CN111047884A - Traffic light control method based on fog calculation and reinforcement learning - Google Patents

Traffic light control method based on fog calculation and reinforcement learning Download PDF

Info

Publication number
CN111047884A
CN111047884A CN201911398203.XA CN201911398203A CN111047884A CN 111047884 A CN111047884 A CN 111047884A CN 201911398203 A CN201911398203 A CN 201911398203A CN 111047884 A CN111047884 A CN 111047884A
Authority
CN
China
Prior art keywords
intersection
traffic
time
vehicles
green light
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
Application number
CN201911398203.XA
Other languages
Chinese (zh)
Inventor
樊秀梅
安萌萌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian University of Technology
Original Assignee
Xian University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xian University of Technology filed Critical Xian University of Technology
Priority to CN201911398203.XA priority Critical patent/CN111047884A/en
Publication of CN111047884A publication Critical patent/CN111047884A/en
Pending legal-status Critical Current

Links

Images

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
    • 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
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G08G1/0141—Measuring and analyzing of parameters relative to traffic conditions for specific applications for traffic information dissemination
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/095—Traffic lights

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a traffic light control method based on fog calculation and reinforcement learning, which comprises the steps of firstly collecting traffic conditions of intersections by fog nodes of the intersections and broadcasting the traffic conditions to fog nodes of adjacent intersections, collecting the number of vehicles at the intersections and calculating the green light time required by the vehicles to pass through the intersections, calculating the green light time required by traffic flow of the adjacent intersections to pass through the intersections and the green light time required by traffic flow of the adjacent intersections to pass through the intersections, the crossing integrates the number of vehicles at the crossing and the adjacent crossings and calculates the total green light time required by the crossing, the Agent at the crossing applies the calculated green light time length to the traffic signal lamp, the feedback reward signal is observed, and the environmental state is transferred to the state at the next moment.

Description

Traffic light control method based on fog calculation and reinforcement learning
Technical Field
The invention belongs to the technical field of intelligent traffic, and particularly relates to a traffic light control method based on fog calculation and reinforcement learning.
Background
With the acceleration of urbanization construction and the improvement of the living standard of people, vehicles are more and more, so that the urban road is overloaded and the traffic jam is increasingly serious. Every year, all countries in the world cause huge economic losses due to traffic congestion, and thus the traffic congestion seriously restricts economic development. According to survey, traffic accidents caused by traffic jam still rise every year, and a series of energy problems and environmental problems are caused; traffic jam not only causes waste of transportation resources, but also greatly reduces transportation efficiency, excessively consumes social cost and seriously hinders development of cities, so that the traffic jam is urgently relieved.
The internet of vehicles aims to solve the new challenging requirements in the current traffic system field; the vehicles in the internet of vehicles can acquire state information of the vehicles and the environment through devices such as sensors, radio frequency identification technology, road side units and the like, and the information is transmitted, analyzed and processed by utilizing the internet and computer technology, so that intelligent traffic management is realized.
The cloud computing is a highly virtualized network platform, which consists of a large number of edge terminal nodes and routing equipment, provides services such as computing, storage and the like between a cloud and the network terminal equipment, and is essentially decentralized edge computing; the fog computing has the advantages of cognition, high efficiency and low time delay, and information sharing can be realized through a fog platform, so that data can be communicated.
Artificial intelligence can process a large amount of data under complex conditions and is concerned about; the reinforcement learning is developed on the basis of theories such as animal learning and parameter disturbance adaptive control and is one of the methods of machine learning; the learning mode of reinforcement learning is 'trial and error learning', and the intelligent agent evaluates the behavior through reward signals; the goal of the agent is to find the optimal strategy in each discrete state, with the expectation that the maximum discount reward sum will be obtained.
Disclosure of Invention
The invention aims to provide a traffic light control method based on fog calculation and reinforcement learning, and solves the problems that the existing traffic signal light control mode is poor in real-time performance and traffic information of each intersection cannot be shared in real time.
The invention adopts the technical scheme that a traffic light control method based on fog calculation and reinforcement learning is implemented according to the following steps:
step 1, collecting intersection traffic conditions by the intersection fog node and broadcasting the intersection traffic conditions to fog nodes of adjacent intersections;
step 2, the Agent at the intersection collects the number of vehicles at the intersection and calculates the green light time T required by the vehicles to pass through the intersectiong；
Step 3, the Agent at the intersection selects corresponding influence factor values according to the number of vehicles at the adjacent intersections, and calculates the green light time T required by the traffic flow at the adjacent intersections to pass through the intersectionneighbor；
Step 4, calculating the green light time T required by the traffic flow of the adjacent crossing to pass through the crossing by the crossingneighbor；
And 5: the intersection integrates the number of vehicles at the intersection and the adjacent intersections, and calculates the total green light time T required by the intersectiontotal；
Step 6: the Agent at the intersection applies the calculated green light duration to the traffic signal lamp, observes the feedback reward signal and transfers the environment state to the state of the next moment;
and 7: and the Agent at the intersection updates the Q table according to the updating rule.
The invention is also characterized in that:
wherein the step 1 of testing the traffic conditions at the intersection comprises the following steps:
the number of vehicles driving to the next intersection;
testing the time T of the upstream crossing of the crossing to release the traffic flowstart；
Testing the time T of the end of passing the traffic flow at the upstream intersection of the intersectionsend；
Time T for traffic flow of the intersection to reach the next intersectionarriveThe following formula (1):
Figure BDA0002346855820000031
wherein, | Stl1-Stl2I is the distance between two intersections, and v is the average travel speed of the traffic flow.
Wherein, the intersection calculates the green time T needed by the intersection according to the number of the vehicles in the step 2gThe following formula (2):
Tg＝d+2*NumVeh (2)
wherein d is a vehicle start delay time, and NumVeh is the number of vehicles;
wherein in step 2 when T isgWhen the traffic light time is not less than the maximum green light time, the green light time of the traffic light is set as the maximum green light time, and the step 6 is carried out; when T isgNo more than maximum green time, no vehicles at adjacent crossing, TgSetting the actual green time of the intersection, and turning to the step 6; when vehicles exist at the adjacent intersection, turning to the step 3;
the influence factor β value in the step 3 represents the influence degree, the bigger the β value is, the bigger the traffic pressure of the adjacent intersection to the intersection is, the corresponding β value is taken according to the number of vehicles at the adjacent intersection of the intersection, and the β value is used for adjusting the traffic flow of the intersection and the adjacent intersection;
wherein in step 5TtotalThe calculation formula (2) is shown in formula (3):
Ttotal＝Tg+Tarrive+Tneighbor(3)；
wherein in step 5 when T istotalWhen the traffic light time is not less than the maximum green light time, the green light time of the traffic light is set as the maximum green light time, and the step 6 is carried out; when T istotalNot more than the maximum green time, TtotalSetting the actual green time of the intersection, and turning to the step 6;
wherein, in step 7, the Q table is a matrix, and the updating rule is as follows:
Figure BDA0002346855820000041
where s is the status, a is the action, r is the return, α is the learning rate, γ is the discount factor, Qt(s, a) is the Q value of the state selection operation at time t;
and 7, judging whether the Q matrix is converged or not in the step 7, wherein the Q matrix is converged, the procedure is ended, and the step 2 is carried out after the Q matrix is not converged.
The invention has the beneficial effects that:
compared with the traditional traffic light control mode, the traffic light control method based on the fog calculation and the reinforcement learning has the advantages that the green light time can be distributed according to the number of the vehicles at the intersection in real time, the parking time and the parking times can be reduced, the road regulation efficiency is improved, and the traffic jam can be effectively relieved.
Drawings
FIG. 1 is a traffic light management system based on a fog platform for a traffic light control method based on fog calculation and reinforcement learning according to the present invention;
FIG. 2 is a schematic diagram of multiple intersections in a traffic light control method based on fog calculation and reinforcement learning according to the present invention;
FIG. 3 is a flow chart of a multi-intersection traffic light control algorithm in a traffic light control method based on fog calculation and reinforcement learning according to the present invention;
FIG. 4 is a graph of average vehicle throughput in a traffic light control method based on fog calculations and reinforcement learning in accordance with the present invention;
FIG. 5 is a graph of average vehicle delay time in a traffic light control method based on fog calculation and reinforcement learning in accordance with the present invention;
fig. 6 is a graph of average vehicle queue length in a traffic light control method based on fog calculation and reinforcement learning in accordance with the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
Fog computing is between cloud computing and personal computing, is a semi-virtualized and distributed service computing architecture model, is close to the ground, has short time for transmitting information and high computing speed compared with cloud computing, and meets the requirement of high real-time performance of traffic lights; reinforcement learning is one of machine learning, in the process of interacting with the environment, an intelligent body can realize a set target through a learning strategy, Traffic flow has no specific rule, a scheme with a certain rule cannot be adopted to control a Traffic light, and Reinforcement learning can interact with the environment in real time, and has better reliability, so that the invention combines the characteristics of fog calculation and Reinforcement learning to control the Traffic light, namely an FRTL (fog learning Traffic light) control mode, and comprises the following steps:
a traffic light control algorithm is formulated by adopting a Q learning algorithm in reinforcement learning, and the three elements of the Q learning algorithm are states, actions and rewards. The vehicle delay time is directly acquired from a delay time detector in the VISSIM. Positive rewards are taken, i.e. the smaller the delay time, the larger the prize value. Equation (5) is a traffic condition expression, and equation (6) is a reward function:
Figure BDA0002346855820000051
in the formula (5), the first and second groups,
Figure BDA0002346855820000052
is the state of the ith intersection at time t,
Figure BDA0002346855820000053
is the number of vehicles from intersection i to intersection j at time t;
Figure BDA0002346855820000061
the invention provides a traffic light control method based on fog calculation and reinforcement learning, which is implemented by the following steps:
step 1, as shown in fig. 1, a fog node deployed at each intersection collects traffic information of the intersection, and broadcasts the traffic information of the intersection to fog nodes of adjacent intersections through a fog computing platform for information sharing, wherein the broadcast information includes:
the number of vehicles driving to the next intersection;
time T for upstream intersection of road junction to begin releasing traffic flowstart；
Time T for ending releasing traffic flow at upstream crossing of road junctionend；
Time T for traffic flow of the intersection to reach the next intersectionarriveSee formula (1):
Figure BDA0002346855820000062
in the formula (1), | Stl1-Stl2L is the distance between two intersections, v is the average travel speed of the traffic flow;
step 2: the Agent at the intersection calculates the required green light time T according to the number of the vehicles per segSee formula (2):
Tg＝d+2*NumVeh (2)
in the formula (2), d is the vehicle start delay time, and NumVeh is the number of vehicles;
if T isgIf the maximum green light time is more than or equal to the maximum green light time, the green light time of the traffic light is set as the maximum green light time, the step 6 is carried out, and if T is greater than or equal to the maximum green light time, the stepgNot more than the maximum green time, if there is no vehicle at the adjacent crossing, then T is setgSetting the actual green time of the intersection, and turning to the step 6; if the adjacent crossing has the vehicle, go to step 3;
step 3, the Agent at the intersection selects a corresponding influence factor β value from the table 1 according to the number of vehicles at the adjacent intersection, and calculates the green light time T required by the traffic flow at the adjacent intersection to pass through the intersectionneighbor；
TABLE 1 β corresponding relationship table of vehicle number of adjacent crossing
β value Number of vehicles at adjacent crossing
0.8 NumVeh>60
0.7 50<NumVeh≤60
0.6 40<NumVeh≤50
0.5 30<NumVeh≤40
0.4 20<NumVeh≤30
0.3 10<NumVeh≤20
0.2 NumVeh≤10
Because the traffic flow of the adjacent crossing can cause traffic pressure to the crossing, as shown in fig. 2, an influence factor β is adopted to represent the influence degree, the larger the β value is, the larger the traffic pressure of the adjacent crossing to the crossing is, the table of correspondence between the β value and the number of vehicles at the adjacent crossing is shown in table 1, the β value is used for adjusting the traffic flow of the crossing and the adjacent crossing, and the corresponding value is taken according to the number of the vehicles at the adjacent crossing, so that the traffic jam caused by the number of the vehicles passing through the adjacent crossing is avoided;
step 4, calculating the green light time T required by the traffic flow of the adjacent intersection to pass through the intersection by adopting a formula (3) at the intersectionneighbor；
Step 5, the intersection integrates the number of vehicles at the intersection and the adjacent intersections, and calculates the total green light time T required by the intersectiontotalThe calculation formula is shown as formula (3);
Ttotal＝Tarrive+Tg+Tneighbor(3)
if T istotalAnd if the maximum green light time is greater than or equal to the maximum green light time, setting the traffic light green light time as the maximum green light time, and turning to the step 6. If T istotalNot more than the maximum green time, TtotalSetting the actual green time of the intersection, and turning to the step 6;
and 6, the Agent at the intersection applies the calculated green light duration to the traffic signal lamp, observes the fed-back reward signal and transfers the environment state to the state at the next moment.
Step 7, the Agent at the intersection updates the Q table according to the update rule, the Q table is a behavior criterion table, that is, a matrix, as shown in table 2, S represents the environment state, a represents the behavior, and the reward obtained by selecting a certain behavior in the state can be calculated according to the Q value, and the update rule is shown in formula (6):
Figure BDA0002346855820000081
in equation (4), s is the status, a is the action, r is the reward value, α is the learning rate, γ is the discount factor, Qt(s, a) is the Q value of the state selection operation at time t;
finally, judging whether the Q matrix is converged or not, wherein the Q matrix is converged, the procedure is finished, and the Q matrix is not converged and then the step 2 is carried out;
TABLE 2Q Table
Q a1 a2
S 1 1 3
S 2 2 4
As shown in fig. 3, according to the traffic light control algorithm flow chart, a VISSIM-exception VBA-MATLAB integrated simulation platform is used to simulate the proposed adaptive traffic light control mode, and finally, the simulation result is analyzed, and the parameter settings are respectively shown in table 3 and table 4:
TABLE 3 learning System parameter settings
Parameter name Parameter value
ε 0.3
α 0.6
γ 0.5
Table 4 VISSIM parameter settings
Road network arrangement VISSIM parameter setting
Traffic regulations Right-hand traffic
Intersection shape Cross-shaped
Width of road 3.5m
Number of entrances to driveways Bidirectional single lane
Number of vehicles per lane 200 to 2000 vehicles
Simulated duration 3600s
Random seed 15
Speed simulation 10.0Sim.sec./s
Simulated resolution 5Time steps(s)/Sim.sec.
Traffic signal lamp 16 (4 each crossing)
Number of phases Two phases (east west straight going, south north straight going)
Flow detector Stop line of each approach lane
Based on the simulation conditions, the following simulation scenarios are carried out:
and evaluating the performance of the FRTL control mode by using three indexes of average vehicle throughput, average vehicle delay time and average vehicle queuing length at the intersection.
Example 1
As shown in fig. 4, the abscissa is the number of vehicles per hour of each entering lane, and the ordinate is the average throughput, and it can be seen from the figure that when the number of vehicles is less than 300veh/h, the average vehicle throughput at the intersection under the three control modes is almost equal, and when the number of vehicles is between 300veh/h and 600veh/h, the intersection throughput under the trunk control mode and the FRTL control mode is slightly higher than that under the time-sharing control mode, because the trunk control mode only considers the traffic flow of the trunk and ignores the influence of the traffic flow of the non-trunk on the trunk, when the number of vehicles exceeds 1500veh/h, the regulating effect of the trunk control mode and the time-sharing control mode is very small, and the maximum regulating effect is basically achieved, but when the number of vehicles is more than 600veh/h, the average vehicle throughput at the intersection under the FRTL control mode is maximum, the regulation and control effect is best;
example 2
As shown in fig. 5, the abscissa is the number of vehicles per hour, and the ordinate is the average vehicle delay time, and it can be seen from the figure that, when the number of vehicles is about 1500veh/h, the average vehicle delay time in the FRTL control mode is slightly higher than that in the main road control mode, which is caused by a small traffic flow of the secondary road in the time period (the main road control mode sets priority to the road, so there is a concept of the secondary road; the FRTL control mode considers that the road has the same priority), but the average vehicle delay time in the FRTL control mode is the lowest as a whole;
example 3
As shown in fig. 6, the abscissa is the number of vehicles per hour per lane, and it can be seen from the figure that the queuing length of vehicles at multiple intersections is longer than that at single intersection, which is caused by the mutual influence of vehicles at adjacent intersections, and the simulation result shows that the control effect of the three control modes is not very different when the traffic flow is relatively small, and the average queuing length of vehicles is very small. Along with the increase of the traffic flow, when the vehicle input is 1500veh/h or 2000veh/h, the FRTL control mode has a good control effect on the reduction of the queuing length, and the average vehicle queuing length under the FRTL control mode is the shortest as a whole.

Claims (9)

1. A traffic light control method based on fog calculation and reinforcement learning is characterized by comprising the following steps:
step 1, collecting intersection traffic conditions by the intersection fog node and broadcasting the intersection traffic conditions to fog nodes of adjacent intersections;
step 2, the Agent at the intersection collects the number of vehicles at the intersection and calculates the green light time T required by the vehicles to pass through the intersectiong；
Step 3, the Agent at the intersection selects corresponding influence factor values according to the number of vehicles at the adjacent intersections, and calculates the green light time T required by the traffic flow at the adjacent intersections to pass through the intersectionneighbor；
Step 4, calculating the green light time T required by the traffic flow of the adjacent crossing to pass through the crossing by the crossingneighbor；
Step 5, the intersection integrates the number of vehicles at the intersection and the adjacent intersections, and calculates the total green light time T required by the intersectiontotal；
Step 6, the Agent at the intersection applies the calculated green light duration to the traffic signal lamp, observes the feedback reward signal, and shifts the environment state to the state of the next moment;
and 7, updating the Q table by the Agent at the intersection according to the updating rule.
2. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1, wherein the step 1 of testing the traffic conditions at the intersection comprises:
the number of vehicles driving to the next intersection;
testing the time T of the upstream crossing of the crossing to release the traffic flowstart；
Testing the time T of the end of passing the traffic flow at the upstream intersection of the intersectionsend；
Time T for traffic flow of the intersection to reach the next intersectionarriveThe following formula (1):
Figure FDA0002346855810000011
wherein, | Stl1-Stl2I is the distance between two intersections, and v is the average travel speed of the traffic flow.
3. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1, wherein the intersection calculates the green light time T required by itself according to the number of vehicles at the intersection in step 2gThe following formula (2):
Tg＝d+2*NumVeh (2)
where d is the vehicle start delay time and NumVeh is the number of vehicles.
4. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1 or 3, wherein T is the time T in step 2gWhen the traffic light time is not less than the maximum green light time, the green light time of the traffic light is set as the maximum green light time, and the step 6 is carried out; when T isgNo more than maximum green time, no vehicles at adjacent crossing, TgSetting the actual green time of the intersection, and turning to the step 6; and when the adjacent crossing has the vehicle, turning to the step 3.
5. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1, wherein the influence factor β value in step 3 represents the influence degree, the larger the β value is, the larger the traffic pressure of the adjacent intersection to the intersection is, the corresponding β value is taken according to the number of vehicles at the adjacent intersection of the intersection, and the β value is used for adjusting the traffic flow of the intersection and the adjacent intersection.
6. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1, wherein T in step 5totalThe calculation formula (2) is shown in formula (3):
Ttotal＝Tg+Tarrive+Tneighbor(3)。
7. the traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1 or 6, wherein T is the time T in step 5totalWhen the traffic light time is not less than the maximum green light time, the green light time of the traffic light is set as the maximum green light time, and the step 6 is carried out; when T istotalNot more than the maximum green time, TtotalAnd 6, setting the actual green time of the intersection and turning to the step 6.
8. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1, wherein the Q table in step 7 is a matrix, i.e. Q matrix, and the updating rule is as follows:
Figure FDA0002346855810000031
where s is the status, a is the action, r is the return, α is the learning rate, γ is the discount factor, Qt(s, a) is the Q value of the state selection operation at time t.
9. The traffic light control method based on fog calculation and reinforcement learning as claimed in claim 1 or 8, wherein in step 7, whether the Q matrix is converged is judged, the Q matrix is converged, the process is finished, and the Q matrix is not converged and the process goes to step 2.
CN201911398203.XA 2019-12-30 2019-12-30 Traffic light control method based on fog calculation and reinforcement learning Pending CN111047884A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911398203.XA CN111047884A (en) 2019-12-30 2019-12-30 Traffic light control method based on fog calculation and reinforcement learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911398203.XA CN111047884A (en) 2019-12-30 2019-12-30 Traffic light control method based on fog calculation and reinforcement learning

Publications (1)

Publication Number Publication Date
CN111047884A true CN111047884A (en) 2020-04-21

Family

ID=70241723

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911398203.XA Pending CN111047884A (en) 2019-12-30 2019-12-30 Traffic light control method based on fog calculation and reinforcement learning

Country Status (1)

Country Link
CN (1) CN111047884A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111899535A (en) * 2020-06-19 2020-11-06 广东省智能制造研究所 Traffic signal control method and system
CN113012013A (en) * 2021-02-09 2021-06-22 北京工业大学 Cooperative edge caching method based on deep reinforcement learning in Internet of vehicles
CN113362603A (en) * 2021-07-15 2021-09-07 山东交通学院 Regional intersection traffic control method and system based on edge calculation
CN113409598A (en) * 2021-06-08 2021-09-17 智道网联科技(北京)有限公司 Cooperative linkage self-adaptive timing method and device for regional road traffic signal lamps
CN113870588A (en) * 2021-08-20 2021-12-31 深圳市人工智能与机器人研究院 Traffic light control method based on deep Q network, terminal and storage medium
CN114464001A (en) * 2022-01-30 2022-05-10 同济大学 Urban multi-intersection multilayer distribution control system and method under cooperative vehicle and road environment
CN114550456A (en) * 2022-02-28 2022-05-27 重庆长安汽车股份有限公司 Urban traffic jam scheduling method based on reinforcement learning
CN114999164A (en) * 2022-08-05 2022-09-02 深圳支点电子智能科技有限公司 Intelligent traffic early warning processing method and related equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104933876A (en) * 2015-06-03 2015-09-23 浙江师范大学 Control method of self-adaptive smart city intelligent traffic signals
CN105046987A (en) * 2015-06-17 2015-11-11 苏州大学 Pavement traffic signal lamp coordination control method based on reinforcement learning
CN108492592A (en) * 2018-03-01 2018-09-04 西安理工大学 A kind of traffic light intelligent control system and its control method based on mist calculating

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104933876A (en) * 2015-06-03 2015-09-23 浙江师范大学 Control method of self-adaptive smart city intelligent traffic signals
CN105046987A (en) * 2015-06-17 2015-11-11 苏州大学 Pavement traffic signal lamp coordination control method based on reinforcement learning
CN108492592A (en) * 2018-03-01 2018-09-04 西安理工大学 A kind of traffic light intelligent control system and its control method based on mist calculating

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
安萌萌: "基于强化学习的交通灯智能调控研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111899535A (en) * 2020-06-19 2020-11-06 广东省智能制造研究所 Traffic signal control method and system
CN113012013A (en) * 2021-02-09 2021-06-22 北京工业大学 Cooperative edge caching method based on deep reinforcement learning in Internet of vehicles
CN113012013B (en) * 2021-02-09 2024-05-28 北京工业大学 Collaborative edge caching method based on deep reinforcement learning in Internet of vehicles
CN113409598A (en) * 2021-06-08 2021-09-17 智道网联科技(北京)有限公司 Cooperative linkage self-adaptive timing method and device for regional road traffic signal lamps
CN113409598B (en) * 2021-06-08 2022-12-13 智道网联科技(北京)有限公司 Cooperative linkage self-adaptive timing method and device for regional road traffic signal lamps
CN113362603A (en) * 2021-07-15 2021-09-07 山东交通学院 Regional intersection traffic control method and system based on edge calculation
CN113870588A (en) * 2021-08-20 2021-12-31 深圳市人工智能与机器人研究院 Traffic light control method based on deep Q network, terminal and storage medium
CN113870588B (en) * 2021-08-20 2022-12-30 深圳市人工智能与机器人研究院 Traffic light control method based on deep Q network, terminal and storage medium
CN114464001A (en) * 2022-01-30 2022-05-10 同济大学 Urban multi-intersection multilayer distribution control system and method under cooperative vehicle and road environment
CN114550456A (en) * 2022-02-28 2022-05-27 重庆长安汽车股份有限公司 Urban traffic jam scheduling method based on reinforcement learning
CN114550456B (en) * 2022-02-28 2023-07-04 重庆长安汽车股份有限公司 Urban traffic jam scheduling method based on reinforcement learning
CN114999164A (en) * 2022-08-05 2022-09-02 深圳支点电子智能科技有限公司 Intelligent traffic early warning processing method and related equipment

Similar Documents

Publication Publication Date Title
CN111047884A (en) Traffic light control method based on fog calculation and reinforcement learning
CN108510764B (en) Multi-intersection self-adaptive phase difference coordination control system and method based on Q learning
CN112700664B (en) Traffic signal timing optimization method based on deep reinforcement learning
CN107507430B (en) Urban intersection traffic control method and system
CN111951549B (en) Self-adaptive traffic signal lamp control method and system in networked vehicle environment
CN101789178B (en) Optimized control method for traffic signals at road junction
CN106960584A (en) A kind of traffic control method and device of self adaptation crossroad traffic signal lamp
CN106205156A (en) A kind of crossing self-healing control method for the sudden change of part lane flow
CN113947900A (en) Intelligent network connection express way ramp cooperative control system
CN103593535A (en) Urban traffic complex self-adaptive network parallel simulation system and method based on multi-scale integration
Kong et al. Urban arterial traffic two-direction green wave intelligent coordination control technique and its application
CN109887289A (en) A kind of network vehicle flowrate maximization approach of urban traffic network model
CN111681433A (en) Intersection traffic signal lamp timing optimization method and device
US10891855B2 (en) Method to schedule intelligent traffic lights in real time based on digital infochemicals
CN109544913A (en) A kind of traffic lights dynamic timing algorithm based on depth Q e-learning
AU2021103022A4 (en) A Method of Controlling Traffic Light Based on Fog Computing and Reinforcement Learning
Guerrero-Ibanez et al. A policy-based multi-agent management approach for intelligent traffic-light control
CN113780624B (en) Urban road network signal coordination control method based on game equilibrium theory
CN109816978B (en) Regional group traffic guidance system and method considering dynamic response behaviors of drivers
CN114038216B (en) Signal lamp control method based on road network division and boundary flow control
CN105551250A (en) Method for discriminating urban road intersection operation state on the basis of interval clustering
CN105743783B (en) Car networking network node screening technique based on BS-TS and autoencoder network
Shen et al. Study on road network traffic coordination control technique with bus priority
CN113299078A (en) Multi-mode traffic trunk line signal coordination control method and device based on multi-agent cooperation
Teo et al. Agent-based traffic flow optimization at multiple signalized intersections

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200421