CN110503834A - The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN - Google Patents
The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN Download PDFInfo
- Publication number
- CN110503834A CN110503834A CN201910851015.1A CN201910851015A CN110503834A CN 110503834 A CN110503834 A CN 110503834A CN 201910851015 A CN201910851015 A CN 201910851015A CN 110503834 A CN110503834 A CN 110503834A
- Authority
- CN
- China
- Prior art keywords
- crossing
- traffic
- global
- multiple intersections
- data
- 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
- 238000000034 method Methods 0.000 title claims abstract description 13
- 238000012549 training Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 230000006399 behavior Effects 0.000 claims 1
- 238000004891 communication Methods 0.000 abstract description 7
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000003993 interaction Effects 0.000 abstract description 3
- 230000007246 mechanism Effects 0.000 abstract description 3
- 238000012217 deletion Methods 0.000 abstract description 2
- 230000037430 deletion Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- XNKARWLGLZGMGX-UHFFFAOYSA-N ethyl 4-(4-chloro-2-methylphenoxy)butanoate Chemical compound CCOC(=O)CCCOC1=CC=C(Cl)C=C1C XNKARWLGLZGMGX-UHFFFAOYSA-N 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G08G1/0125—Traffic data processing
-
- 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
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- 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
- G08G1/0125—Traffic data processing
- G08G1/0133—Traffic data processing for classifying traffic situation
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/07—Controlling traffic signals
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention provides a kind of intelligent traffic administration system decision-making techniques that Multiple Intersections collaboration is realized based on GAN.The present invention realizes the intelligent traffic administration system decision-making mechanism of Multiple Intersections collaboration, realizes the real-time traffic intelligent control of the Multiple Intersections under global information deletion condition.Under complicated Multiple Intersections traffic scene, the interaction of status information often has very big communication delay between crossing, and the accurate acquisition of global traffic information is also highly difficult.In order to reduce the communication load between crossing, realize the coordinated management between crossing, the present invention, which passes through, utilizes generation confrontation network technology, global traffic state information is generated according to respective local state information in each intersection, it is then based on the global traffic state information of generation, realizes the intelligent traffic administration system decision of Multiple Intersections collaboration.
Description
Technical field
The present invention relates to a kind of based on the intelligent traffic administration system decision-making mechanism for generating confrontation network implementations Multiple Intersections collaboration.
Background technique
With increasing for number of vehicles, urban transportation blocking will cause very huge economic loss and environmental pollution etc. and seriously ask
Topic.Wisdom traffic management system is the method for effectively reducing urban transportation blocking.Pass through modern communications information technology and optimization control
System is theoretical, carries out real-time traffic administration control according to the traffic environment information of acquisition, to improve driving experience degree, reduce and hand over
Logical congestion reduces traffic accident and reduces vehicle disposal of pollutants, is the important goal of wisdom traffic management.
In the changeable traffic scene of actual complex, in order to guarantee the real-time of traffic control, traditional intelligent transportation pipe
Reason is mainly carried out the traffic signals intelligent control of part by each traffic intersection according to the environmental information on periphery.In H.Wei,
G.Zheng,H.Yao,and Z.Li,"IntelliLight:A reinforcement learning approach for
intelligent traffic light control,"in Proc.ACM Conference on Knowledge
Discovery and Data Mining(KDD),London,United Kingdom,Aug.2018.;X.Liang,X.Du,
G.Wang,and Z.Han,“A deep reinforcement learning network for traffic light
cycle control,”IEEE Transactions on Vehicular Technology,vol.68,no.2,pp.1243-
Author utilizes deeply learning art in 1253, Jan.2019., is extracted, is passed through according to the state of intersection surrounding enviroment
The mode of study satisfy the need oral sex messenger carry out real-time control.Although above-mentioned depth learning technology can largely reduce traffic
The calculating dimension of control decision, but the cooperation between intersection difficult to realize, can not take into account the Global Optimality of network
It goes.For single intersection, in order to obtain global traffic environment information, the information generally required between intersection is handed over
Mutually.However, limitation of the interaction of the traffic environment information between intersection due to the communication resource, there is very high communication to prolong
When.Also, the acquisition of accurate complete traffic environment information is also highly difficult.
Summary of the invention
The purpose of the present invention is: the information exchange between crossing is reduced, while realizing the intelligent transportation pipe of Multiple Intersections collaboration
Reason.
In order to achieve the above object, Multiple Intersections collaboration is realized based on GAN the technical solution of the present invention is to provide a kind of
Intelligent traffic administration system decision-making technique, which comprises the following steps:
Step 1, using the historical traffic data at global K crossing, obtain generating pair by the training of stochastic gradient descent method
Anti- network G, K crossing overall situation traffic shape can be generated by the traffic state data at the single crossing of input by generating confrontation network G
State data, training objective are as follows:
In formula, D () indicates to differentiate network used in training process;G () indicates to generate confrontation network;X indicates true
The historical traffic data at K real crossing;Z indicates to generate the input data of confrontation network G, is single crossing in K crossing
Status information;Ex[] indicates the desired value to stochastic variable x;Ez[] indicates the desired value to stochastic variable z;
Step 2, the real-time traffic states data for obtaining each crossing in K crossing, with k-th of tunnel any in K crossing
Real-time traffic states data s of the mouth in t momentkIt (t) is input, the generation confrontation network G obtained according to step 1 training generates complete
The global traffic state data o at K crossing of officek(t);
The global traffic state data o that each crossing is generated according to step 2 in step 3, K crossingkIt (t) is Design of State
Global gain function, k-th of crossing are r in the global gain function of t momentk(t), deeply learning algorithm, study are utilized
The optimal traffic administration decision at each crossing is obtained, k-th of crossing is a in the optimal traffic administration decision of t momentk(t)。
Preferably, in step 3, the deeply learning algorithm is Deep Q-Network algorithm.
Preferably, in step 3, the global gain function rk(t) waiting time to reduce vehicle;The optimal traffic
Administrative decision akIt (t) is the conversion of control belisha beacon.
The invention proposes one kind based on confrontation network is generated, and realizes the intelligent traffic administration system decision machine of Multiple Intersections collaboration
System, realizes the real-time traffic intelligent control of the Multiple Intersections under global information deletion condition.In complicated Multiple Intersections traffic scene
Under, the interaction of status information often has very big communication delay, and the accurate acquisition of global traffic information between crossing
It is also highly difficult.In order to reduce the communication load between crossing, the coordinated management between crossing is realized, the present invention, which passes through, utilizes life
At confrontation network technology, global traffic state information is generated according to respective local state information in each intersection, so
Global traffic state information based on generation afterwards realizes the intelligent traffic administration system decision of Multiple Intersections collaboration.
Detailed description of the invention
Fig. 1 a is single crossing traffic schematic diagram of a scenario;
Fig. 1 b is nine intersection overall situation traffic scene schematic diagrames;
Fig. 2 a is single traffic state at road cross data;
Fig. 2 b makes a living into the Multiple Intersections overall situation traffic state data of network generation;
Fig. 3 is crossing decision process schematic diagram.
Specific embodiment
Present invention will be further explained below with reference to specific examples.It should be understood that these embodiments are merely to illustrate the present invention
Rather than it limits the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, those skilled in the art
Member can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Range.
The present invention proposes a kind of based on generation confrontation network, the intelligent traffic administration system decision-making mechanism of realization Multiple Intersections collaboration.
Each crossing is first with confrontation network technology is generated, and according to the local traffic status information at current crossing, generation obtains the overall situation
Multiple Intersections traffic state information.Then, the global traffic state information based on generation, it is real using deeply learning art
The traffic administration decision of existing Multiple Intersections collaboration.
Specifically, the present invention the following steps are included:
Step 1, using the historical traffic data at global K crossing, obtain generating pair by the training of stochastic gradient descent method
Anti- network G, K crossing overall situation traffic shape can be generated by the traffic state data at the single crossing of input by generating confrontation network G
State data, training objective are as follows:
In formula, D () indicates to differentiate network used in training process;G () indicates to generate confrontation network;X indicates true
The historical traffic data at K real crossing;Z indicates to generate the input data of confrontation network G, is single crossing in K crossing
Status information;Ex[] indicates the desired value to stochastic variable x;Ez[] indicates the desired value to stochastic variable z;
Step 2, the real-time traffic states data for obtaining each crossing in K crossing, with k-th of tunnel any in K crossing
Real-time traffic states data s of the mouth in t momentkIt (t) is input, the generation confrontation network G obtained according to step 1 training generates complete
The global traffic state data o at K crossing of officek(t);
The global traffic state data o that each crossing is generated according to step 2 in step 3, K crossingkIt (t) is Design of State
Global gain function, k-th of crossing are r in the global gain function of t momentk(t), such askIt (t) is when reducing the waiting of vehicle
Long, using deeply learning algorithm, such as Deep Q-Network (DQN) algorithm, study obtains the optimal traffic at each crossing
Administrative decision, k-th of crossing are a in the optimal traffic administration decision of t momentk(t), such askIt (t) is control belisha beacon
Conversion.
What Fig. 2 a was indicated is traffic state data figure corresponding to Fig. 1 a, specifically, white data point represents in Fig. 2 a
The location information of the intersection Fig. 1 a vehicle.Fig. 2 b is then the traffic state data of the multiple crossings overall situation generated, data pair
What is answered is the traffic state data of 9 intersection traffic scenes shown in Fig. 1 b.
As shown in Fig. 2, training obtains after generating confrontation network, in t moment, by the traffic shape for inputting current single crossing
State data sk(t), it as shown in Figure 2 a, exports to obtain the global traffic state data o of Multiple Intersections using generation networkk(t), as schemed
Shown in 2b.Single crossing is finally according to the global traffic state data o of generationk(t), real-time traffic administration decision a is madek
(t)。
As shown in figure 3, the local traffic status information s at current crossing is observed in each intersectionk(t), it is entered into
Trained generation network generates the traffic state information that network exports the Multiple Intersections overall situation generated, finally according to generation
Global traffic state information ok(t), study obtains the optimal traffic administration decision a of Multiple Intersections collaborationk(t)。
Claims (3)
1. a kind of intelligent traffic administration system decision-making technique for realizing Multiple Intersections collaboration based on GAN, which is characterized in that including following step
It is rapid:
Step 1, using the historical traffic data at global K crossing, obtain generating confrontation net by the training of stochastic gradient descent method
Network G, K crossing overall situation traffic behavior number can be generated by the traffic state data at the single crossing of input by generating confrontation network G
According to training objective are as follows:
In formula, D () indicates to differentiate network used in training process;G () indicates to generate confrontation network;X indicates true K
The historical traffic data at a crossing;Z indicates to generate the input data of confrontation network G, is the state letter at single crossing in K crossing
Breath;Ex[] indicates the desired value to stochastic variable x;Ez[] indicates the desired value to stochastic variable z;
Step 2, the real-time traffic states data for obtaining each crossing in K crossing, with k-th of crossing any in K crossing in t
The real-time traffic states data s at momentkIt (t) is input, the generation confrontation network G obtained according to step 1 training generates overall situation K
The global traffic state data o at crossingk(t);
The global traffic state data o that each crossing is generated according to step 2 in step 3, K crossingk(t) global for Design of State
Revenue function, k-th of crossing are r in the global gain function of t momentk(t), using deeply learning algorithm, study is obtained
The optimal traffic administration decision at each crossing, k-th of crossing are a in the optimal traffic administration decision of t momentk(t)。
2. a kind of intelligent traffic administration system decision-making technique for realizing Multiple Intersections collaboration based on GAN as described in claim 1, feature
It is, in step 3, the deeply learning algorithm is Deep Q-Network algorithm.
3. a kind of intelligent traffic administration system decision-making technique for realizing Multiple Intersections collaboration based on GAN as described in claim 1, feature
It is, in step 3, the global gain function rk(t) waiting time to reduce vehicle;The optimal traffic administration decision ak
It (t) is the conversion of control belisha beacon.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910851015.1A CN110503834A (en) | 2019-09-10 | 2019-09-10 | The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910851015.1A CN110503834A (en) | 2019-09-10 | 2019-09-10 | The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110503834A true CN110503834A (en) | 2019-11-26 |
Family
ID=68591560
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910851015.1A Pending CN110503834A (en) | 2019-09-10 | 2019-09-10 | The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110503834A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109214422A (en) * | 2018-08-02 | 2019-01-15 | 深圳先进技术研究院 | Parking data method for repairing and mending, device, equipment and storage medium based on DCGAN |
CN109215349A (en) * | 2018-10-26 | 2019-01-15 | 同济大学 | Traffic flow forecasting method when long based on deep learning |
CN109493599A (en) * | 2018-11-16 | 2019-03-19 | 南京航空航天大学 | A kind of Short-time Traffic Flow Forecasting Methods based on production confrontation network |
CN110021168A (en) * | 2019-04-18 | 2019-07-16 | 上海科技大学 | The stepped strategy method of real-time intelligent traffic management is realized under a kind of car networking |
CN110147709A (en) * | 2018-11-02 | 2019-08-20 | 腾讯科技(深圳)有限公司 | Training method, device, terminal and the storage medium of vehicle attribute model |
CN110164128A (en) * | 2019-04-23 | 2019-08-23 | 银江股份有限公司 | A kind of City-level intelligent transportation analogue system |
-
2019
- 2019-09-10 CN CN201910851015.1A patent/CN110503834A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109214422A (en) * | 2018-08-02 | 2019-01-15 | 深圳先进技术研究院 | Parking data method for repairing and mending, device, equipment and storage medium based on DCGAN |
CN109215349A (en) * | 2018-10-26 | 2019-01-15 | 同济大学 | Traffic flow forecasting method when long based on deep learning |
CN110147709A (en) * | 2018-11-02 | 2019-08-20 | 腾讯科技(深圳)有限公司 | Training method, device, terminal and the storage medium of vehicle attribute model |
CN109493599A (en) * | 2018-11-16 | 2019-03-19 | 南京航空航天大学 | A kind of Short-time Traffic Flow Forecasting Methods based on production confrontation network |
CN110021168A (en) * | 2019-04-18 | 2019-07-16 | 上海科技大学 | The stepped strategy method of real-time intelligent traffic management is realized under a kind of car networking |
CN110164128A (en) * | 2019-04-23 | 2019-08-23 | 银江股份有限公司 | A kind of City-level intelligent transportation analogue system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112700639B (en) | Intelligent traffic path planning method based on federal learning and digital twins | |
Aljohani et al. | An algorithm for accessing traffic database using wireless technologies | |
CN109612488B (en) | Big data micro-service-based mixed travel mode path planning system and method | |
CN104809895A (en) | Adjacent intersection arterial road coordinate control model and optimization method thereof | |
CN102867409A (en) | Road traffic cooperative control method for urban central area | |
CN107134137A (en) | A kind of Dynamic User-Optimal Route Choice method for considering real time information | |
CN109002923A (en) | A kind of intercity multimode travel route planing method | |
CN111009140B (en) | Intelligent traffic signal control method based on open-source road condition information | |
CN113053120A (en) | Traffic signal lamp scheduling method and system based on iterative learning model predictive control | |
CN113657645B (en) | Space-time connection enhanced 3DCNN traffic prediction method based on attention mechanism | |
CN113409576B (en) | Bayesian network-based traffic network dynamic prediction method and system | |
Wu et al. | Traffic signal networks control optimize with PSO algorithm | |
CN113392577A (en) | Regional boundary main intersection signal control method based on deep reinforcement learning | |
CN109584564A (en) | A kind of letter prosecutor case being applicable in more equipment implements optimization method and its system and device | |
CN110021168B (en) | Grading decision method for realizing real-time intelligent traffic management under Internet of vehicles | |
CN110503834A (en) | The intelligent traffic administration system decision-making technique of Multiple Intersections collaboration is realized based on GAN | |
CN107424407A (en) | A kind of vehicle dispatch system and vehicle scheduling scheme generation method | |
CN115063988B (en) | Cross-traffic subarea signal cooperative control method for emergency vehicle priority | |
CN115331460B (en) | Large-scale traffic signal control method and device based on deep reinforcement learning | |
CN107195192B (en) | A kind of recognition methods of vehicle-mounted orientation operating status | |
CN110097757A (en) | A kind of intersection group critical path recognition methods based on depth-first search | |
Zhang et al. | Traffic Guidance Based on LSTM Neural Network and Dual Tracking Dijkstra Algorithm | |
CN110164153A (en) | A kind of adaptive timing method of traffic signals | |
CN106092109A (en) | The road of the efficient pathfinding of a kind of electronic chart connects logically method and application | |
Shan et al. | A novel personalized dynamic route recommendation approach based on pearson similarity coefficient in cooperative vehicle-infrastructure systems |
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: 20191126 |
|
RJ01 | Rejection of invention patent application after publication |