CN106600993B - A kind of crossing vehicle shunting amount prediction technique based on RFID data - Google Patents
A kind of crossing vehicle shunting amount prediction technique based on RFID data Download PDFInfo
- Publication number
- CN106600993B CN106600993B CN201710086397.4A CN201710086397A CN106600993B CN 106600993 B CN106600993 B CN 106600993B CN 201710086397 A CN201710086397 A CN 201710086397A CN 106600993 B CN106600993 B CN 106600993B
- Authority
- CN
- China
- Prior art keywords
- prediction
- data
- period
- vehicle
- crossing
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/07—Controlling traffic signals
- G08G1/08—Controlling traffic signals according to detected number or speed of vehicles
Abstract
The crossing vehicle shunting amount prediction technique based on RFID data that the present invention relates to a kind of, comprising: step 1 collects RFID data;Step 2, data prediction remove repeated data;Step 3 differentiates that calculating crossing turns to vehicle flowrate based on the Vehicular turn that time phase divides;Step 4 extracts feature construction feature set;Step 5 uses training dataset training prediction model;Step 6, using test data set as prediction model input prediction, it is exported.The present invention mainly according to the RFID data of vehicle, the prediction of crossing vehicle shunting amount is realized using machine learning algorithm, provides decision-making foundation for the intelligent control of intelligent traffic lamp.
Description
Technical field
The present invention relates to intelligent transportation field, a kind of crossing vehicle shunting amount prediction technique based on RFID data.
Background technique
Nowadays the livable property for judging a city is largely the traffic condition for relying on city, the friendship of city bad luck
Lead to and brings great inconvenience to the trip of people.How more intelligent control and plan urban transportation becomes grinding for today's society
Study carefully hot spot, therefore intelligent transportation is come into being.The present invention passes through mould based on vehicle RFID data by machine learning algorithm
Crossing vehicle shunting amount size is predicted in type training, it is therefore an objective to realize that the intelligent control of traffic lights provides decision-making foundation.Mesh
Before the article published or the main data source of patent be GPS data, magnetic induction coil data and image data, it is right
RFID data research is less.In addition the volume forecasting on the single direction of crossing is mostly rested on to the prediction of intersection vehicle flux, it is right
The shunt volume prediction few people of each crossing all directions are related to.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of crossing vehicle shunting amount prediction side based on RFID data
Method predicts vehicle shunt volume size in all directions of crossing with realizing, final to mention for the intelligent control of traffic lights
For decision-making foundation.
The purpose of the present invention is achieved through the following technical solutions, a kind of crossing vehicle shunting based on RFID data
Measure prediction technique, key step are as follows:
Step 1 collects vehicle RFID data;
Step 2, data prediction, removal repeats and singular data;
Step 3, based on pretreated data, judge crossing Vehicular turn based on time phase division methods, and
Calculate all directions vehicle flowrate size in crossing in each signal period;
Step 4, all directions vehicle flowrate size being calculated according to step 3 are in conjunction with publicly-owned data construction feature
Collection, every characteristic the last one field that this feature integrates is as tag field, the vehicle of each steering under mark different time sections
Uninterrupted;
Step 5, using step 4 the data obtained as training set, using machine learning algorithm training one prediction crossing it is each
The prediction model of direction vehicle flowrate size;
Step 6, the signal period new for one, according to step 4 construction feature collection and using the feature set newly constructed as
The input of prediction model, prediction model output are the vehicle shunting amount size of corresponding direction in the prediction signal period.
Further, Vehicular turn recognition methods specific steps are as follows:
Step a, provide that each signal period is T, the signal period starts timing for the green light on some direction of signal lamp and arrives
Next time green light start between time interval;
Step b, regulation traffic lights duration within each signal period can be divided into multiple phases, be expressed as p1, p2 ...
pn;
Step c, each crossing divides East, West, South, North by direction, has the collection point RFID to use respectively in each direction
RE、RW、RS、RNIt indicates, it may determine that by the information of vehicles that the collection point RFID each in specific phase time section is got
It turns to;
Step d, Vehicular turn is identified according to step c, and big by the vehicle flowrate in each signal period T statistics all directions
It is small.
Further, in step 4, the building process of feature set are as follows:
Step 4.1 fully considers influence of the historical data to predetermined period, f from time seriest+1Indicate prediction week
Phase t+1 bogie car uninterrupted considers first five signal period vehicle flowrate size of predetermined period: ft、ft-1、ft-2、ft-3、
ft-4;The relevance between prediction result and the previous day same time period is considered, before choosing the previous day predetermined period and predetermined period
Five period vehicle flowrate sizes:When considering that prediction result is identical as before one week
Between relevance between section, predetermined period and predetermined period first five period vehicle flowrate size before choosing one week:
Step 4.2 considers the influence structure of time factor, weather conditions, working day factor and Temperature Factor to prediction result
Feature time, min_temp, max_temp, weather, weekday, month are built, predetermined period hourage, pre- is respectively represented
Survey daily minimal tcmperature, prediction daily maximum temperature, prediction day weather, prediction work day or weekend and prediction month day;
Then step 4.3, the feature in conjunction with selected by step 4.1 and step 4.2 add the final feature set packet of direction signs
Containing 24 features.
Due to using the technology described above, the invention has the following advantages that
1. using the RFID data of vehicle, acquisition of information is accurately simple.
2. accurately predicting that crossing turns to the size of vehicle flowrate by advanced data mining technology.
Detailed description of the invention
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
The detailed description of one step, in which:
Fig. 1 is crossing vehicle shunting amount prediction technique overall flow figure.
Specific embodiment
Below in conjunction with attached drawing, a preferred embodiment of the present invention will be described in detail.
Fig. 1 is the present invention about crossing vehicle shunting amount prediction technique overall flow figure.The described method includes:
Step 1 collects vehicle RFID data, and data mainly include that the collection point RFID number, vehicle unique identification, RFID are adopted
Set point location information and RFID data acquisition time stamp.
Step 2, RFID data cleaning, removal repeats and singular data.By the number of same vehicle under each collection point RFID
According to sorting in temporal sequence, a lesser time threshold is set, deletes and repeats when adjacent data time interval is less than threshold value
Data retain the first data.Unusual operation is gone to remove obvious abnormal data.
Step 3, based on pretreated data, judge crossing Vehicular turn based on time phase division methods, and
Calculate all directions vehicle flowrate size in crossing in each signal period.This method specifically calculates step are as follows:
A, provide that each signal period is T, i.e. green light on some direction of signal lamp starts timing and opens to green light next time
Time interval between beginning.
B, regulation traffic lights duration within each signal period can be divided into multiple phases, be expressed as p1, p2 ... pn (one
As n=2,4,8).One of the most common is four phases (n=4) division, and if p1 is that thing keeps straight on plus turns right, p2 is north and south straight trip
Add right-hand rotation, p3 turns left for thing, and p4 turns left for north and south.The transit time of each phase is not overlapped.
C, each crossing divides (usually four direction) east (E), western (W), southern (S), northern (N) by direction.In each direction
There is the collection point RFID to use R respectivelyE、RW、RS、RNIt indicates, because the transit time between multiple phases is not overlapped,
The information of vehicles that each collection point RFID is got in specific phase time section may determine that its steering.Such as it is provided in p1 phase
Thing straight trip adds right-hand rotation direction vehicle pass-through, then the R in the transit time section of p1 phaseEThe information of vehicles of acquisition be from west to
The vehicle of east straight trip, is indicated with W-E.For another example p4 phase provides North and South direction left-hand rotation direction vehicle pass-through, then in p4 phase
R in transit time sectionEThe information of vehicles of acquisition is the vehicle by north orientation east, is indicated with N-E.
D, Vehicular turn is identified according to step c, and by the vehicle flowrate size in each signal period T statistics all directions.
Step 4, all directions vehicle flowrate being calculated according to step 3 are in conjunction with publicly-owned data construction feature collection, often
The last one field of characteristic is tag field, identifies the vehicle flowrate size of each steering under different time sections.Publicly-owned number
According to mainly include weather, temperature, the date, whether weekend etc..
The building process of feature set are as follows:
A, influence of the historical data to predetermined period is fully considered from time series.ft+1Indicate predetermined period t+1
Bogie car uninterrupted considers first five signal period vehicle flowrate size of predetermined period: ft、ft-1、ft-2、ft-3、ft-4.Consider pre-
The relevance surveyed between result and the previous day same time period chooses the previous day predetermined period and predetermined period first five period vehicle
Uninterrupted:Consider prediction result and before one week between same time period
Predetermined period and predetermined period first five period vehicle flowrate size before relevance is chosen one week:
B, consider the influence construction feature of time factor, weather conditions, working day factor and Temperature Factor to prediction result
time,min_temp,max_temp,weather,weekday,month.Respectively represent predetermined period hourage, prediction day most
Low temperature, prediction daily maximum temperature, prediction day weather, prediction work day or weekend and prediction month day.
C, the feature in conjunction with selected by step a and step b then, in addition the final feature set of direction signs include 24 spies
Sign.
Step 5, using step 4 the data obtained as training set, key data source be vehicle RFID and can disclose obtain
The data such as weather, temperature and date.Using machine learning algorithm training, one can predict that crossing all directions vehicle flowrate is big
Small prediction model.
Step 6, the signal period new for one can according to step 4 construction feature and the feature set that will newly construct as
The input of prediction model, prediction model output are the vehicle shunting amount size of corresponding direction in the prediction signal period.
A kind of crossing vehicle shunting amount prediction technique based on RFID data provided by the invention mainly utilizes crossing vehicle
RFID data then realized in conjunction with corresponding prediction model to new by the processes such as building of data prediction, feature set
The prediction of crossing vehicle shunting amount in traffic signal cycles.Purpose is to provide decision-making foundation for the intelligent control of traffic lights.
Finally, it is stated that preferred embodiment above is only used to illustrate the technical scheme of the present invention and not to limit it, although logical
It crosses above preferred embodiment the present invention is described in detail, however, those skilled in the art should understand that, can be
Various changes are made to it in form and in details, without departing from claims of the present invention limited range.
Claims (2)
1. a kind of crossing vehicle shunting amount prediction technique based on RFID data, it is characterised in that: key step are as follows:
Step 1 collects vehicle RFID data;
Step 2, data prediction, removal repeats and singular data;
Step 3, based on pretreated data, judge crossing Vehicular turn based on time phase division methods, and calculate
All directions vehicle flowrate size in crossing in each signal period;
Step 4, all directions vehicle flowrate size being calculated according to step 3, should in conjunction with publicly-owned data construction feature collection
Feature set includes 24 features, each feature is indicated with a field, and wherein the last one field is tag field, and mark is not
With the vehicle flowrate size of steering each under the period, the publicly-owned data include weather, temperature, the date and whether weekend;
Step 5, using step 4 the data obtained as training set, use machine learning algorithm training one prediction crossing all directions
The prediction model of vehicle flowrate size;
Step 6, the signal period i.e. next cycle in current demand signal period predicted for needs, use the method structure of step 4
A characteristic is built, but this characteristic does not include the field of the last one mark vehicle uninterrupted, then uses this
The input for the prediction model that characteristic is trained as step 5, prediction model output are corresponding in the prediction signal period
The vehicle shunting amount size in direction;
In step 4, the building process of feature set are as follows:
Step 4.1 fully considers influence of the historical data to predetermined period, f from time seriest+1Indicate predetermined period t+1
Certain bogie car uninterrupted considers first five signal period vehicle flowrate size of predetermined period: ft、ft-1、ft-2、ft-3、ft-4;Consider
Relevance between prediction result and the previous day same time period chooses the previous day predetermined period and first five period of predetermined period
Vehicle flowrate size:Consider prediction result and before one week between same time period
Relevance, predetermined period and predetermined period first five period vehicle flowrate size before choosing one week:
Step 4.2 considers that the influence of time factor, weather conditions, working day factor and Temperature Factor to prediction result constructs spy
Time, min_temp, max_temp, weather, weekday, month are levied, predetermined period hourage, prediction day are respectively represented
The lowest temperature, prediction daily maximum temperature, prediction day weather, prediction work day or weekend and prediction month day;
Then it includes 24 that step 4.3, the feature in conjunction with selected by step 4.1 and step 4.2 add the final feature set of direction signs
Class field.
2. a kind of crossing vehicle shunting amount prediction technique based on RFID data according to claim 1, it is characterised in that:
Vehicular turn recognition methods specific steps are as follows:
Step a, provide that each signal period is T, the signal period is that green light on some direction of signal lamp starts timing to next
Secondary green light start between time interval;
Step b, regulation traffic lights duration within each signal period can be divided into multiple phases, be expressed as p1, p2 ... pn;Step
Rapid c, each crossing divide East, West, South, North by direction, have the collection point RFID to use R respectively in each directionE、RW、RS、RN
It indicates, its steering is judged by the information of vehicles that the collection point RFID each in specific phase time section is got;
Step d, Vehicular turn is identified according to step c, and by the vehicle flowrate size in each signal period T statistics all directions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710086397.4A CN106600993B (en) | 2017-02-17 | 2017-02-17 | A kind of crossing vehicle shunting amount prediction technique based on RFID data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710086397.4A CN106600993B (en) | 2017-02-17 | 2017-02-17 | A kind of crossing vehicle shunting amount prediction technique based on RFID data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106600993A CN106600993A (en) | 2017-04-26 |
CN106600993B true CN106600993B (en) | 2019-10-01 |
Family
ID=58587664
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710086397.4A Active CN106600993B (en) | 2017-02-17 | 2017-02-17 | A kind of crossing vehicle shunting amount prediction technique based on RFID data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106600993B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107730890B (en) * | 2017-11-09 | 2021-04-20 | 一石数字技术成都有限公司 | Intelligent transportation method based on traffic flow speed prediction in real-time scene |
CN108205890B (en) * | 2017-12-29 | 2021-03-09 | 迈锐数据(北京)有限公司 | Traffic data processing method and device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2351453A1 (en) * | 1976-05-11 | 1977-12-09 | Thomson Csf | Accelerated time traffic flow simulator - uses several element shift registers to represent sections of road network controlled by traffic lights |
CN101833867A (en) * | 2010-04-22 | 2010-09-15 | 赵乎 | Traffic management system based on radio frequency identification technology and method thereof |
CN102637365A (en) * | 2011-02-15 | 2012-08-15 | 成都西谷曙光数字技术有限公司 | System and method for realizing urban traffic intelligence by utilizing i-RFID (interactive-radio frequency identification) |
CN103310057A (en) * | 2013-06-14 | 2013-09-18 | 广州市公共交通数据管理中心 | Microscopic traffic simulation running method and device |
CN103927891A (en) * | 2014-04-29 | 2014-07-16 | 北京建筑大学 | Crossroad dynamic turning proportion two-step prediction method based on double Bayes |
CN105046956A (en) * | 2015-06-24 | 2015-11-11 | 银江股份有限公司 | Traffic flow simulating and predicting method based on turning probability |
CN106023621A (en) * | 2016-07-22 | 2016-10-12 | 山东交通学院 | Urban intelligent traffic guidance method based on RFID and WeChat platform |
CN106205150A (en) * | 2016-07-20 | 2016-12-07 | 安徽建筑大学 | A kind of car networking road condition monitoring system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9978270B2 (en) * | 2014-07-28 | 2018-05-22 | Econolite Group, Inc. | Self-configuring traffic signal controller |
-
2017
- 2017-02-17 CN CN201710086397.4A patent/CN106600993B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2351453A1 (en) * | 1976-05-11 | 1977-12-09 | Thomson Csf | Accelerated time traffic flow simulator - uses several element shift registers to represent sections of road network controlled by traffic lights |
CN101833867A (en) * | 2010-04-22 | 2010-09-15 | 赵乎 | Traffic management system based on radio frequency identification technology and method thereof |
CN102637365A (en) * | 2011-02-15 | 2012-08-15 | 成都西谷曙光数字技术有限公司 | System and method for realizing urban traffic intelligence by utilizing i-RFID (interactive-radio frequency identification) |
CN103310057A (en) * | 2013-06-14 | 2013-09-18 | 广州市公共交通数据管理中心 | Microscopic traffic simulation running method and device |
CN103927891A (en) * | 2014-04-29 | 2014-07-16 | 北京建筑大学 | Crossroad dynamic turning proportion two-step prediction method based on double Bayes |
CN105046956A (en) * | 2015-06-24 | 2015-11-11 | 银江股份有限公司 | Traffic flow simulating and predicting method based on turning probability |
CN106205150A (en) * | 2016-07-20 | 2016-12-07 | 安徽建筑大学 | A kind of car networking road condition monitoring system |
CN106023621A (en) * | 2016-07-22 | 2016-10-12 | 山东交通学院 | Urban intelligent traffic guidance method based on RFID and WeChat platform |
Non-Patent Citations (2)
Title |
---|
基于BP神经网络的路口短时交通流量预测方法;尚宁 等;《计算机应用与软件》;20060228;第23卷(第2期);正文第32-34页 * |
集成BP神经网络预测模型的研究与应用;赵会敏 等;《电信科学》;20160228(第2期);正文第60-67页 * |
Also Published As
Publication number | Publication date |
---|---|
CN106600993A (en) | 2017-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103325245B (en) | Method for predicting space-time traveling track of blacklisted vehicle | |
Mahler et al. | Reducing idling at red lights based on probabilistic prediction of traffic signal timings | |
Mandhare et al. | Intelligent road traffic control system for traffic congestion: a perspective | |
CN105118294B (en) | A kind of Short-time Traffic Flow Forecasting Methods based on state model | |
Li et al. | Studying the benefits of carpooling in an urban area using automatic vehicle identification data | |
CN104157139B (en) | A kind of traffic congestion Forecasting Methodology and method for visualizing | |
CN103838846B (en) | Emergency guiding method and emergency guiding system for individual on basis of big data | |
CN103198104B (en) | A kind of public transport station OD acquisition methods based on city intelligent public transit system | |
Huang et al. | A novel bus-dispatching model based on passenger flow and arrival time prediction | |
Yousef et al. | Intelligent traffic light scheduling technique using calendar-based history information | |
CN105489056B (en) | A kind of parking facilities' forecasting method based on OD matrixes | |
CN105551239A (en) | Travelling time prediction method and device | |
CN107369318A (en) | A kind of speed predicting method and device | |
CN104575038A (en) | Intersection signal control method considering priority of multiple buses | |
CN106017496A (en) | Real-time navigation method based on road condition | |
Wang et al. | Early warning of burst passenger flow in public transportation system | |
CN110491158A (en) | A kind of bus arrival time prediction technique and system based on multivariate data fusion | |
CN106600993B (en) | A kind of crossing vehicle shunting amount prediction technique based on RFID data | |
CN110118567A (en) | Trip mode recommended method and device | |
EP4060642A1 (en) | Method and system of predictive traffic flow and of traffic light control | |
Reddy et al. | Bus travel time prediction under high variability conditions | |
CN107195177A (en) | Based on Forecasting Methodology of the distributed memory Computational frame to city traffic road condition | |
CN111554085A (en) | Public transportation integrated travel intelligent service device and application method thereof | |
CN103268707A (en) | Signal regulating method for pedestrian crossing road section of bus prior passage | |
CN106558217B (en) | A kind of method, apparatus and server obtaining parking lay-by information |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |