CN106530721A - Dynamic prediction method for flow value of intersection in each flow direction based on transfer matrix - Google Patents
Dynamic prediction method for flow value of intersection in each flow direction based on transfer matrix Download PDFInfo
- Publication number
- CN106530721A CN106530721A CN201611238843.0A CN201611238843A CN106530721A CN 106530721 A CN106530721 A CN 106530721A CN 201611238843 A CN201611238843 A CN 201611238843A CN 106530721 A CN106530721 A CN 106530721A
- Authority
- CN
- China
- Prior art keywords
- flow
- prediction
- state
- matrix
- ratio
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
Abstract
The invention discloses a dynamic prediction method for a flow value of an intersection in each flow direction based on a transfer matrix. The dynamic prediction method comprises the steps of 1) respectively acquiring the real-time flow direction proportion and the road section flow according to entrance lane geomagnetic data and road section microwave data, 2) determining an initial probability vector and a state transition matrix, 3) fusing a weighted moving average method to calibrate the initial probability vector and the state transmission prediction matrix, 4) predicting the flow proportion of the intersection, and 5) predicting the flow value of the intersection in each flow direction. The dynamic prediction method performs scientific and reasonable dynamic prediction on the flow value of the intersection in each flow direction by utilizing traffic data and being combined with mathematical methods such as the weighted moving average method and the state transition matrix, and improves the scientificity and the practicability of prediction.
Description
Technical field
The present invention relates to traffic control field, and in particular to a kind of crossing based on transfer matrix respectively flows to flow value and moves
State Forecasting Methodology.
Background technology
In the big data epoch, real-time anticipation is made to traffic noise prediction using multi-source data information efficiently and is particularly weighed
Will.Scientific forecasting is carried out to traffic circulation using existing traffic data, and then formulates rational management and control scheme, be to alleviate city
One effective measures of traffic congestion.
In recent years, with the continuous growth of vehicle guaranteeding organic quantity, throwing of each city of China to intelligent transportation management and control facility
Enter and also continue to increase, the growth of the traffic data in explosion type of magnanimity.Although now can be using valid data to link flow
It is predicted, but still lacks a kind of each method for flowing to flow value of predicting at the intersection so that the number at key node
Effectively do not bring into play according to value.
The flow value that respectively flows to of crossing is predicted, substantially using the earth magnetism of early stage period crossover mouth entrance driveway
Data and section microwave data, using mathematical methods such as the method for weighted moving average, state-transition matrixes, to lower a period of time in crossing
The left-hand rotation of section, straight trip, the volume of traffic turned right are predicted.As shown in Fig. 2 obtaining left in real time by the geomagnetic data of entrance driveway
Turn, straight trip, right-turn volume, and further process and flowed to ratio in real time, it is real by the microwave detector detection on section
When link flow, using the mathematical methods such as the method for weighted moving average, state-transition matrix set up based on seasonal effect in time series it is real-time
Forecast model, respectively flows to flow value by what historical data was input into that the model is obtained subsequent period.
The content of the invention
Object of the present invention is to provide a kind of crossing based on transfer matrix respectively flows to flow value dynamic prediction side
Method, using earth magnetism, microwave data, with reference to mathematical methods such as the method for weighted moving average, state-transition matrixes, to next in crossing
The left-hand rotation of period, straight trip, turn right the volume of traffic be predicted, improve crossing is respectively flowed to flow value prediction science and
Practicality.
The technical solution used in the present invention is：
The present invention is comprised the following steps：
A) obtained according to earth magnetism and microwave data respectively and flow to ratio and link flow in real time
A) obtained by entrance driveway geomagnetic data and flow to ratio in real time
Here six parameters are included：Left-hand rotation flow L (tn), keep straight on flow S (tn), right-hand rotation flow R (tn), flow direction of turning left ratio
Example kL(tn), straight trip flows to ratio kS(tn), right-hand rotation flows to ratio kR(tn)；
B) real-time section flow is obtained by section microwave data
Here include following three parameter：To section microwave detector apart from L, vehicle crosses this section of L apart from institute to earth magnetism
Average time t0, section microwave is in (tn-t0) the volume of traffic Q (t that detect of periodn-t0)。
B) determine probability vector sum state-transition matrix
A) determine probability vector
Here following several parameters are included:tnPeriod probability vector N (tn), left-hand rotation flows to ratio kL(tn), straight trip stream
To ratio kS(tn), right-hand rotation flows to ratio kR(tn)；
B) determine state-transition matrix
Here calculating parameter includes：Left turn state L, straight-going state S, right turn state R, in tnTo tn+1Period, state i
To the transition probability p of state jij(tn), branch value k of state i to state jij(tn), state i turning to all states (L, S, R)
Shifting value ki.(tn), state-transition matrix An。
C) diffusion-weighted moving average method demarcates probability predicted vector and state branch prediction matrix
It is for reaching the purpose for carrying out dynamic prediction using historical data, with reference to the method for weighted moving average, continuous with three
Probability vector demarcates the probability predicted vector of subsequent period for one group of data, shifts square with three continuous states
Battle array demarcates the state branch prediction matrix of subsequent period for one group of data, and step includes：
A) diffusion-weighted moving average method demarcates probability predicted vector
Here calculating parameter includes：tnThe probability predicted vector N ' (t of periodn), three weight coefficients α, β, γ；
B) diffusion-weighted moving average method demarcates state branch prediction matrix
Here calculating parameter includes：tnPeriod is to tn+1The state branch prediction matrix of periodThree weight coefficient α,
β、γ。
D) crossing flows to scale prediction
Here calculating parameter includes：tnThe flow direction prediction ratio vector N of period*(tn), left-hand rotation flows to prediction ratioStraight trip flows to prediction ratioRight-hand rotation flows to prediction ratio
E) crossing respectively flows to flow value prediction
Here include following three parameter：Left-hand rotation predicted flow rate L*(tn), predicted flow rate S of keeping straight on*(tn), pre- flow measurement of turning right
Amount R*(tn)；
Section microwave is in (tn-t0) volume of traffic that detects of period is multiplied by the t that prediction is obtainednPeriod crossover mouth flows to ratio,
The traffic prediction value can be turned left, keep straight on, turned right.
The present invention is respectively flowed in flow value dynamic prediction method in crossing, the geomagnetic data in foundation crossing inlet road,
Section microwave data, and the mathematical methods such as the method for weighted moving average, state-transition matrix are combined, to subsequent period in crossing
The volume of traffic turn left, kept straight on, turn right carries out dynamic prediction.
Beneficial effect of the present invention：
1) each flow direction stream of the present invention using the detection data of crossing earth magnetism and upstream section microwave to crossing inlet
Amount is predicted, and on the basis of tradition flows to scale prediction, increased the accuracy guarantee of total flow, improves volume forecasting
Degree of accuracy.
2) what the present invention turned left, and kept straight on, turning right flows to composition of proportions probability vector, and flat using weighting movement
Equal method, is one group of data demarcating the state branch prediction matrix of subsequent period with three continuous state-transition matrixes, according to
Time serieses progressively recursion, reaches the purpose of dynamic prediction with this.
3) present invention is multiplied by state with the probability predicted vector demarcated and turns when carrying out crossing and flowing to scale prediction
Move prediction matrix to obtain flowing to prediction ratio vector, three components flowed in prediction ratio vector are predicted crossing
That what is turned left, keeps straight on, turning right flows to ratio.
Description of the drawings
Fig. 1 is the inventive method flow chart；
Fig. 2 is the layout diagram of earth magnetism and microwave equipment in the present invention；
Fig. 3 is dynamic calibration probability predicted vector schematic diagram of the present invention；
Fig. 4 is dynamic calibration state branch prediction matrix schematic diagram of the present invention.
Specific embodiment
The invention will be further described with reference to the accompanying drawings and detailed description：
As shown in figure 1, a kind of crossing based on transfer matrix respectively flows to flow value dynamic prediction method, step includes：
Obtained by entrance driveway geomagnetic data and section microwave data respectively flow in real time ratio and link flow, determine probability to
Amount and state-transition matrix, diffusion-weighted moving average method demarcate probability predicted vector and state branch prediction matrix, friendship
Prong flows to scale prediction, crossing and respectively flows to flow value prediction.
A) ground magnetic machine is laid in crossing inlet road, lay microwave equipment, by entrance driveway geomagnetic data on section
Acquisition flows to ratio in real time, obtains real-time section flow by section microwave data, is that crossing respectively flows to flow value dynamic prediction
Basic data needed for obtaining：
A) obtained by entrance driveway geomagnetic data and flow to ratio in real time
Here six parameters are included：Left-hand rotation flow L (tn), keep straight on flow S (tn), right-hand rotation flow R (tn), flow direction of turning left ratio
Example kL(tn), straight trip flows to ratio kS(tn), right-hand rotation flows to ratio kR(tn)；
The ground magnetic machine laid by crossing inlet road, detects tnThe left-hand rotation flow L (t of periodn), straight trip flow S
(tn), right-hand rotation flow R (tn), obtain tnPeriod turn left the ratio that flows to be：
Straight trip the ratio that flows to be：
Turn right the ratio that flows to be：
B) real-time section flow is obtained by section microwave data
Here include following three parameter：To section microwave detector apart from L, vehicle crosses this section of L apart from institute to earth magnetism
Average time t0, section microwave is in (tn-t0) the volume of traffic Q (t that detect of periodn-t0)；
As shown in Fig. 2 ground magnetic machine is embedded at crossing inlet road, on distance ground, the upstream section of magnetic machine L is laid
Microwave detector, vehicle is through t average time0Ground magnetic machine is reached from microwave detector, thus in tnPeriod ground magnetic machine inspection
The vehicle for measuring is (tn-t0) the volume of traffic Q (t that detected by section microwave equipment of periodn-t0)。
B) determine probability vector sum state-transition matrix
A) determine probability vector
Here following several parameters are included:tnPeriod probability vector N (tn), left-hand rotation flows to ratio kL(tn), straight trip stream
To ratio kS(tn), right-hand rotation flows to ratio kR(tn)；
Probability vector flows to composition of proportions, t by what is turned left, keep straight on, turn rightnThe probability vector of period is：
N(tn)=(kL(tn),kS(tn),kR(tn))
B) determine state-transition matrix
Here calculating parameter includes：Left turn state L, straight-going state S, right turn state R, in tnTo tn+1Period, state i
To the transition probability p of state jij(tn), branch value k of state i to state jij(tn), state i turning to all states (L, S, R)
Shifting value ki.(tn), state-transition matrix An；
In tnTo tn+1Period, branch value from state i to state j divided by branch value from state i to all states (L, S, R)
Transition probability is obtained, because state i can be shifted to itself, therefore ki.(tn)=ki(tn), state i to the transition probability of state j is：
State-transition matrix is made up of state transition probability, tnTo tn+1The state-transition matrix of period is：
C) diffusion-weighted moving average method demarcates probability predicted vector and state branch prediction matrix
It is for reaching the purpose for carrying out dynamic prediction using historical data, with reference to the method for weighted moving average, continuous with three
Probability vector demarcates the probability predicted vector of subsequent period for one group of data, shifts square with three continuous states
Battle array demarcates the state branch prediction matrix of subsequent period for one group of data, and step includes：
A) diffusion-weighted moving average method demarcates probability predicted vector
Here calculating parameter includes：tnThe probability predicted vector N ' (t of periodn), three weight coefficients α, β, γ；
Using the method for weighted moving average, subsequent period is demarcated with three continuous probability vectors as one group of data
Probability predicted vector, reaches the purpose of dynamic prediction, t with thisnPeriod probability predicted vector is：
N′(tn)=α N (tn-3)+βN(tn-2)+γN(tn-1)
As shown in figure 3, t0Period probability vector is N (t0), t1Period probability vector is N (t1), t2At the beginning of period
Beginning probability vector is N (t2), t3Period probability vector is N (t3), then t3Period probability predicted vector is：
N′(t3)=α N (t0)+βN(t1)+γN(t2)
t4Period probability predicted vector is：
N′(t4)=α N (t1)+βN(t2)+γN(t3)
According to the method described above, dynamic calibration is progressively carried out to probability predicted vector；
B) diffusion-weighted moving average method demarcates state branch prediction matrix
Here calculating parameter includes：tnPeriod is to tn+1The state branch prediction matrix of periodThree weight coefficient α,
β、γ；
Using the method for weighted moving average, it is one group of data demarcating subsequent period with three continuous state-transition matrixes
State branch prediction matrix, reaches the purpose of dynamic prediction, t with thisnPeriod state branch prediction matrix is：
As shown in figure 4, t0Period state-transition matrix is A0, t1Period state-transition matrix is A1, t2Period state turns
Shifting matrix is A2, t3Period state-transition matrix is A3, then t3Period state branch prediction matrix is：
t4Period state branch prediction matrix is：
According to the method described above, dynamic calibration is progressively carried out to state branch prediction matrix.
D) crossing flows to scale prediction
Here calculating parameter includes：tnThe flow direction prediction ratio vector N of period*(tn), left-hand rotation flows to prediction ratioStraight trip flows to prediction ratioRight-hand rotation flows to prediction ratio
Probability predicted vector and state branch prediction matrix according to having demarcated is obtained and flows to prediction ratio vector,
tnThe flow direction of period predicts that ratio vector is：
Flow to the component in prediction ratio vectorThe t for as being predictednPeriod crossover mouth
That what is turned left, keeps straight on, turning right flows to ratio.
E) crossing respectively flows to flow value prediction
Here include following three parameter：Left-hand rotation predicted flow rate L*(tn), predicted flow rate S of keeping straight on*(tn), pre- flow measurement of turning right
Amount R*(tn)；
Using section microwave in (tn-t0) the volume of traffic Q (t that detect of periodn-t0) it is multiplied by tnThe flow direction prediction ratio of period
Vector, you can obtain tnPeriod crossover mouth respectively flows to traffic prediction value：
Q(tn-t0)*N*(tn)=(L*(tn),S*(tn),R*(tn))
L in above-mentioned formula*(tn)、S*(tn)、R*(tn) it is tnPeriod crossover mouth turns left, keeps straight on, the flow of right-hand rotation is pre-
Measured value.
It should be pointed out that for those skilled in the art, 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 be regarded as protection scope of the present invention.
Claims (3)
1. a kind of crossing based on transfer matrix respectively flows to flow value dynamic prediction method, it is characterised in that：Including following step
Suddenly：
A）Obtained according to earth magnetism and microwave data respectively and flow to ratio and link flow in real time；
B）Determine probability vector sum state-transition matrix；
C）Diffusion-weighted moving average method demarcates probability predicted vector and state branch prediction matrix；
D）Crossing flows to scale prediction；
E）Crossing respectively flows to flow value prediction；
Step A）Obtained according to earth magnetism and microwave data respectively and flow to ratio and link flow in real time, comprised the steps：
a）Obtained by entrance driveway geomagnetic data and flow to ratio in real time；
b）Real-time section flow is obtained by section microwave data；
Step B）Determine probability vector sum state-transition matrix, comprise the steps：
a）Determine probability vector；
b）Determine state-transition matrix；
Step C）Diffusion-weighted moving average method demarcates probability predicted vector and state branch prediction matrix, including such as
Lower step：
a）Diffusion-weighted moving average method demarcates probability predicted vector；
b）Diffusion-weighted moving average method demarcates state branch prediction matrix.
2. a kind of crossing based on transfer matrix according to claim 1 respectively flows to flow value dynamic prediction method, its
It is characterised by：Step C）The step of b）The middle utilization method of weighted moving average, with three continuous state-transition matrixes as one
Group data are demarcating the state branch prediction matrix of subsequent period, and according to time serieses progressively recursion, reached with this dynamic pre-
The purpose of survey.
3. a kind of crossing based on transfer matrix according to claim 1 respectively flows to flow value dynamic prediction method, its
It is characterised by：Step D）In when carrying out crossing and flowing to scale prediction, be multiplied by shape with the probability predicted vector demarcated
State branch prediction matrix obtains flowing to prediction ratio vector, and three components flowed in prediction ratio vector are predicted friendship
What prong turned left, and kept straight on, turning right flows to ratio.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611238843.0A CN106530721B (en) | 2016-12-28 | 2016-12-28 | A kind of intersection based on transfer matrix respectively flows to flow value dynamic prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611238843.0A CN106530721B (en) | 2016-12-28 | 2016-12-28 | A kind of intersection based on transfer matrix respectively flows to flow value dynamic prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106530721A true CN106530721A (en) | 2017-03-22 |
| CN106530721B CN106530721B (en) | 2017-08-29 |
Family
ID=58338200
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611238843.0A Active CN106530721B (en) | 2016-12-28 | 2016-12-28 | A kind of intersection based on transfer matrix respectively flows to flow value dynamic prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106530721B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109766642A (en) * | 2019-01-15 | 2019-05-17 | 电子科技大学 | One kind is from evolution traffic network topological modelling approach |
| CN109993980A (en) * | 2019-02-21 | 2019-07-09 | 平安科技（深圳）有限公司 | Traffic flow forecasting method, device, computer equipment and storage medium |
| CN110164127A (en) * | 2019-04-04 | 2019-08-23 | 中兴飞流信息科技有限公司 | A kind of method, apparatus and server of vehicle flowrate prediction |
| CN110245423A (en) * | 2019-06-14 | 2019-09-17 | 重庆大学 | Discharge relation analysis method between a kind of freeway toll station |
| CN110363984A (en) * | 2019-06-25 | 2019-10-22 | 讯飞智元信息科技有限公司 | Traffic flow forecasting method and equipment |
| CN111344757A (en) * | 2017-09-19 | 2020-06-26 | 大陆汽车系统公司 | Adaptive traffic control system and method for operating the same |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10051777A1 (en) * | 2000-10-19 | 2002-04-25 | Daimler Chrysler Ag | Method for producing a dynamic forecast of traffic conditions for a road system around individual vehicles allows each vehicle to make an autonomic, continuous time-cum-location record for a parameter of traffic conditions |
| CN102629418A (en) * | 2012-04-09 | 2012-08-08 | 浙江工业大学 | Fuzzy kalman filtering-based traffic flow parameter prediction method |
| 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 |
-
2016
- 2016-12-28 CN CN201611238843.0A patent/CN106530721B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10051777A1 (en) * | 2000-10-19 | 2002-04-25 | Daimler Chrysler Ag | Method for producing a dynamic forecast of traffic conditions for a road system around individual vehicles allows each vehicle to make an autonomic, continuous time-cum-location record for a parameter of traffic conditions |
| CN102629418A (en) * | 2012-04-09 | 2012-08-08 | 浙江工业大学 | Fuzzy kalman filtering-based traffic flow parameter prediction method |
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| 张伟: "城市交通流数据优化感知关键技术研究", 《CNKI中国博士学位论文全文数据库》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111344757A (en) * | 2017-09-19 | 2020-06-26 | 大陆汽车系统公司 | Adaptive traffic control system and method for operating the same |
| CN109766642A (en) * | 2019-01-15 | 2019-05-17 | 电子科技大学 | One kind is from evolution traffic network topological modelling approach |
| CN109993980A (en) * | 2019-02-21 | 2019-07-09 | 平安科技（深圳）有限公司 | Traffic flow forecasting method, device, computer equipment and storage medium |
| CN110164127A (en) * | 2019-04-04 | 2019-08-23 | 中兴飞流信息科技有限公司 | A kind of method, apparatus and server of vehicle flowrate prediction |
| CN110164127B (en) * | 2019-04-04 | 2021-06-25 | 中兴飞流信息科技有限公司 | Traffic flow prediction method and device and server |
| CN110245423A (en) * | 2019-06-14 | 2019-09-17 | 重庆大学 | Discharge relation analysis method between a kind of freeway toll station |
| CN110245423B (en) * | 2019-06-14 | 2023-01-31 | 重庆大学 | Method for analyzing flow relation between highway toll stations |
| CN110363984A (en) * | 2019-06-25 | 2019-10-22 | 讯飞智元信息科技有限公司 | Traffic flow forecasting method and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106530721B (en) | 2017-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106530721B (en) | A kind of intersection based on transfer matrix respectively flows to flow value dynamic prediction method | |
| CN107230351B (en) | A kind of Short-time Traffic Flow Forecasting Methods based on deep learning | |
| Liu et al. | Development of a VISSIM simulation model for U-turns at unsignalized intersections | |
| CN104866654B (en) | A kind of construction method of integrated urban dynamic traffic emulation platform | |
| CN104298841B (en) | A kind of Flood Forecasting Method and system based on historical data | |
| CN110223517A (en) | Short-term traffic flow forecast method based on temporal correlation | |
| CN103208033B (en) | Access passenger flow forecasting method for urban rail transit new line under network condition | |
| CN104281780B (en) | Linear resesvoir is detained and confluxes and net watershed (many sub basin) method of confluxing | |
| CN106997669A (en) | A kind of method of the judgement traffic congestion origin cause of formation of feature based importance | |
| CN107705556A (en) | A kind of traffic flow forecasting method combined based on SVMs and BP neural network | |
| CN107991873A (en) | A kind of metro stray current leakage class Forecasting Methodology based on convolutional neural networks and BP neural network | |
| CN108695902A (en) | A kind of Cascade Reservoirs ecology-mutual feedback regulation and control method of power generation dynamic | |
| CN101877169B (en) | Data fusion system and method for controlling balance of multi-intersection traffic flow of trunk road | |
| CN103761138A (en) | Parameter correction method for traffic simulation software | |
| CN108876021A (en) | A kind of Medium-and Long-Term Runoff Forecasting method and system | |
| CN105427004A (en) | Optimization arrangement method of rapid road network traffic sensing network | |
| CN107705558A (en) | A kind of Bayes OD matrix estimation methods of more data types | |
| CN108877219B (en) | Traffic flow diversion prediction method after left-forbidden intersection of city | |
| CN107862877A (en) | A kind of urban traffic signal fuzzy control method | |
| CN106875060A (en) | A kind of flood real-time correction method based on global algorithms of automatic optimization | |
| CN114970377B (en) | Method and system for field flood forecasting based on Xinanjiang and deep learning coupling model | |
| CN108629970A (en) | Intersection signal parameter optimization method based on the search of Monte Carlo tree | |
| CN107945510A (en) | A kind of section detecting method for considering transport need and road network operational efficiency | |
| CN109886461A (en) | A kind of Runoff Forecast method and device | |
| CN108734958A (en) | A kind of traffic speed prediction technique |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |