CN106846834A - A kind of traffic control optimization method based on deep learning - Google Patents
A kind of traffic control optimization method based on deep learning Download PDFInfo
- Publication number
- CN106846834A CN106846834A CN201710044128.1A CN201710044128A CN106846834A CN 106846834 A CN106846834 A CN 106846834A CN 201710044128 A CN201710044128 A CN 201710044128A CN 106846834 A CN106846834 A CN 106846834A
- Authority
- CN
- China
- Prior art keywords
- traffic
- traffic control
- data
- intersection
- deep learning
- 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
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 invention discloses a kind of traffic control optimization method based on deep learning, belong to traffic control optimization field, the traffic control optimization method of deep learning should be based on, including:1)The extraction of the intersection traffic operational factor such as intersection capacity, delay, stop frequency, queue length;2)With reference to various intersection traffic operational factors, a System of Comprehensive Evaluation is set up;3)Using genetic algorithm, acquisition is optimal the traffic control strategy of comprehensive evaluation index;4)Using deep learning algorithm, constantly study adjustment is carried out to the traffic control optimisation strategy under different traffic, realize final urban traffic adaptive control.The present invention considers various traffic parameters, ensure that the reasonability of evaluation index, the use of genetic algorithm and deep learning simultaneously, solve the problems such as label training samples number is inadequate, algorithm the convergence speed is relatively slow and local optimum phenomenon occurs, it is ensured that the accuracy and reasonability of traffic control optimization.
Description
Technical field
The present invention relates to urban traffic control optimization method, especially a kind of traffic control optimization side based on deep learning
Method, belongs to traffic control optimization field.
Background technology
With the continuous propulsion of automobile process, automobile quantity is skyrocketed through.However, the growth rate of urban highway traffic
There is implacable contradiction with urban developing landuse scale, the problems such as bringing traffic congestion, traffic accident and take place frequently, to one
Society, the economic development in city are all great restrictive factors.
In recent years, with the continuous quickening of urbanization process, intelligent transportation system(ITS)Flourished,
As the current important solutions for alleviating Traffic Problems.In the research of ITS, intellectual traffic control is with induction
System is a heat subject, it is desirable to realize that it is excellent that urban traffic signal is controlled with inducible system using intellectual traffic control
Change, improve the traffic organization efficiency in city.
The operation of traffic is complicated and changeable, and the change road traffic system with space is also being continually changing over time.
Therefore road traffic system has an opening, randomness and dynamic, such as peak period on and off duty or traffic accident, heavy rain occurs
During severe snow, it is possible that road Severe blockage even results in the traffic system paralysis in whole city.Influence road traffic system
Series of factors be uncertain and with paroxysmal, therefore comprehensive multi-party complexity is needed in traffic optimization control process
Factor.However, in face of excessively complicated data, it is excellent that traditional traffic data processing method cannot match current traffic control
The demand of change.Therefore rational index system how is set up to seek traffic optimization control optimal solution be current to set about solution
Key certainly, so as to really realize representing for actual road conditions comprehensively.
In recent years, deep learning has come into the visual field of people, is model and magnanimity training data by building many hidden layers
To improve the performance of the unstructured datas such as treatment image, text, language, this has important meaning to treatment traffic base data
Justice.Therefore, how the big data learning characteristic in deep learning algorithm system to be combined with Intelligent traffic management systems, is to grind
Study carefully an important direction of traffic control optimization.
However, not there is a kind of effective traffic control optimization method based on deep learning, city can be really realized
The traffic control optimization of self study.
The content of the invention
For the defect that above-mentioned prior art is present, the present invention provides a kind of traffic control optimization side based on deep learning
Method, the method is comprised the following steps:1)The extraction of intersection traffic operational factor, obtains the friendship of each intersection of city in real time
Logical operational factor, including intersection capacity, delay, stop frequency, queue length;2)Comprehensive traffic postitallation evaluation index
Definition, using intersection capacity, delay, stop frequency, the queue length parameter that can obtain, sets up an overall merit
Index system;3)Traffic control optimization under current traffic condition, using genetic algorithm, acquisition is optimal comprehensive evaluation index
Traffic control strategy;4)The traffic control optimization of self study, using deep learning algorithm, to the traffic under different traffic
Control optimisation strategy carries out constantly study adjustment, realizes final urban traffic adaptive control.
Concrete technical scheme of the invention is as follows:
Step 1, the extraction of intersection traffic operational factor
Intersection traffic operational factor of the present invention includes intersection capacity, delay, stop frequency, length of queuing up
Degree, specific extracting mode is as follows:
Step 1.1, using the mass data sources such as city Floating Car gps data, video monitoring data and mobile phone signaling data, warp
Data anastomosing algorithm is crossed, the traffic basic parameters such as the magnitude of traffic flow, travel speed and the transit time of urban road are extracted;
Step 1.2, using city road network data, road infrastructure data, traffic signal timing data etc., realizes microcosmic traffic base
Plinth road digital;
Step 1.3, based on the digitlization microcosmic traffic basis road that step 1.2 is obtained, the true city obtained using step 1.1
Road traffic basic parameter, simulates real urban microscopic traffic circulation state;
Step 1.4, based on the urban microscopic traffic circulation state that step 1.3 is simulated, exports each cross-channel from operation result
Mouth is in current demand signal lamp with intersection capacity at present(C), delay(D), stop frequency(N), queue length(L)Data.
Step 2, the definition of comprehensive traffic postitallation evaluation index
Defined comprehensive traffic postitallation evaluation index of the invention(E)It is each the intersection energy for combining step 1 acquisition
Power(C), delay(D), stop frequency(S), queue length(L)Four indexs, specific calculation is as follows:
Step 2.1, a certain intersectioniComprehensive evaluation index, wherein,f()Represent that overall assessment refers to
Mark and the traffic capacity, delay, stop frequency, the relation function of queue length;
Step 2.2, obtains urban transportation postitallation evaluation index, i.e., after the overall assessment index of comprehensive whole intersection:
Wherein,nRepresent intersection total quantity.
Traffic control optimization under step 3, current traffic condition
The present invention utilizes genetic algorithm, and acquisition is optimal the traffic control strategy of comprehensive evaluation index, and specific practice is as follows:
Step 3.1, for current traffic condition under, random generation N kinds intersection control parameter combined strategy;
Step 3.2, N kind comprehensive traffic postitallation evaluation indexs are obtained by N kinds combined strategy using the method for step 2, while calculating
Each tactful weight(α), wherein,
Step 3.3, the weight size according to N kind strategies is reselected to the N kind strategies of step 3.1, and every kind of strategy is selected
The probability selected is the weight of step 3.2(α), generate newest N kind strategies;
Step 3.4, the N kinds strategy of step 3.3 is reconfigured according to certain probability, completes the variation of partial strategy, raw
Into another N kind strategies;
Step 3.5, brings the result of step 3.4 into step 3.2, carries out successive ignition;
Step 3.6, the similitude in iterations reaches predetermined value or N kind strategies reaches to a certain degree, completes to calculate, now
Intersection control parameter be combined as optimal traffic control result.
Step 4:The traffic control optimization of self study
The present invention utilizes deep learning algorithm, constantly study is carried out to the traffic control optimisation strategy under different traffic and is adjusted
It is whole, final urban traffic adaptive control is realized, specific practice is as follows:
Step 4.1, is trained using a large amount of historical datas, will history different traffic data bring step 3 into, obtain not
With the optimal traffic control result under traffic behavior, and stamp corresponding label;
Step 4.2, in actual trip, the label data that actual traffic state data data and step 4.1 are trained is carried out generally
Rate is matched, and obtains the joint probability distribution function of actual observation data and label dataf(p);
Step 4.3, the joint probability distribution obtained using step 4.2 completes the estimation of prior probability and posterior probability, and probability is estimated
Traffic control strategy high is haggled over for optimal policy;
Step 4.4, in traffic control optimization from now on, the result that step 4.3 is obtained treats as historical data, expansion step 4.1
In sample size, and repeat aforesaid operations, complete self study traffic control optimization.
The beneficial effects of the invention are as follows:This is based on the traffic control optimization method of deep learning, using a large amount of multi-source datas
The traffic circulation state in city is obtained, and the multiple parameters of urban intersection are obtained by way of urban microscopic traffic simulation and referred to
Mark, comprehensive multiple index obtains city overall operation evaluation index, using the index, by genetic algorithm and deep learning algorithm
Realize that urban traffic control optimizes.The method considers various traffic parameters, ensure that the reasonability of evaluation index, while losing
The use of propagation algorithm and deep learning solve label training samples number not enough, algorithm the convergence speed is relatively slow and occurs local
The problems such as optimal phenomenon, it is ensured that the accuracy and reasonability of traffic control optimization.
Brief description of the drawings
Fig. 1 is traffic control optimization method general flow chart of the present invention based on deep learning.
Fig. 2 is that intersection traffic operational factor of the present invention extracts flow chart.
Fig. 3 is comprehensive traffic postitallation evaluation index calculation flow chart of the present invention.
Fig. 4 is traffic control optimisation strategy formulation flow chart under current traffic condition of the present invention.
Fig. 5 is self study traffic control Optimizing Flow figure of the present invention.
Specific embodiment
Feature of the invention and other correlated characteristics are described in further detail below in conjunction with accompanying drawing:
As shown in figure 1, the present invention provides a kind of traffic control optimization method based on deep learning, the method includes following step
Suddenly:1)The extraction of intersection traffic operational factor, obtains the traffic circulation parameter of each intersection of city, including intersect in real time
The mouth traffic capacity, delay, stop frequency, queue length;2)The definition of traffic circulation evaluation index, using the intersection that can be obtained
The mouth traffic capacity, delay, stop frequency, queue length parameter, set up an assessment indicator system for totality;3)Current traffic shape
Traffic control optimization under state, using genetic algorithm, acquisition is optimal the traffic control strategy of comprehensive evaluation index;4)Learn by oneself
The traffic control optimization of habit, using deep learning algorithm, is carried out constantly to the traffic control optimisation strategy under different traffic
Study adjustment, realizes final urban traffic adaptive control.
As shown in Fig. 2 intersection traffic operational factor extract comprise the following steps that:
Step 1.1, using the mass data sources such as city Floating Car gps data, video monitoring data and mobile phone signaling data, warp
Data anastomosing algorithm is crossed, the traffic basic parameters such as the magnitude of traffic flow, travel speed and the transit time of urban road are extracted;
Step 1.2, using city road network data, road infrastructure data, traffic signal timing data etc., realizes microcosmic traffic base
Plinth road digital;
Step 1.3, based on the digitlization microcosmic traffic basis road that step 1.2 is obtained, the true city obtained using step 1.1
Road traffic basic parameter, simulates real urban microscopic traffic circulation state;
Step 1.4, based on the urban microscopic traffic circulation state that step 1.3 is simulated, exports each cross-channel from operation result
Mouth is in current demand signal lamp with intersection capacity at present(C), delay(D), stop frequency(N), queue length(L)Data.
As shown in figure 3, the calculation process of Metropolitan Integrative Traffic postitallation evaluation index is as follows:
Step 2.1, a certain intersectioniOverall assessment index, wherein,f()Represent that overall assessment refers to
Mark and the traffic capacity, delay, stop frequency, the relation function of queue length;
Step 2.2, obtains urban transportation postitallation evaluation index, i.e., after the overall assessment index of comprehensive whole intersection:
Wherein,nRepresent intersection total quantity.
As shown in figure 4, the formulation flow of the traffic control optimisation strategy under current traffic condition is as follows:
Step 3.1, for current traffic condition under, random generation N kinds intersection control parameter combined strategy;
Step 3.2, N kind traffic circulation evaluation indexes are obtained by N kinds combined strategy using the method for step 2, while calculating each
The weight of individual strategy(α), wherein,
Step 3.3, the weight size according to N kind strategies is reselected to the N kind strategies of step 3.1, and every kind of strategy is selected
The probability selected is the weight of step 3.2(α), generate newest N kind strategies;
Step 3.4, the N kinds strategy of step 3.3 is reconfigured according to certain probability, completes the variation of partial strategy, raw
Into another N kind strategies;
Step 3.5, brings the result of step 3.4 into step 3.2, carries out successive ignition;
Step 3.6, the similitude in iterations reaches predetermined value or N kind strategies reaches to a certain degree, completes to calculate, now
Intersection control parameter be combined as optimal traffic control result.
As shown in figure 5, the traffic control Optimizing Flow of city self study is as follows:
Step 4.1, is trained using a large amount of historical datas, will history different traffic data bring step 3 into, obtain not
With the optimal traffic control result under traffic behavior, and stamp corresponding label;
Step 4.2, in actual trip, the label data that actual traffic state data data and step 4.1 are trained is carried out generally
Rate is matched, and obtains the joint probability distribution of actual observation data and label dataf(p);
Step 4.3, the joint probability distribution obtained using step 4.2 completes the estimation of prior probability and posterior probability, and probability is estimated
Traffic control strategy high is haggled over for optimal policy;
Step 4.4, in traffic control optimization from now on, the result that step 4.3 is obtained treats as historical data, expansion step 4.1
In sample size, and repeat aforesaid operations, complete self study traffic control optimization.
Claims (5)
1. a kind of traffic control optimization method based on deep learning, it is characterised in that including:1)Intersection traffic operational factor
Extraction;2)The definition of comprehensive traffic postitallation evaluation index;3)The formulation of the traffic control optimisation strategy under current traffic condition;
4)The formulation of the traffic control optimisation strategy of system-wide net self study.
2. the traffic control optimization method based on deep learning according to claim 1, it is characterised in that described intersection
The logical operational factor of oral sex includes intersection capacity(C), delay(D), stop frequency(N), queue length(L), it is specific to extract
Method is as follows:
Step 1.1, using the mass data sources such as city Floating Car gps data, video monitoring data and mobile phone signaling data, warp
Data anastomosing algorithm is crossed, the traffic basic parameters such as the magnitude of traffic flow, travel speed and the transit time of urban road are extracted;
Step 1.2, using city road network data, road infrastructure data, traffic signal timing data etc., realizes microcosmic traffic base
Plinth road digital;
Step 1.3, based on the digitlization microcosmic traffic basis road that step 1.2 is obtained, the true city obtained using step 1.1
Road traffic basic parameter, simulates real urban microscopic traffic circulation state;
Step 1.4, based on the urban microscopic traffic circulation state that step 1.3 is simulated, exports each intersection from operation result
Mouth is in current demand signal lamp with intersection capacity at present(C), delay(D), stop frequency(N), queue length(L)Data.
3. the traffic control optimization method based on deep learning according to claim 1, it is characterised in that described synthesis
Traffic circulation evaluation index is the comprehensive crossover mouthful traffic capacity(C), delay(D), stop frequency(N), queue length(L)Four fingers
What mark was obtained, concrete mode is as follows:
Step 2.1, a certain intersectioniOverall assessment index, wherein,f()Represent overall assessment index
With the traffic capacity, delay, stop frequency, queue length relation function;
Step 2.2, obtains urban transportation postitallation evaluation index, i.e., after the overall assessment index of comprehensive whole intersection:
Wherein,nRepresent intersection total quantity.
4. the traffic control optimization method based on deep learning according to claim 1, it is characterised in that described traffic
The formulation of optimisation strategy is controlled to realize that specific method is as follows using genetic algorithm:
Step 3.1, for current traffic condition under, random generation N kinds intersection control parameter combined strategy;
Step 3.2, N kind traffic circulation evaluation indexes are obtained by N kinds combined strategy using the method for step 2, while calculating each
The weight of individual strategy(α), wherein,
Step 3.3, the weight size according to N kind strategies is reselected to the N kind strategies of step 3.1, and every kind of strategy is selected
The probability selected is the weight of step 3.2(α), generate newest N kind strategies;
Step 3.4, the N kinds strategy of step 3.3 is reconfigured according to certain probability, completes the variation of partial strategy, raw
Into another N kind strategies;
Step 3.5, brings the result of step 3.4 into step 3.2, carries out successive ignition;
Step 3.6, the similitude in iterations reaches predetermined value or N kind strategies reaches to a certain degree, completes to calculate, now
Intersection control parameter be combined as optimal traffic control result.
5. the traffic control optimization method based on deep learning according to claim 1, it is characterised in that described system-wide
The traffic control optimisation strategy of net self study is formulated to be realized using deep learning algorithm, and specific practice is as follows:
Step 4.1, is trained using a large amount of historical datas, will history different traffic data bring step 3 into, obtain not
With the optimal traffic control result under traffic behavior, and stamp corresponding label;
Step 4.2, in actual trip, the label data that actual traffic state data data and step 4.1 are trained is carried out generally
Rate is matched, and obtains the joint probability distribution of actual observation data and label dataf(p);
Step 4.3, the joint probability distribution obtained using step 4.2 completes the estimation of prior probability and posterior probability, and probability is estimated
Traffic control strategy high is haggled over for optimal policy;
Step 4.4, in traffic control optimization from now on, the result that step 4.3 is obtained treats as historical data, expansion step 4.1
In sample size, and repeat aforesaid operations, complete constantly study self adaptation traffic control optimization.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710044128.1A CN106846834A (en) | 2017-01-21 | 2017-01-21 | A kind of traffic control optimization method based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710044128.1A CN106846834A (en) | 2017-01-21 | 2017-01-21 | A kind of traffic control optimization method based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106846834A true CN106846834A (en) | 2017-06-13 |
Family
ID=59119851
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710044128.1A Pending CN106846834A (en) | 2017-01-21 | 2017-01-21 | A kind of traffic control optimization method based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106846834A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107331169A (en) * | 2017-09-01 | 2017-11-07 | 山东创飞客交通科技有限公司 | Urban road intersection signal time distributing conception evaluation method and system under saturation state |
CN108053661A (en) * | 2017-12-19 | 2018-05-18 | 青岛海信网络科技股份有限公司 | A kind of method and device of traffic control |
CN109272756A (en) * | 2018-11-07 | 2019-01-25 | 同济大学 | A kind of signal-control crossing queue length estimation method |
CN109284869A (en) * | 2018-10-08 | 2019-01-29 | 北方工业大学 | Urban intersection flow estimation method based on floating car data |
WO2019047905A1 (en) * | 2017-09-11 | 2019-03-14 | 阿里巴巴集团控股有限公司 | Road traffic analysis system, method and apparatus |
CN110114806A (en) * | 2018-02-28 | 2019-08-09 | 华为技术有限公司 | Signalized control method, relevant device and system |
CN110322687A (en) * | 2018-03-30 | 2019-10-11 | 杭州海康威视系统技术有限公司 | The method and apparatus for determining target intersection running state information |
CN110335478A (en) * | 2019-07-10 | 2019-10-15 | 江苏航天大为科技股份有限公司 | Across sub-district inter-linked controlling method based on deep learning |
CN110349416A (en) * | 2018-04-04 | 2019-10-18 | 百度(美国)有限责任公司 | The traffic light control system based on density for automatic driving vehicle (ADV) |
CN110491146A (en) * | 2019-08-21 | 2019-11-22 | 浙江工业大学 | A kind of traffic signal control scheme real-time recommendation method based on deep learning |
WO2020118517A1 (en) * | 2018-12-11 | 2020-06-18 | 深圳先进技术研究院 | Method for establishing and issuing evaluation indicator system for traffic management and control service indexes |
CN111311038A (en) * | 2018-12-11 | 2020-06-19 | 深圳先进技术研究院 | Evaluation method of traffic management and control service index |
CN111696348A (en) * | 2020-06-05 | 2020-09-22 | 南京云创大数据科技股份有限公司 | Multifunctional intelligent signal control system and method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5668717A (en) * | 1993-06-04 | 1997-09-16 | The Johns Hopkins University | Method and apparatus for model-free optimal signal timing for system-wide traffic control |
CN101789178A (en) * | 2009-01-22 | 2010-07-28 | 中国科学院自动化研究所 | Optimized control method for traffic signals at road junction |
CN104616498A (en) * | 2015-02-02 | 2015-05-13 | 同济大学 | Markov chain and neural network based traffic congestion state combined prediction method |
CN104657752A (en) * | 2015-03-17 | 2015-05-27 | 银江股份有限公司 | Deep learning-based safety belt wearing identification method |
CN106295874A (en) * | 2016-08-08 | 2017-01-04 | 上海交通大学 | Traffic flow parameter Forecasting Methodology based on deep belief network |
-
2017
- 2017-01-21 CN CN201710044128.1A patent/CN106846834A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5668717A (en) * | 1993-06-04 | 1997-09-16 | The Johns Hopkins University | Method and apparatus for model-free optimal signal timing for system-wide traffic control |
CN101789178A (en) * | 2009-01-22 | 2010-07-28 | 中国科学院自动化研究所 | Optimized control method for traffic signals at road junction |
CN104616498A (en) * | 2015-02-02 | 2015-05-13 | 同济大学 | Markov chain and neural network based traffic congestion state combined prediction method |
CN104657752A (en) * | 2015-03-17 | 2015-05-27 | 银江股份有限公司 | Deep learning-based safety belt wearing identification method |
CN106295874A (en) * | 2016-08-08 | 2017-01-04 | 上海交通大学 | Traffic flow parameter Forecasting Methodology based on deep belief network |
Non-Patent Citations (4)
Title |
---|
李云: "基于遗传算法的动态路径优化", 《中国优秀硕士学位论文全文数据库工程科技П辑》 * |
李振龙,等: "基于遗传算法的交叉口信号控制多目标优化", 《计算机运用》 * |
李苗苗,等: "基于vissim仿真的城市交叉口信号配时优化分析", 《洛阳理工学院学报(自然科学版)》 * |
陈姗: "城市道路平面交叉口优化设计与评价", 《中国优秀硕士学位论文全文数据库工程科技П辑》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107331169A (en) * | 2017-09-01 | 2017-11-07 | 山东创飞客交通科技有限公司 | Urban road intersection signal time distributing conception evaluation method and system under saturation state |
CN109493620B (en) * | 2017-09-11 | 2022-05-24 | 阿里巴巴集团控股有限公司 | Traffic road condition analysis system, method and device |
WO2019047905A1 (en) * | 2017-09-11 | 2019-03-14 | 阿里巴巴集团控股有限公司 | Road traffic analysis system, method and apparatus |
CN109493620A (en) * | 2017-09-11 | 2019-03-19 | 阿里巴巴集团控股有限公司 | A kind of traffic analysis system, method and device |
CN108053661A (en) * | 2017-12-19 | 2018-05-18 | 青岛海信网络科技股份有限公司 | A kind of method and device of traffic control |
WO2019165616A1 (en) * | 2018-02-28 | 2019-09-06 | 华为技术有限公司 | Signal light control method, related device, and system |
CN110114806A (en) * | 2018-02-28 | 2019-08-09 | 华为技术有限公司 | Signalized control method, relevant device and system |
CN110322687A (en) * | 2018-03-30 | 2019-10-11 | 杭州海康威视系统技术有限公司 | The method and apparatus for determining target intersection running state information |
CN110349416A (en) * | 2018-04-04 | 2019-10-18 | 百度(美国)有限责任公司 | The traffic light control system based on density for automatic driving vehicle (ADV) |
CN109284869A (en) * | 2018-10-08 | 2019-01-29 | 北方工业大学 | Urban intersection flow estimation method based on floating car data |
CN109284869B (en) * | 2018-10-08 | 2022-03-15 | 北方工业大学 | Urban intersection flow estimation method based on floating car data |
CN109272756A (en) * | 2018-11-07 | 2019-01-25 | 同济大学 | A kind of signal-control crossing queue length estimation method |
WO2020118517A1 (en) * | 2018-12-11 | 2020-06-18 | 深圳先进技术研究院 | Method for establishing and issuing evaluation indicator system for traffic management and control service indexes |
CN111311038A (en) * | 2018-12-11 | 2020-06-19 | 深圳先进技术研究院 | Evaluation method of traffic management and control service index |
CN110335478A (en) * | 2019-07-10 | 2019-10-15 | 江苏航天大为科技股份有限公司 | Across sub-district inter-linked controlling method based on deep learning |
CN110491146A (en) * | 2019-08-21 | 2019-11-22 | 浙江工业大学 | A kind of traffic signal control scheme real-time recommendation method based on deep learning |
CN111696348A (en) * | 2020-06-05 | 2020-09-22 | 南京云创大数据科技股份有限公司 | Multifunctional intelligent signal control system and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106846834A (en) | A kind of traffic control optimization method based on deep learning | |
Wang et al. | A BP neural network model optimized by mind evolutionary algorithm for predicting the ocean wave heights | |
CN110517482B (en) | Short-term traffic flow prediction method based on 3D convolutional neural network | |
Liao et al. | Urban sprawl scenario simulations based on cellular automata and ordered weighted averaging ecological constraints | |
Li et al. | Research on optimized GA-SVM vehicle speed prediction model based on driver-vehicle-road-traffic system | |
CN110570672B (en) | Regional traffic signal lamp control method based on graph neural network | |
Li et al. | Short-term power generation energy forecasting model for small hydropower stations using GA-SVM | |
CN102323970A (en) | Hydrological time series prediction method based on multiple-factor wavelet neural network model | |
CN105740643A (en) | Self-adaptive PM<2.5>concentration speculating method based on city region grid | |
CN112907970B (en) | Variable lane steering control method based on vehicle queuing length change rate | |
Xu et al. | Short-term traffic flow prediction model of wavelet neural network based on mind evolutionary algorithm | |
CN106845080B (en) | Based on the modified Scene Tourist meteorological disaster intelligent Forecasting of difference | |
CN104732300A (en) | Neural network wind power short-term forecasting method based on fuzzy partition theory | |
CN105678034A (en) | Pedestrian crossing time model and signalized intersection pedestrian crosswalk width optimizing method | |
CN104598765A (en) | Building energy consumption prediction method based on elastic adaptive neural network | |
CN111582541A (en) | Firefly algorithm-based inland inundation model prediction method | |
Lu et al. | Analyzing the evolution of rare events via social media data and k-means clustering algorithm | |
Li et al. | Scenario-based urban growth simulation by incorporating ecological-agricultural-urban suitability into a future land use simulation model | |
CN103699933A (en) | Traffic signal timing optimization method based on minimum spanning tree clustering genetic algorithm | |
CN112182951B (en) | Land utilization planning method considering multiple uncertainties | |
CN114120670A (en) | Method and system for traffic signal control | |
CN103268416B (en) | Wind speed prediction method based on STCP-BP | |
CN110021168B (en) | Grading decision method for realizing real-time intelligent traffic management under Internet of vehicles | |
Yu et al. | Minimize pressure difference traffic signal control based on deep reinforcement learning | |
CN116758767B (en) | Traffic signal lamp control method based on multi-strategy reinforcement learning |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170613 |