CN102156822A - Pedestrian traffic data assembly multi-step forecasting method - Google Patents
Pedestrian traffic data assembly multi-step forecasting method Download PDFInfo
- Publication number
- CN102156822A CN102156822A CN2011101093777A CN201110109377A CN102156822A CN 102156822 A CN102156822 A CN 102156822A CN 2011101093777 A CN2011101093777 A CN 2011101093777A CN 201110109377 A CN201110109377 A CN 201110109377A CN 102156822 A CN102156822 A CN 102156822A
- Authority
- CN
- China
- Prior art keywords
- pedestrian traffic
- traffic data
- data
- pedestrian
- prediction
- 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
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention provides a pedestrian traffic data assembly multi-step forecasting method. The method comprises the following steps: S1, providing a pedestrian traffic data long-term trend development mode with a longitudinal time sequence, and making multi-step forecasting on whole-day data of intraday pedestrian traffic data by utilizing the longitudinal time sequence; S2, making multi-step forecasting on short-term preset data of the intraday pedestrian traffic data by utilizing the transverse time sequence of the pedestrian traffic data long-term trend development mode; S3, sampling pedestrian traffic real-time data, and comparing the pedestrian traffic real-time data with the data at the moment in the long-term trend multi-step forecasting so as to acquire a forecasting error; and S4, comparing the forecasting error with an error threshold, calling a forecasting result in the long-term trend multi-step forecasting if the forecasting error of the long-term trend multi-step forecasting data is smaller than the error threshold, or calling the data at the moment in the short-term present multi-step forecasting as the forecasting result if the forecasting error is more than the error threshold.
Description
Technical field
The present invention relates to a kind of pedestrian traffic data predication method, relate in particular to a kind of pedestrian traffic data combination multistep forecasting method.
Background technology
Crowded place density of personnel early warning system, be to rely on the video equipment that is installed in the different location, gather the video image of monitored area in real time, by the real-time crowd's passenger flow that reflects in the video image is carried out data statistics and analysis, thereby realize the crowd density early warning prediction of crowded place.If conditions permit, in wishing to begin every day, administrative authority can understand the pedestrian traffic state of following each sampling time interval on the same day, so that rational plan is made in the pedestrian traffic management to a day in advance, therefore need the pedestrian traffic state of a kind of pedestrian traffic data predication method with the prediction whole day.
Summary of the invention
The objective of the invention is to propose a kind of pedestrian traffic data combination multistep forecasting method, can predict the pedestrian traffic state of whole day, so that rational plan is made in one day pedestrian traffic management in advance.
In order to achieve the above object, the invention provides a kind of pedestrian traffic data combination multistep forecasting method, this method comprises: step S1: a group traveling together's traffic data secular trend development model is provided, it has a vertical time series, utilizes this vertical time series that the whole day data of the pedestrian traffic data on the same day are made multi-step prediction; Step S2: described pedestrian traffic data secular trend development model also has a horizontal time series, utilizes this horizontal time series that the short-term trend of the times data of the pedestrian traffic data on the same day are made multi-step prediction; Step S3: sampling pedestrian traffic real time data, these moment data of pedestrian traffic real time data and described secular trend multi-step prediction are compared, to obtain predicated error; Step S4: with described predicated error and error threshold comparison; If the predicated error of secular trend multi-step prediction data is then called predicting the outcome of secular trend multi-step prediction less than described error threshold; If predicated error, is then called these moment data of described short-term trend of the times multi-step prediction greater than described error threshold as predicting the outcome.
Pedestrian traffic data combination multistep forecasting method of the present invention adopts the vertical time series of moving average model(MA model) utilization that the pedestrian traffic data are carried out multi-step prediction among the wherein said step S1, formula is
In the formula:
The vertical predicted value of pedestrian traffic data for current sampling time interval; z
K(t) be K the historical data in pedestrian traffic data long-run development pattern front on the contained same day; N is the contained historical data quantity of pedestrian traffic data long-run development pattern.
Pedestrian traffic data combination multistep forecasting method of the present invention, pedestrian traffic data short-term trend of the times multistep forecasting method comprises among the wherein said step S2: the S21 Forecasting Methodology is chosen with S22 prediction step number and is determined.
Pedestrian traffic data combination multistep forecasting method of the present invention, Forecasting Methodology is chosen and is adopted the horizontal time series of moving average model(MA model) utilization to carry out pedestrian traffic data short-term trend of the times multi-step prediction among the wherein said step S21.
Pedestrian traffic data combination multistep forecasting method of the present invention adopts predicated error continuously less than the maximum sampling time interval quantity of error threshold among the wherein said step S22, the prediction step number of determining as off-line.
Pedestrian traffic data combination multistep forecasting method of the present invention, the index of behavioral characteristics by obtaining the data time sequence among the wherein said step S22, comprise undulatory property characteristic exponent, tendency characteristic exponent and uncertain characteristic exponent, calculate online definite prediction step number with neural network model again.
Pedestrian traffic data combination multistep forecasting method of the present invention, the computing method of wherein said undulatory property characteristic exponent, tendency characteristic exponent and uncertain characteristic exponent suc as formula
A is the undulatory property characteristic exponent in the formula, and M calculates selected correlation time of interval quantity, z for the undulatory property characteristic exponent
K 'Be the measured value of k ' individual correlation time of interval pedestrian's traffic data, z
EBe the mean value of interval pedestrian's traffic data correlation time, δ is a zoom factor;
B is the tendency characteristic exponent in the formula, z
K '+dIt is k '+d the correlation time of the measured value of pedestrian's traffic data at interval; D is that slope calculates interval quantity correlation time, and η is a zoom factor;
C is uncertain characteristic exponent in the formula, f
K 'Be the blur level of interval pedestrian's traffic data correlation time, λ is a zoom factor.
Can understand the pedestrian traffic state of following each sampling time interval on the same day in begin every day by the present invention, and then obtain pedestrian traffic status predication result, need not to obtain the data on the same day, help administrative authority rational plan is made in one day pedestrian traffic management in advance.
Description of drawings
Fig. 1 is the process flow diagram of pedestrian traffic data combination multistep forecasting method of the present invention.
Embodiment
The invention will be further described in conjunction with the embodiments below with reference to accompanying drawing.
According to time organizational form difference, the present invention is divided into horizontal time series and vertical time series two classes with the time series of pedestrian traffic data.Wherein, laterally time series is meant the data sequence of arranging by arbitrary day time sequencing; Vertically time series is meant in chronological sequence pedestrian traffic data sequence of same period of series arrangement certain day.
The pedestrian traffic data time sequence of particular spatial location has secular trend, the short-term trend of the times and random fluctuation three specific characters usually.(1) secular trend, the specific region generally has more stable socio-economic activity pattern, promptly go to work, go to school, activity such as shopping has certain rules in time and spatial distributions, causes different same date in week (Monday, Tuesdays ... Sunday) pedestrian traffic pattern has stronger similarity.The present invention is not with same monitoring position, the characteristics of the same supplemental characteristic time series of same date with similarity are called secular trend.(2) the short-term trend of the times, because the influence of factors such as pedestrian traffic incident, the phenomenon of secular trend may appear departing from pedestrian's rule of specific region in short-term, the present invention is referred to as the pedestrian traffic seasonal effect in time series short-term trend of the times.(3) random fluctuation, except secular trend, the short-term trend of the times, also there is tangible random fluctuation in the pedestrian traffic data, in order to eliminate random fluctuation to the influence that the pedestrian traffic management decision produces, it suitably should be carried out filtering.
For a certain date, if the evolution of pedestrian traffic data time sequence relatively meets the long-run development pattern, then think a kind of normal pedestrian's traffic behavior, otherwise, be called the normal pedestrian traffic state in boundary.The predictability of normal pedestrian's traffic behavior is stronger, can carry out early warning to it, and unusual pedestrian traffic state generally be difficult to prediction, can only carry out Realtime Alerts to it.
For reliable information basis being provided can for the judgement of pedestrian traffic state, the present invention is by analyzing the vertical time series of pedestrian traffic data, and the off-line of design data long-run development pattern makes up and the online updating method.
By observing the pedestrian traffic data as can be known, often there is difference in same position not on the same day data time scale, causes data long-run development pattern to be difficult to extract.Therefore, the present invention proofreaies and correct the time scale of raw data earlier, is benchmark with 00:00:00 promptly, is step-length with the time scale, with same position not on the same day pedestrian traffic raw data time scale artificially align.For example time series 00:00:04,00:00:10,00:00:16, carry out index correction according to the 6s time scale after, become 00:00:06,00:00:12,00:00:18.
For a certain monitoring position, different weeks, of even date pedestrian traffic pattern was more similar, and in view of the above, the present invention's design is based on the long-run development mode construction method on nature date.Specifically, at arbitrary sampling time interval, the screening secular trend comparatively similar continuous some weeks the phase same date the pedestrian traffic data, and with it as judging whether next week of even date data meet the foundation of secular trend.
Long-run development pattern that it should be noted that the pedestrian traffic data is a relative notion, can be not unalterable, but have certain evolution property in season.That is to say that for relatively contiguous week, the long-run development pattern of pedestrian traffic data is more stable, but for time span for bigger week, the long-run development pattern of pedestrian traffic data has bigger difference.Therefore, make up the long-run development pattern week quantity can not be excessive, generally adopt 4-5 week to be advisable.
Now introduce a kind of structure and update method of pedestrian traffic data long-run development pattern, described pedestrian traffic data have time scale and time scale, and comprise pedestrian traffic raw data and pedestrian traffic real time data, this method comprises: step S1 ': obtain many days pedestrian traffic raw data at least one position and storage; Step S2 ': described pedestrian traffic raw data is carried out time scale proofread and correct; Step S3 ': arbitrary day pedestrian traffic original data sequence of arranging in chronological order of screening same position and conduct be time series laterally, the pedestrian traffic original data sequence of the same time scale of arranging in chronological order of screening same position certain day and as vertical time series, based on described two sequences to make up pedestrian traffic data long-run development pattern; Step S4 ': sampling pedestrian traffic real time data is also carried out pre-service; Step S5 ': the pedestrian traffic raw data in pretreated pedestrian traffic real time data and the described vertical time series is compared and upgrade vertical time series according to comparison result; Step S6 ': repeating step S4 ' and step S5 ', thus finish the structure and the renewal of pedestrian traffic data long-run development pattern.
As shown in Figure 1, be the process flow diagram of pedestrian traffic data combination multistep forecasting method of the present invention.The present invention is a kind of pedestrian traffic data combination multistep forecasting method, this method comprises: step S1: a group traveling together's traffic data secular trend development model is provided, it has a vertical time series, utilizes this vertical time series that the whole day data of the pedestrian traffic data on the same day are made multi-step prediction; Step S2: described pedestrian traffic data secular trend development model also has a horizontal time series, utilizes this horizontal time series that the short-term trend of the times data of the pedestrian traffic data on the same day are made multi-step prediction; Step S3: sampling pedestrian traffic real time data, these moment data of pedestrian traffic real time data and described secular trend multi-step prediction are compared, to obtain predicated error; Step S4: with described predicated error and error threshold comparison; If the predicated error of secular trend multi-step prediction data is then called predicting the outcome of secular trend multi-step prediction less than described error threshold; If predicated error, is then called these moment data of described short-term trend of the times multi-step prediction greater than described error threshold as predicting the outcome.
The present invention utilizes moving average model(MA model) to carry out the multi-step prediction of pedestrian traffic data, specifically suc as formula shown in (6-2).
In the formula:
---the vertical predicted value of pedestrian traffic data of current sampling time interval;
z
K(t)---K the historical data in pedestrian traffic data long-run development pattern front on the contained same day;
N---the contained historical data quantity of pedestrian traffic data long-run development pattern;
If the variation of pedestrian traffic data breaks away from the long-run development pattern, the prediction that then utilizes the long-run development pattern to be done certainly will produce bigger error.Therefore, in order predicting in conjunction with the data on the same day, to react the pedestrian traffic situation on the same day, thereby to improve the precision of prediction of pedestrian traffic data, the present invention proposes the short-term trend of the times multistep forecasting method of pedestrian traffic data.
Pedestrian traffic data short-term trend of the times multistep forecasting method mainly is divided into Forecasting Methodology and chooses and predict that step number determines two links.The present invention adopts moving average model(MA model), utilizes horizontal time series to carry out pedestrian traffic data short-term trend of the times multi-step prediction.The present invention adopts predicated error continuously less than the maximum sampling time interval quantity of threshold value, the prediction step number of determining as off-line.
The pedestrian traffic data time sequence of different periods has different behavioral characteristics, simultaneously, have the corresponding different prediction step number of pedestrian traffic data time sequence of Different Dynamic feature, therefore, there are certain corresponding relation in the behavioral characteristics of data time sequence and its prediction step number.
The present invention makes up undulatory property characteristic index, tendency characteristic index and the uncertain characteristic index of pedestrian traffic data time sequence, the behavioral characteristics that is used to quantize to express the data time sequence.The computing method of pedestrian traffic data fluctuations, tendency and uncertain characteristic exponent suc as formula (7-1), (7-2) and (7-3).
In the formula: A---the undulatory property characteristic exponent;
M---undulatory property characteristic exponent is calculated selected correlation time of interval quantity;
z
K '---the measured value of k ' individual correlation time of interval pedestrian's traffic data;
z
E---correlation time is the mean value of pedestrian's traffic data at interval;
δ---zoom factor.
In the formula: B---the tendency characteristic exponent;
z
K '+d---k '+d correlation time be the measured value of pedestrian's traffic data at interval;
D---slope calculates interval quantity correlation time;
η---zoom factor.
In the formula: C---uncertain characteristic exponent;
f
K '---correlation time is the blur level of pedestrian's traffic data at interval;
λ---zoom factor.
By the measured data analysis as can be known, all there is stronger correlationship in three kinds of data characteristics indexs with the prediction step number.
In view of artificial nerve network model merges advantage aspect the estimation at many-one, and model is through after the off-line training, arithmetic speed is very fast, can satisfy the pedestrian traffic Data Dynamic and analyze ageing requirement, and this paper selects for use the BP neural network model to carry out determining of on-line prediction step number.
The precision of pedestrian traffic data short-term trend of the times multi-step prediction is totally higher, but also has the bigger part of predicated error, and reason mainly is divided into two kinds of situations: the first, when the radix of pedestrian traffic data hour, the prediction relative error certainly will be bigger; The second, when data sequence sudden change amplitude was excessive, predicated error certainly will be bigger.
Because the pedestrian traffic data of every day both may show secular trend, also may show the short-term trend of the times.For this reason, the present invention proposes pedestrian traffic data combination multistep forecasting method, so that adapt to the variation of one day pedestrian traffic situation.
The above only is preferred embodiment of the present invention, non-limitation protection scope of the present invention, and the equivalent structure that all utilizations instructions of the present invention and accompanying drawing content are done changes, and all is contained in protection scope of the present invention.
Claims (7)
1. pedestrian traffic data combination multistep forecasting method is characterized in that this method comprises:
Step S1: a group traveling together's traffic data secular trend development model is provided, and it has a vertical time series, utilizes this vertical time series that the whole day data of the pedestrian traffic data on the same day are made multi-step prediction;
Step S2: described pedestrian traffic data secular trend development model also has a horizontal time series, utilizes this horizontal time series that the short-term trend of the times data of the pedestrian traffic data on the same day are made multi-step prediction;
Step S3: sampling pedestrian traffic real time data, these moment data of pedestrian traffic real time data and described secular trend multi-step prediction are compared, to obtain predicated error;
Step S4: with described predicated error and error threshold comparison; If the predicated error of secular trend multi-step prediction data is then called predicting the outcome of secular trend multi-step prediction less than described error threshold; If predicated error, is then called these moment data of described short-term trend of the times multi-step prediction greater than described error threshold as predicting the outcome.
2. pedestrian traffic data combination multistep forecasting method as claimed in claim 1 is characterized in that, adopts the vertical time series of moving average model(MA model) utilization that the pedestrian traffic data are carried out multi-step prediction among the described step S1, and formula is
In the formula:
The vertical predicted value of pedestrian traffic data for current sampling time interval; z
K(t) be K the historical data in pedestrian traffic data long-run development pattern front on the contained same day; N is the contained historical data quantity of pedestrian traffic data long-run development pattern.
3. pedestrian traffic data combination multistep forecasting method as claimed in claim 1 is characterized in that, pedestrian traffic data short-term trend of the times multistep forecasting method comprises among the described step S2: the S21 Forecasting Methodology is chosen with S22 prediction step number and is determined.
4. pedestrian traffic data combination multistep forecasting method as claimed in claim 3 is characterized in that, Forecasting Methodology is chosen and adopted the horizontal time series of moving average model(MA model) utilization to carry out pedestrian traffic data short-term trend of the times multi-step prediction among the described step S21.
5. pedestrian traffic data combination multistep forecasting method as claimed in claim 3 is characterized in that, adopts predicated error continuously less than the maximum sampling time interval quantity of error threshold among the described step S22, the prediction step number of determining as off-line.
6. pedestrian traffic data combination multistep forecasting method as claimed in claim 3, it is characterized in that, the index of behavioral characteristics by obtaining the data time sequence among the described step S22, comprise undulatory property characteristic exponent, tendency characteristic exponent and uncertain characteristic exponent, calculate online definite prediction step number with neural network model again.
7. pedestrian traffic data combination multistep forecasting method as claimed in claim 6 is characterized in that, the computing method of described undulatory property characteristic exponent, tendency characteristic exponent and uncertain characteristic exponent suc as formula
A is the undulatory property characteristic exponent in the formula, and M calculates selected correlation time of interval quantity, z for the undulatory property characteristic exponent
K 'Be the measured value of k ' individual correlation time of interval pedestrian's traffic data, z
EBe the mean value of interval pedestrian's traffic data correlation time, δ is a zoom factor;
B is the tendency characteristic exponent in the formula, z
K '+dIt is k '+d the correlation time of the measured value of pedestrian's traffic data at interval; D is that slope calculates interval quantity correlation time, and η is a zoom factor;
C is uncertain characteristic exponent in the formula, f
K 'Be the blur level of interval pedestrian's traffic data correlation time, λ is a zoom factor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011101093777A CN102156822A (en) | 2011-04-28 | 2011-04-28 | Pedestrian traffic data assembly multi-step forecasting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011101093777A CN102156822A (en) | 2011-04-28 | 2011-04-28 | Pedestrian traffic data assembly multi-step forecasting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102156822A true CN102156822A (en) | 2011-08-17 |
Family
ID=44438316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011101093777A Pending CN102156822A (en) | 2011-04-28 | 2011-04-28 | Pedestrian traffic data assembly multi-step forecasting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102156822A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103840970A (en) * | 2014-01-24 | 2014-06-04 | 珠海多玩信息技术有限公司 | Method and device for obtaining running status of service |
| CN107480784A (en) * | 2017-06-28 | 2017-12-15 | 青岛科技大学 | A kind of mobile phone signaling data pedestrian traffic trajectory predictions method based on deep learning |
| CN110494902A (en) * | 2017-02-03 | 2019-11-22 | 西门子交通有限责任公司 | For managing the system, apparatus and method of the traffic in geographical location |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739814A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | SCATS coil data-based traffic state online quantitative evaluation and prediction method |
-
2011
- 2011-04-28 CN CN2011101093777A patent/CN102156822A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739814A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | SCATS coil data-based traffic state online quantitative evaluation and prediction method |
Non-Patent Citations (1)
| Title |
|---|
| 牛世峰等: "基于纵向时间序列的快速路交通事件检测算法", 《哈尔滨工业大学学报》, no. 02, 28 February 2011 (2011-02-28) * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103840970A (en) * | 2014-01-24 | 2014-06-04 | 珠海多玩信息技术有限公司 | Method and device for obtaining running status of service |
| CN103840970B (en) * | 2014-01-24 | 2017-09-15 | 珠海多玩信息技术有限公司 | A kind of method and device for obtaining service operation state |
| CN110494902A (en) * | 2017-02-03 | 2019-11-22 | 西门子交通有限责任公司 | For managing the system, apparatus and method of the traffic in geographical location |
| CN107480784A (en) * | 2017-06-28 | 2017-12-15 | 青岛科技大学 | A kind of mobile phone signaling data pedestrian traffic trajectory predictions method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103729550B (en) | Multiple-model integration Flood Forecasting Method based on propagation time cluster analysis | |
| CN104408913B (en) | A kind of traffic flow three parameter real-time predicting method considering temporal correlation | |
| CN104715292A (en) | City short-term water consumption prediction method based on least square support vector machine model | |
| CN105702029A (en) | Express way traffic state prediction method taking spatial-temporal correlation into account at different times | |
| CN103024762A (en) | Service feature based communication service forecasting method | |
| Merabtene et al. | Risk assessment for optimal drought management of an integrated water resources system using a genetic algorithm | |
| CN103092699A (en) | Cloud computing resource pre-distribution achievement method | |
| CN102568205A (en) | Traffic parameter short-time prediction method based on empirical mode decomposition and classification combination prediction in abnormal state | |
| CN103810532B (en) | The method of Optimizing City drainage system operation conditions | |
| CN105305426B (en) | Mapreduce two-step short-period load prediction method based on deviation control mechanism | |
| CN107895014B (en) | Time series bridge monitoring data analysis method based on MapReduce framework | |
| CN105184402A (en) | Personalized user short-term load forecasting algorithm based on decision-making tree | |
| CN102930155A (en) | Method and device for acquiring early-warming parameters of power demands | |
| CN105335796A (en) | System and method for predicting wind farm output power climbing event | |
| JP5949979B1 (en) | Information processing apparatus, information processing system, information processing method, and program | |
| CN104599500A (en) | Grey entropy analysis and Bayes fusion improvement based traffic flow prediction method | |
| CN103699947A (en) | Meta learning-based combined prediction method for time-varying nonlinear load of electrical power system | |
| CN105005822A (en) | Optimal step length and dynamic model selection based ultrahigh arch dam response prediction method | |
| CN114202253A (en) | Charging station load adjustable potential evaluation method and system, storage medium and server | |
| CN102142192B (en) | Method and system for forewarning pedestrian traffic congestion based on combined multi-step prediction | |
| CN102156822A (en) | Pedestrian traffic data assembly multi-step forecasting method | |
| CN114254833A (en) | Reservoir water level prediction and scheduling method based on multiple linear regression and meteorological data | |
| CN102912760B (en) | Lowest navigable water level generating method of downstream river of reservoir | |
| CN102130955B (en) | System and method for generating alternative service set of composite service based on collaborative filtering | |
| CN106875681A (en) | A kind of traffic congestion tendency Forecasting Methodology and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20110817 |