N≤
When 50,σ ≠ s (i), i.e., can not be calculated to obtain population standard deviation, now, constructed variable by sample standard deviationThe following period on similar date ith enters standee under then giving confidence level 1 α
Flow confidential interval byCalculate.
Beneficial effect：The urban track traffic AFC system that the present invention is provided enters the station volume of the flow of passengers method for detecting abnormality in real time, leads to
Cross using the Support vector regression model amount of being entered the station passenger flow estimation in real time, during according to training set working day and each nonworkdays
Section regression criterion sequence statistic distribution character, can determine the volume of the flow of passengers abnormality detection threshold values that enters the station in real time, can effectively be collapsible into
The abnormality detection scope of standee's flow, and the rate of false alarm of data exception detection is reduced, detection energy of the reinforcing to abnormal passenger flow data
Power, it is ensured that obtain the accuracy and promptness of passenger flow data in real time, is passenger information service system, realtime passenger flow estimation and big
The application such as passenger flow early warning is supported there is provided reliable data, so as to enhance the service ability of subway system.
Brief description of the drawings
Fig. 1 is implementing procedure figure of the invention；
Fig. 2 is 5 to the 18 big temporary dwelling palace station amount of the entering the station distribution maps of August in 2013；
Fig. 3 is the amount of the entering the station abnormality detection threshold values comparison diagram of on December 29th, 2013；
Fig. 4 is the amount of the entering the station abnormality detection threshold values comparison diagram of on December 31st, 2013.
Specific embodiment
The present invention is further described below in conjunction with the accompanying drawings.
As shown in Figure 1 for a kind of urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time, literary grace is with changing
The small data sets arithmetic for entering calculates the Lyapunov indexes of the passenger flow time series that enters the station, and verifies the chaotic characteristic of the sequence；Using C_C
Method calculates the time delay and smallest embedding dimension number of the passenger flow time series that enters the station, and carries out phase space reconfiguration to the sequence, raw
Into model training, checking and test sample collection；And Support vector regression model parameter is entered using grid data service on a large scale
Row optimizing；Then using the volume of the flow of passengers that enters the station of chaos Support vector regression model prediction day part, with reference to hypothesis testing method,
The stochastic variable of specific distribution is obeyed with the regression criterion construction of training set under the period using the similar date, when calculating each successively
Corresponding confidential interval of the passenger flow estimation residual error under respective confidence that enter the station of section, so the volume of the flow of passengers that actually entered the station threshold values
The upper limit and lower threshold, to obtain more effective abnormality detection scope.
The present invention is made further instructions with reference to embodiment.
This example data come from the big temporary dwelling palace station of No. two lines of Line of Nanjing Subway track traffic January 20 30 days to 2014 July in 2013
Day 5：3023：The passenger flow data that enters the station between 29, the time granularity of the passenger flow data that enters the station takes 15 minutes, entering in the time period
The mathematical notation of standee's flow data is X={ x_{i} i=1,2 ..., 12600 }.
Step1, sequence chaotic characteristic judge
Choose the Time Subseries X={ x of length K=3000_{i} i=1,2 ..., 3000 }, try to achieve the portion using C_C methods
Divide the time delay τ and Embedded dimensions m of time series, calculate to obtain τ=3, m=15.
Step2, sequence chaotic characteristic judge
According to τ=3, m=15 tries to achieve the maximum Lyapunov exponent λ of the sequence using the improved method of small data sets arithmetic_{1}
=0.06 ＞ 0, therefore the subway station amount of the entering the station time series has chaotic characteristic.
Step3, enter the station Passenger flow forecast model in real time
Phase space reconfiguration is carried out to former chaos time sequence, and with 30 days in Septembers, 2013 data of 24 days of July in 2013
Used as training data, used as checking data, on December 29th, 2013 is extremely on the December 28th, 25 days 1 of September in 2013 data
The data on January 20 in 2014 after being standardized conversion to training, checking and test data set, are used as test data
Support vector regression model penalty coefficient C, insensitive coefficient ε and index footpath are determined using grid search optimization method on a large scale
To base kernel functional parameter λ, optimizing optimized after model parameter C=360, ε=3, λ=0.03.
The residual distribution of Step4, predicted value and actual value
August in 2013 is chosen to be visualized as shown in Fig. 2 can learn to August No. 18 passenger flow data that enters the station of two weeks within 5th
Workaday passenger flow Changing Pattern is roughly the same, and the passenger flow Changing Pattern between nonworkdays is also roughly the same, working day and nonwork
The Trip distribution situation for making day differs greatly；And from calculating, training sample concentrates working day residual with nonworkdays day part
The statistical parameter value difference of difference sequence is not larger, therefore the model training residual error data of day part will distinguish working day and nonworkdays,
Obtain the day part amount of entering the station residual error data in model training sample using Support vector regression model, so obtain working day with
The statistical parameter value of nonworkdays day part model prediction residual error, i.e. sample average, sample variance, sample number.In order to test this
Model to the effect of the two class date amount of entering the station abnormality detections, now on December 29th, 2013 to entering the station between 20 days January in 2014
Volume of the flow of passengers data carry out validity check, and be given two days on December 29,31st, 2013 (i.e. Sunday and Tuesday) threshold value setting and
The specific calculating process of abnormality detection.
29 days 6 December in 2013 is obtained using the Support vector regression model for training:30 to 6:44 periods (period 5)
The passenger flow forecast value that enters the station isThe volume of the flow of passengers that actually enters the station is y=27, and such date (nonworkdays) this period is trained
Sample residual sequence carries out KS inspections, and to judge its distribution situation, it is assumed that its Normal Distribution, taking significance is
0.05, it is 0.995 to be calculated with SPSS Statistics and must check probable value, much larger than significance, meanwhile, it is each to remaining
Period residual sequence data carry out KS inspections, its inspection probable value all be more than significance, so it is believed that such date it is each
The residual sequence Normal Distribution of period.Such training sample residual sequence sample mean of period on date 5Sample
This standard difference s (i)=20.44.
31 days 6 December in 2013 is obtained using the Support vector regression model for training:30 to 6:44 periods (period 5)
The passenger flow forecast value that enters the station isThe volume of the flow of passengers that actually enters the station is y=120, and such date (nonworkdays) this period is trained
Sample residual sequence carries out KS inspections, and to judge its distribution situation, it is assumed that its Normal Distribution, taking significance is
0.05, it is 0.665 to be calculated with SPSS Statistics and must check probable value, much larger than significance, meanwhile, it is each to remaining
Period residual sequence data carry out KS inspections, its inspection probable value all be more than significance, so it is believed that such date it is each
The residual sequence Normal Distribution of period.Such training sample residual sequence sample mean of period on date 5
Sample standard deviation s (i)=15.12.
Step5, determination are entered the station passenger flow confidential interval
To 29 days 6 December in 2013:30 to 6:44 periods (period 5) are analyzed, sample number N=27 ＜ 50, therefore, the period
The sample standard deviation of residual sequence is larger with population standard deviation deviation, and parameter is substituted into (5) Shi Ke get, and the period enters the station the volume of the flow of passengers
It is to enter the station the confidential interval of 1 α=0.99, i.e. this period volume of the flow of passengers valid value range for [0,87] in confidence level, its left and right end points
Value is required lower threshold and the threshold values upper limit, and the threshold values of other periods is calculated by that analogy, and the day volume of the flow of passengers that enters the station is examined extremely
Side valve value is as shown in Figure 3.
To 31 days 6 December in 2013:30 to 6:44 periods (period 5) are analyzed, sample number N=68 ＞ 50, therefore, the period
The sample standard deviation of residual sequence is approximately equal to population standard deviation, and parameter is substituted into (4) Shi Ke get, and the period amount of entering the station is in confidence level
It is the confidential interval of 1 α=0.99, i.e. this period enters the station volume of the flow of passengers valid value range for [81,158], and its left and right endpoint value is
Required lower threshold and the threshold values upper limit, the threshold values of other periods are calculated by that analogy, and enter the station volume of the flow of passengers abnormality detection threshold values in the day
As shown in Figure 4.
Step6, realtime enter the station volume of the flow of passengers abnormality detection and processing method
If certain station for being obtained in real time from AFC system, a certain period enter the station passenger flow data in threshold range,
Accept and believe the value；It is no, then the passenger flow data exception that enters the station is can determine that, rear data are now rounded using model predication valueAs this
The volume of the flow of passengers that enters the station in real time of period.
The above is only the preferred embodiment of the present invention, it should be pointed out that：For the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (2)
1. a kind of urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time, it is characterised in that：Including following step
Suddenly：
(1) phase space reconfiguration of chaos time sequence：Parttime sequence X is intercepted from the former time series of the passenger flow data that enters the station
={ x_{i} i=1,2 ..., K }, the time delay τ and Embedded dimensions m of the parttime sequence are tried to achieve using C_C methods；If the portion
Divide time series not chaos, then intercept parttime sequence again；If the parttime sequence chaos, based on the parttime
Sequence pair original time series carries out phase space reconfiguration；x_{i}It is ith volume of the flow of passengers data that enter the station of sample, K is the parttime sequence
The number of samples that row are included.
(2) sequence chaotic characteristic judges：According to time delay τ and Embedded dimensions m that step (1) is tried to achieve, using improved decimal
The Lyapunov indexes of the parttime sequence are calculated according to amount method：If Lyapunov indexes are for just, then it represents that the parttime sequence
Row chaos；Otherwise, the parttime sequence not chaos is represented；
(3) enter the station Passenger flow forecast model in real time：Training sample set is intercepted out from the former time series after phase space reconfiguration and is tested
Card sample set, each column data to each sample set importation carries out standardized normal distribution conversion；By the training sample after conversion
Collection is brought into Support vector regression model and is trained, at the same using grid search optimization method on a large scale determine to support to
Amount machine regression model penalty coefficient C, insensitive coefficient ε and index Radial basis kernel function parameter lambda；
(4) residual distribution of predicted value and actual value：Checking sample set after conversion is brought into Support vector regression model
In be trained, the checking sample period is entered the station using the Support vector regression model after training to carry out single step pre for the volume of the flow of passengers
Survey, with y_{j}I () represents the volume of the flow of passengers that actually enters the station of the ith period of jth day, withRepresent that the prediction of the ith period of jth day is entered the station
The volume of the flow of passengers,The predicated error of the ith period of jth day is represented, e (i) represents the pre of period on similar date ith
Residual sequence is surveyed, then e_{j}I () meets normal distribution, i.e.,：
E (i)~N (μ, σ^{2})
Wherein, μ is the population mean of the ith period of similar date prediction residual, and σ is the ith period of similar date prediction residual
Population standard deviation, N (μ, σ^{2}) expression average be μ, variance be σ^{2}Normal distribution sequence；
Checking sample set is predicted using the Support vector regression model after training, the period on similar date ith for obtaining
The sample mean of prediction residualIt is respectively with sample standard deviation s (i)：
Wherein, N is the sample size of the ith period of similar date prediction residual；
(5) the passenger flow confidential interval that enters the station is determined：Understand that carrying out following classification considers according to mathematical statistics relevant knowledge：
1. as N ＞ 50,σ≈s(i)；Under given confidence level 1 α, the following period on similar date ith enters the station passenger flow
Measuring confidential interval is：
Wherein,For the following period on similar date ith enters the station passenger flow forecast value, Z_{α/2}For on standardized normal distributionPoint
Site；
2. when N≤50,σ≠s(i)；Under given confidence level 1 α, the following period on similar date ith enters the station passenger flow
Measuring confidential interval is：
Wherein, t_{α/2}(N1) in t (N1) distributionQuantile；
Because the passenger flow numerical quantity that enters the station in real time is for integer, therefore the lower bound of the confidential interval to 1. and being 2. calculated is needed to take upwards
The whole, upper bound rounds downwards, and the confidential interval left end point as lower threshold after rounding, right endpoint is the threshold values upper limit；
(6) enter the station volume of the flow of passengers abnormality detection and processing method in real time：If certain station, a certain obtained in real time from AFC system
The volume of the flow of passengers that enters the station of period accepts and believe the value in the range of corresponding confidence interval threshold, then；Otherwise, judge that the volume of the flow of passengers that enters the station is different
Often, the volume of the flow of passengers that enters the station is predicted using the Support vector regression model after training, accepts and believe the predicted value after rounding.
2. urban track traffic AFC system according to claim 1 enters the station volume of the flow of passengers method for detecting abnormality, its feature in real time
It is：In the step (5), as N ＞ 50,σ ≈ s (i), i.e., now
Then constructed variableThen the following period on similar date ith enters under given confidence level 1 α
Standee's flow confidential interval byCalculate；When N≤50,σ ≠ s (i), i.e., can not
Population standard deviation, now, constructed variable are calculated to obtain by sample standard deviationThen give
Under confidence level 1 α the following period on similar date ith enter the station volume of the flow of passengers confidential interval by
Calculate.
Priority Applications (1)
Application Number  Priority Date  Filing Date  Title 

CN201710055877.4A CN106803137A (en)  20170125  20170125  Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

CN201710055877.4A CN106803137A (en)  20170125  20170125  Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time 
Publications (1)
Publication Number  Publication Date 

CN106803137A true CN106803137A (en)  20170606 
Family
ID=58987305
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

CN201710055877.4A Pending CN106803137A (en)  20170125  20170125  Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time 
Country Status (1)
Country  Link 

CN (1)  CN106803137A (en) 
Cited By (4)
Publication number  Priority date  Publication date  Assignee  Title 

CN108876029A (en) *  20180611  20181123  南京航空航天大学  A kind of passenger flow forecasting based on the adaptive chaos firefly of double populations 
CN109462517A (en) *  20181024  20190312  云南电网有限责任公司信息中心  A kind of method, system and the equipment of the data monitoring towards digital electric network business 
CN110443657A (en) *  20190819  20191112  泰康保险集团股份有限公司  Customer traffic data processing method, device, electronic equipment and readable medium 
CN111121946A (en) *  20191218  20200508  东南大学  Method for accurately determining abnormal value at multiple points in large dynamic range and large discrete single area 
Citations (3)
Publication number  Priority date  Publication date  Assignee  Title 

CN103268519B (en) *  20130426  20160525  哈尔滨工程大学  Based on the powersystem shortterm load forecasting method and the device that improve Lyapunov index 
CN106096226A (en) *  20160527  20161109  腾讯科技（深圳）有限公司  A kind of data assessment method, device and server 
WO2016180323A1 (en) *  20150512  20161117  杭州海康威视数字技术股份有限公司  Method and device for calculating customer traffic volume 

2017
 20170125 CN CN201710055877.4A patent/CN106803137A/en active Pending
Patent Citations (3)
Publication number  Priority date  Publication date  Assignee  Title 

CN103268519B (en) *  20130426  20160525  哈尔滨工程大学  Based on the powersystem shortterm load forecasting method and the device that improve Lyapunov index 
WO2016180323A1 (en) *  20150512  20161117  杭州海康威视数字技术股份有限公司  Method and device for calculating customer traffic volume 
CN106096226A (en) *  20160527  20161109  腾讯科技（深圳）有限公司  A kind of data assessment method, device and server 
NonPatent Citations (2)
Title 

张成: "城市轨道交通客流特性分析", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * 
薛向红、薛小锋: "基于相空间重构—SVM的公交客流量预测研究", 《江苏技术师范学院学报》 * 
Cited By (4)
Publication number  Priority date  Publication date  Assignee  Title 

CN108876029A (en) *  20180611  20181123  南京航空航天大学  A kind of passenger flow forecasting based on the adaptive chaos firefly of double populations 
CN109462517A (en) *  20181024  20190312  云南电网有限责任公司信息中心  A kind of method, system and the equipment of the data monitoring towards digital electric network business 
CN110443657A (en) *  20190819  20191112  泰康保险集团股份有限公司  Customer traffic data processing method, device, electronic equipment and readable medium 
CN111121946A (en) *  20191218  20200508  东南大学  Method for accurately determining abnormal value at multiple points in large dynamic range and large discrete single area 
Similar Documents
Publication  Publication Date  Title 

CN106803137A (en)  Urban track traffic AFC system enters the station volume of the flow of passengers method for detecting abnormality in real time  
CN104134349B (en)  A kind of public transport road conditions disposal system based on traffic multisource data fusion and method  
CN102496069B (en)  Cable multimode safe operation evaluation method based on fuzzy analytic hierarchy process (FAHP)  
CN105069537B (en)  A kind of construction method of combination type air Quality Prediction  
CN104778837A (en)  Multitime scale forecasting method for road traffic running situation  
CN100454354C (en)  Road traffic accident multihappening section identifying method  
CN106845768B (en)  Bus travel time model construction method based on survival analysis parameter distribution  
CN101783075A (en)  System for forecasting traffic flow of urban ringshaped roads  
CN102074124A (en)  Dynamic bus arrival time prediction method based on support vector machine (SVM) and Hinfinity filtering  
CN103984994B (en)  Method for predicting urban rail transit passenger flow peak duration  
CN103136539A (en)  Grounding grid corrosion rate level prediction method  
CN105678481A (en)  Pipeline health state assessment method based on random forest model  
CN103606274A (en)  Urban road traffic state assessment method  
CN104298881A (en)  Bayesian network model based public transit environment dynamic change forecasting method  
CN104021430A (en)  Method for analyzing uncertainty of passenger flow of urban mass transit terminal  
CN107766636A (en)  A kind of urban intersection safe evaluation method based on extreme value theory and microscopic simulation  
CN104765974A (en)  Urban rail transit passenger flow density index calculating method  
CN111400371A (en)  Voltage correlation verificationbased user variable relationship identification method  
CN104598984A (en)  Fuzzy neural network based fault prediction method  
CN103440410A (en)  Main variable individual defect probability forecasting method  
CN104050319A (en)  Method for realtime online verification of complex traffic control algorithm  
CN103942251A (en)  Method and system for inputting high altitude meteorological data into database based on multiple quality control methods  
CN108898311A (en)  A kind of data quality checking method towards intelligent distribution network repairing dispatching platform  
CN102521698B (en)  The critical process identification of largescale construction engineering quality and method for supervising  
CN104575049A (en)  Viaduct ramp intelligent inducing control method and device based on array radars 
Legal Events
Date  Code  Title  Description 

PB01  Publication  
PB01  Publication  
SE01  Entry into force of request for substantive examination  
SE01  Entry into force of request for substantive examination  
RJ01  Rejection of invention patent application after publication 
Application publication date: 20170606 

RJ01  Rejection of invention patent application after publication 