CN113781784B - Intelligent traffic light and control method thereof - Google Patents
Intelligent traffic light and control method thereof Download PDFInfo
- Publication number
- CN113781784B CN113781784B CN202111317452.9A CN202111317452A CN113781784B CN 113781784 B CN113781784 B CN 113781784B CN 202111317452 A CN202111317452 A CN 202111317452A CN 113781784 B CN113781784 B CN 113781784B
- Authority
- CN
- China
- Prior art keywords
- road section
- current road
- condition information
- current
- road condition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
-
- 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 an intelligent traffic light and a control method thereof, wherein the control method of the intelligent traffic light comprises the following steps: according to the acquired current road condition information of the previous road section and the current road condition information of the current road section, when the current road section is confirmed to have special vehicles, the traffic lights are controlled to be red lights so as to give way for the special vehicles; the intelligent traffic light control method also predicts the traffic flow of the current road section in the next time period according to the current road condition information of the previous road section, the historical information of the previous road section, the current road condition information of the current road section and the historical road condition information, and determines a traffic light timing strategy according to the traffic flow of the next time period. The influence of the previous road section on the current road condition is fully considered, the data volume input by the deep learning model is increased, the prediction of the traffic flow in the next time period is more accurate, the timing strategy of the traffic light is more reasonable, and the utilization rate of traffic resources is effectively increased.
Description
Technical Field
The invention relates to the technical field of light control, in particular to an intelligent traffic light and a control method thereof.
Background
Traffic congestion increasingly becomes one of the main problems restricting urban and economic development.
How to more effectively utilize traffic resources and shorten travel time is still a problem to be solved urgently.
Disclosure of Invention
The invention mainly aims to provide an intelligent traffic light control method, aiming at effectively improving the utilization rate of traffic resources by controlling the timing strategy of a traffic light.
In order to achieve the purpose, the invention provides an intelligent traffic light control method, which comprises the following steps:
acquiring current road condition information of a previous road section, historical information of the previous road section, current road condition information of the current road section and historical road condition information;
according to the current road condition information of the previous road section and the current road condition information of the current road section, when the current road section is confirmed to have special vehicles, the traffic lights are controlled to be red lights so as to give way for the special vehicles;
predicting the traffic flow of the current road section in the next time period according to the current road condition information of the previous road section, the historical information of the previous road section, the current road condition information of the current road section and the historical road condition information;
and determining the traffic light timing strategy according to the traffic flow of the next time period.
In an embodiment, the acquiring current traffic information of a previous road segment, historical traffic information of the previous road segment, current traffic information of the current road segment, and historical traffic information includes:
shooting road condition images at the current time period according to a first preset time interval through a camera of the previous road section, taking a plurality of images shot in the current time period and shooting time as previous road condition information, and storing the previous road condition information as historical road condition information of a subsequent time period;
the method comprises the steps of shooting road condition images at a current time period according to a first preset time interval through a camera of the current road section, taking a plurality of images shot in the current time period and shooting time as current road condition information, and storing the current road condition information as historical road condition information of a subsequent time period.
In an embodiment, when it is determined that there is a special vehicle in the current road section according to the current road condition information of the previous road section and the current road condition information of the current road section, the traffic lights are controlled to be red lights, so that the step of giving way for the special vehicle specifically includes:
setting a first deep learning model, and training the first deep learning model by using historical road condition information of a previous road section and road condition information of a current road section;
and inputting the current road condition information of the previous road section and the current road condition information of the current road section into the first deep learning model, so that the first deep learning model can confirm whether the special vehicle exists in the current road section.
In an embodiment, the first deep learning model comprises:
the first convolution neural network is used for outputting a first characteristic vector according to the current road condition information of the previous road section;
the second convolutional neural network is used for outputting a second feature vector according to the current road condition information of the current road section;
the input end of the first feature fusion layer is respectively connected with the first convolutional neural network and the second convolutional neural network, and the first feature fusion layer is used for fusing a first feature vector and the second feature vector and outputting a first fused feature vector;
and the first output layer is used for outputting and confirming whether the special vehicle exists in the current road section or not according to the first fusion feature vector.
In an embodiment, the step of predicting the traffic flow of the current road section in the next time period according to the current road condition information of the previous road section, the historical information of the previous road section, the current road condition information of the current road section, and the historical road condition information specifically includes:
setting a second deep learning model, and training the second deep learning model by using historical road condition information of a previous road section and historical road condition information of a current road section;
and inputting the current road condition information of the previous road section and the current road condition information of the current road section into the second deep learning model, so that the second deep learning model can predict the traffic flow of the current road section in the next time period.
In an embodiment, the second deep learning model comprises:
the third convolutional neural network is used for sequentially inputting a plurality of images in the previous path of condition information according to a time sequence and sequentially outputting corresponding third eigenvectors;
the cyclic neural network is used for sequentially receiving the third eigenvectors and outputting fourth eigenvectors according to the third eigenvectors;
the fourth convolutional neural network is used for sequentially inputting a plurality of images in the current road condition information according to a time sequence and sequentially outputting corresponding fifth feature vectors;
the input end of the second feature fusion layer is connected with the output end of the recurrent neural network and the output end of the fourth convolutional neural network, and is used for fusing the fourth feature vector and the fifth feature vector and outputting a second fused feature vector;
and the second output layer is used for outputting the traffic flow of the next time period according to the second fusion characteristic vector.
In one embodiment, an input of the third convolutional neural network is connected to an input of the recurrent neural network.
In one embodiment, an output of the third convolutional neural network is connected to an output of the recurrent neural network.
In one embodiment, the intelligent traffic light control method further comprises the following steps:
setting a third deep learning model, and training the third deep learning model by using the road condition information of the current road section;
and inputting the current road condition information of the current road section into the third deep learning model for the third deep learning model to confirm whether a traffic accident occurs in the current road section, and sending an alarm signal through a wireless communication device when the traffic accident occurs in the current road section.
In one embodiment, the recurrent neural network is an episodic memory neural network.
The invention provides an intelligent traffic light, comprising:
a traffic light;
the control module is connected with the traffic light and comprises a memory, a processor and an intelligent traffic light control program which is stored on the memory and can run on the processor, and when the intelligent traffic light control program is executed, the intelligent traffic light control method is realized.
The technical scheme of the invention predicts the traffic flow by acquiring the road condition information of the previous road section and the current road section at the current moment, and the inventor fully considers the vehicle driven to the current road section from the previous road section in the next time period by taking the vehicle condition information of the previous road section into consideration, namely the supplement/influence of the previous road section on the traffic flow of the current road section, and neglects the supplement of the previous road section on the traffic flow of the current road section in the next time period by comparison; compared with the method for independently collecting the road condition information of the current road section, the method for predicting the traffic flow of the traffic light in the deep learning model also collects the road condition information of the previous road section, so that the information quantity, namely the data quantity, input by the deep learning model is increased, the traffic flow of the next time period can be more accurately predicted, the timing strategy for adjusting the traffic light according to the traffic flow is more reasonable, and the utilization rate of traffic resources is effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
FIG. 1 is a flow chart of an embodiment of an intelligent traffic light control method of the present invention;
FIG. 2 is a flow chart of another embodiment of an intelligent traffic light control method of the present invention;
fig. 3 is a schematic diagram of a model structure of the intelligent traffic light control method of the present invention.
The reference numbers illustrate:
| reference numerals | Name (R) | Reference numerals | Name (R) |
| 10 | First |
22 | Recurrent |
| 20 | Second |
23 | Fourth convolutional |
| 11 | First convolutional |
24 | Second |
| 12 | Second convolutional |
25 | |
| 13 | First |
31 | |
| 14 | |
32 | Second input layer |
| 21 | A third convolutional neural network |
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In addition, the descriptions related to "first", "second", etc. in the present invention are only for descriptive purposes and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between various embodiments may be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present invention.
The invention provides an intelligent traffic light control method, which can be applied to intelligent traffic lights, wherein the intelligent traffic lights can comprise traffic lights with three colors of red, green and yellow, a camera and a processor (FPGA and computer/server combination can be used, in the training of a deep learning model, a deep learning model can be trained on the computer/server firstly, then the trained model is compressed and stored in the FPGA, and the deep learning model is operated by the FPGA. The camera can acquire road condition information, the processor can operate the intelligent traffic light control program in the memory to realize the intelligent traffic light control method, so that the traffic flow of the next time period can be predicted according to the road condition of the road, a proper traffic light timing strategy is determined according to the traffic flow information, and the utilization rate of traffic resources is effectively improved.
Referring to fig. 1, in an embodiment of the present invention, the intelligent traffic light control method includes the following steps:
s100, acquiring current road condition information of a previous road section, historical information of the previous road section, current road condition information of the current road section and historical road condition information;
the traffic information may include: the method comprises the following steps of acquiring one or more of road images shot by a camera and audio signals acquired by a sound sensor (microphone). The previous road section and the current road section can be sent to the current road section through the wireless communication device, such as Bluetooth and wifi.
S200, according to the current road condition information of the previous road section and the current road condition information of the current road section, when the current road section is confirmed to have special vehicles, controlling all traffic lights to be red lights so as to give way for the special vehicles;
in the step, a single neural network can be established, the current road condition information of the previous road section and the current road condition information of the current road section are spliced and then input into the deep learning model, so that whether special vehicles exist in the current road section or not is determined in a classified mode.
It is also possible to provide a plurality of parallel neural networks. For example, a deep learning model (using the traffic information of the previous road section as training data) is established, and for the traffic information of the previous road section, a probability of whether the current road section has a special vehicle (police car, emergency 120, fire truck, etc.) is output, another deep learning model is established (using the traffic information of the current road section as training data), and for the information of the current traffic section, another probability of whether the current traffic has a special vehicle is output; the two neural networks run in parallel, the two probabilities are added through a weighting algorithm to obtain the probability of whether special vehicles exist in the current road section, when the probability is greater than 50%, 60%, 70%, 80% or 90% (the probability value can specifically balance sensitivity and accuracy, and is set according to actual requirements), the fact that special vehicles exist in the current road section is determined, and when the special vehicles exist, the traffic lights are set to be red lights to prohibit other vehicles from passing, so that the special vehicles can make way, and the working efficiency of the special vehicles is improved.
S300, predicting the traffic flow of the current road section in the next time period according to the current road condition information of the previous road section, the historical information of the previous road section, the current road condition information of the current road section and the historical road condition information;
in the step, a single neural network can be established, the current road condition information of the previous road section and the current road condition information of the current road section are spliced and then input into the neural network, and the neural network determines the traffic flow according to the spliced road condition information. Of course, a plurality of parallel neural networks may be set, and the road condition information of the current road section and the road condition information of the previous road section are respectively input and run in parallel, which is not described herein again.
S400, determining the traffic light timing strategy according to the traffic flow of the next time period.
The next time period may be a duration of the green light, a time of one cycle of the traffic light, or a preset time period, for example, 15 minutes.
In this step, the green duration of the traffic light can be adjusted appropriately according to the magnitude of the traffic flow. The formula for calculating the traffic flow may be: vehicle flow = vehicle speed per unit time/(vehicle distance + vehicle length). At the moment, the vehicle type does not need to be judged, the unit time in the formula is determined only according to the predicted traffic flow, the timing strategy of the traffic lights is set to be that the green light duration is slightly longer than the unit time, the passing requirement of the vehicle can be met, and the green light duration is properly prolonged under the condition of large traffic flow. That is, the traffic flow and the timing strategy are in one-to-one correspondence, and the timing strategy is determined according to the traffic flow.
A deep learning model can also be set, the traffic flow is input into the deep learning model, and the deep learning model outputs a corresponding timing strategy.
The technical scheme of the invention predicts the traffic flow by acquiring the road condition information of the previous road section and the current road section at the current moment, and the inventor fully considers the vehicle driven to the current road section from the previous road section in the next time period by taking the vehicle condition information of the previous road section into consideration, namely the supplement/influence of the previous road section on the traffic flow of the current road section, and neglects the supplement of the previous road section on the traffic flow of the current road section in the next time period by comparison; compared with the method for independently collecting the road condition information of the current road section, the method for predicting the traffic flow of the traffic light in the deep learning model also collects the road condition information of the previous road section, so that the information quantity, namely the data quantity, input by the deep learning model is increased, the traffic flow of the next time period can be more accurately predicted, the timing strategy for adjusting the traffic light according to the traffic flow is more reasonable, and the utilization rate of traffic resources is effectively improved.
Referring to fig. 2, in an embodiment, the step of acquiring the current traffic information of the previous road section, the historical traffic information of the previous road section, the current traffic information of the current road section, and the historical traffic information includes:
s101, shooting road condition images at the current time period according to a first preset time interval through a camera of the previous road section, taking a plurality of images shot in the current time period and shooting time as previous road condition information, and storing the previous road condition information as historical road condition information of a subsequent time period;
s102, shooting road condition images in the current time period according to a first preset time interval through a camera of the current road section, taking a plurality of images shot in the current time period and shooting time as current road condition information, and storing the current road condition information as historical road condition information in a subsequent time period.
In this embodiment, the first preset time may be 10 seconds, 10 minutes or longer, and the current time period and the next time period may be continuous time periods obtained by dividing time into a plurality of 10-20 minutes, that is, the duration of each time period is 10-20 minutes.
This embodiment is through setting up gather many images as input information in the first preset time, and the information volume is abundanter as the single picture of comparison, and the collection time's of picture is continuous, can fully consider time information, and further, this embodiment still will shoot the time as input information, and the effectual different information of considering different time quantums, traffic flow is different (for example, off duty rush hour traffic flow is different with other time traffic flow).
Referring to fig. 1, in an embodiment, when it is determined that there is a special vehicle in the current road section according to the current road condition information of the previous road section and the current road condition information of the current road section, the traffic lights are controlled to be red lights, so as to provide the special vehicle with the traffic lights specifically:
setting a first deep learning model 10, and training the first deep learning model 10 by using historical road condition information of a previous road section and road condition information of a current road section;
inputting the current road condition information of the previous road section and the current road condition information of the current road section into the first deep learning model 10, so that the first deep learning model 10 can confirm whether the special vehicle exists in the current road section.
It should be noted that the deep learning model does not need to extract features of historical road condition information, and can effectively retain all features of the historical road condition information. The first deep learning model 10 may be any type of deep learning model, or may be a fusion of multiple types of deep learning models.
In this embodiment, the deep learning module may have one input end for inputting the current traffic information of the previous road section and the current traffic information of the current road section after being spliced, or may have two input ends for inputting the current traffic information of the previous road section and the current traffic information of the current road section respectively.
Referring to fig. 3, in one embodiment, the first deep learning model 10 includes:
the first convolution neural network 11 is configured to output a first feature vector according to current road condition information of a previous road section;
the second convolutional neural network 12 is configured to output a second feature vector according to the current road condition information of the current road section;
a first feature fusion layer 13, an input end of which is connected to the first convolutional neural network 11 and the second convolutional neural network 12, respectively, wherein the first feature fusion layer 13 is configured to fuse the first feature vector and the second feature vector and output a first fused feature vector;
and the first output layer 14 is used for outputting and confirming whether the special vehicle exists in the current road section or not according to the first fusion feature vector.
In this embodiment, the first deep learning model 10 includes a first input layer 31 and a second input layer 32, which are used to input the current road condition information of the previous road section and the current road condition information of the current road section, respectively. In this embodiment, a full connection layer may be further disposed between the first feature fusion layer 13 and the first output layer 14, and the information of the first fusion feature vector is integrated to improve the correlation between the first feature vector and the second feature vector. It is to be understood that the first feature vector and the second feature vector do not refer to one feature vector alone, and may be a combination of a plurality of vectors. In order to realize the fusion of the first feature vector and the second feature vector, a linear layer may be disposed at an output end of the first convolutional neural network 11 or the second convolutional neural network 12, the first feature vector or the second feature vector is mapped, and finally, the dimensions of the first feature vector and the second feature vector are consistent, so that the first feature fusion layer 13 fuses the first feature vector and the second feature vector, the first feature fusion layer may adopt a connection function concatenate, the first feature vector and the second feature vector are fused, the connection function may reserve the number of channels, and facilitate back propagation, and the first output layer 14 may adopt a softmax function. The embodiment inputs the road condition information of the previous road section and the current road section into the two neural networks respectively in parallel, and the two neural networks operate in parallel, so that mutual interference is avoided.
Referring to fig. 1, in an embodiment, the step of predicting the traffic flow of the current road section in the next time period according to the current road condition information of the previous road section, the historical information of the previous road section, the current road condition information of the current road section, and the historical road condition information specifically includes:
setting a second deep learning model 20, and training the second deep learning model 20 by using the historical road condition information of the previous road section and the historical road condition information of the current road section;
and inputting the current road condition information of the previous road section and the current road condition information of the current road section into the second deep learning model 20, so that the second deep learning model 20 can predict the traffic flow of the current road section in the next time period.
In this embodiment, the second deep learning model 20 may be any type of deep learning model, or may be a fusion of multiple types of deep learning models.
Referring to fig. 3, in an embodiment, the second deep learning model 20 includes:
the third convolutional neural network 21 is used for sequentially inputting a plurality of images in the previous path of condition information according to a time sequence and sequentially outputting corresponding third feature vectors;
the recurrent neural network 22 is used for sequentially receiving the third eigenvectors and outputting fourth eigenvectors according to the third eigenvectors;
the fourth convolutional neural network 23 is configured to sequentially input a plurality of images in the current road condition information according to a time sequence, and sequentially output a corresponding fifth feature vector;
a second feature fusion layer 24, an input end of which is connected to the output end of the recurrent neural network 22 and the output end of the fourth convolutional neural network 23, and which fuses the fourth feature vector and the fifth feature vector and outputs a second fused feature vector;
and the second output layer 25 is configured to output the traffic flow in the next time period according to the second fused feature vector.
In this embodiment, the second deep learning model 20 further includes a first input layer 31 and a second input layer 32 (which may be shared with the input layer of the first deep learning model 10, and perform the recognition of the special vehicle and the recognition of the traffic flow rate at the same time), and the first input layer 31 and the second input layer 32 respectively input the traffic information of the previous road section and the traffic information of the current road section.
It should be noted that the road condition information is a set of multiple pictures containing a time relationship, the recurrent neural network 22 can integrate the information connection between two consecutive pictures in time to fully extract the time information, and the convolutional neural network can extract the features of the pictures, so that the recurrent neural network 22 can understand the potential features of the input information more easily. In addition, one or more full connection layers may be disposed between the second feature fusion layer 24 and the second output layer 25, and the second fusion feature vector is converted into a space to output features that are easy to output layer classification, so that the second output layer 25 classifies the features.
In the embodiment, the third convolutional neural network 21 and the cyclic neural network 22 are arranged in a unified framework of joint training, and the classification precision is higher compared with that of a single cyclic neural network 22 or convolutional neural network. Specifically, the third convolutional neural network 21 and the cyclic neural network 22 form a first channel, the fourth convolutional neural network 23 forms a second channel, the first channel and the second module run in parallel, the road condition information of the current road section and the road condition information of the previous road section are both used as input, the information of the two road sections is integrated, the input information amount of the second deep learning model 20 is improved, and the detection accuracy of the traffic flow of the current road section is effectively improved.
Referring to fig. 3, in one embodiment, an input of the third convolutional neural network 21 is connected to an input of the recurrent neural network 22.
In the present embodiment, the input information of the third convolutional neural network 21 is simultaneously input to the recurrent neural network 22, so that additional information is added to the recurrent neural network 22, that is, a multi-scale feature is provided for the recurrent neural network 22, and the performance of the recurrent neural network 22 is effectively improved.
Referring to fig. 3, in one embodiment, an output of the third convolutional neural network 21 is connected to an output of the recurrent neural network 22.
In this embodiment, that is, the output of the third convolutional neural network 21 is also used as the output of the first channel, and it should be noted that a linear layer is further provided between the output of the third convolutional neural network 21 and the output of the recurrent neural network 22, and the linear layer maps the output (third eigenvector) of the third convolutional neural network 21 to one dimension with the output (fourth eigenvector) of the recurrent neural network 22, so that the second feature fusion layer 24 can conveniently fuse the third eigenvector and the fourth eigenvector.
In one embodiment, the recurrent neural network 22 is an episodic memory neural network.
The long-and-short-term memory neural network is an improved recurrent neural network 22, which can solve the problem that the recurrent neural network 22 cannot handle long-distance dependence.
In one embodiment, the intelligent traffic light control method further comprises the following steps:
setting a third deep learning model, and training the third deep learning model by using the road condition information of the current road section;
in this embodiment, the third deep learning model may include an input layer, a convolutional layer, a fully-connected layer, and an output layer.
And taking the road condition information of the current road section as a data set, dividing the data set into a training set and a testing set according to a certain proportion, training the third deep learning model by using the training set, testing the performance of the third deep learning model by using the testing set, and storing the third deep learning model in a memory.
And inputting the current road condition information of the current road section into the third deep learning model for the third deep learning model to confirm whether a traffic accident occurs in the current road section, and sending an alarm signal through a wireless communication device when the traffic accident occurs in the current road section.
The sending of the alarm signal can be performed by a communication device, and the communication device for transmitting the road condition information of the previous road section can be the same communication device or different communication devices.
The communication device may be a wireless communication device, such as wifi, bluetooth, mobile communication device, or a wired communication device, which is not limited herein.
The present invention also proposes an intelligent traffic light, comprising:
a traffic light;
the control module is connected with the traffic light and comprises a memory, a processor and an intelligent traffic light control program which is stored on the memory and can run on the processor, and when the intelligent traffic light control program is executed, the intelligent traffic light control method is realized.
The specific scheme of the intelligent traffic light control method refers to the above embodiments, and since the intelligent traffic light adopts all the technical schemes of all the above embodiments, all the beneficial effects brought by the technical schemes of the above embodiments are at least achieved, and are not repeated herein.
The above description is only an alternative embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications and equivalents of the present invention, which are made by the contents of the present specification and the accompanying drawings, or directly/indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (8)
1. An intelligent traffic light control method is characterized by comprising the following steps:
acquiring current road condition information of a previous road section, historical information of the previous road section, current road condition information of the current road section and historical road condition information of the current road section;
according to the current road condition information of the previous road section and the current road condition information of the current road section, when the current road section is confirmed to have special vehicles, the traffic lights are controlled to be red lights so as to give way for the special vehicles;
predicting the traffic flow of the next time section of the current road section according to the current road condition information of the previous road section, the historical road condition information of the previous road section, the current road condition information of the current road section and the historical road condition information of the current road section;
determining the traffic light timing strategy according to the traffic flow of the next time period;
according to the current road condition information of the previous road section and the current road condition information of the current road section, when the current road section is confirmed to have special vehicles, the traffic lights are controlled to be red lights, and the step of giving way for the special vehicles is specifically as follows:
setting a first deep learning model, and training the first deep learning model by using historical road condition information of a previous road section and historical road condition information of a current road section;
inputting the current road condition information of the previous road section and the current road condition information of the current road section into the first deep learning model, and allowing the first deep learning model to confirm whether the special vehicle exists in the current road section;
the first deep learning model comprises:
the first convolution neural network is used for outputting a first characteristic vector according to the current road condition information of the previous road section;
the second convolutional neural network is used for outputting a second feature vector according to the current road condition information of the current road section;
the input end of the first feature fusion layer is respectively connected with the first convolutional neural network and the second convolutional neural network, and the first feature fusion layer is used for fusing the first feature vector and the second feature vector and outputting a first fused feature vector;
and the first output layer is used for outputting and confirming whether the special vehicle exists in the current road section or not according to the first fusion feature vector.
2. The intelligent traffic light control method according to claim 1, wherein the acquiring current traffic information of a previous road section, historical traffic information of the previous road section, current traffic information of the current road section, and historical traffic information of the current road section comprises:
shooting road condition images at a current time period according to a first preset time interval through a camera of a previous road section, taking a plurality of images shot in the current time period and shooting time as current road condition information of the previous road section, and storing the current road condition information as historical road condition information of the previous road section in a subsequent time period;
the method comprises the steps of shooting road condition images at a current time period according to a first preset time interval through a camera of the current road section, taking a plurality of images shot in the current time period and shooting time as current road condition information of the current road section, and storing the current road condition information as historical road condition information of the current road section in a subsequent time period.
3. The intelligent traffic light control method according to claim 1, wherein the step of predicting the traffic flow for the next time segment of the current road segment based on the current road condition information of the previous road segment, the historical road condition information of the previous road segment, the current road condition information of the current road segment, and the historical road condition information of the current road segment specifically comprises:
setting a second deep learning model, and training the second deep learning model by using historical road condition information of a previous road section and historical road condition information of a current road section;
and inputting the current road condition information of the previous road section and the current road condition information of the current road section into the second deep learning model, so that the second deep learning model can predict the traffic flow of the current road section in the next time period.
4. The intelligent traffic light control method of claim 3, wherein the second deep learning model comprises:
the third convolutional neural network is used for sequentially inputting a plurality of images in the current road condition information of the previous road section according to the time sequence and sequentially outputting corresponding third feature vectors;
the cyclic neural network is used for sequentially receiving the third eigenvectors and outputting fourth eigenvectors according to the third eigenvectors;
the fourth convolutional neural network is used for sequentially inputting a plurality of images in the current road condition information of the current road section according to a time sequence and sequentially outputting corresponding fifth feature vectors;
the input end of the second feature fusion layer is connected with the output end of the recurrent neural network and the output end of the fourth convolutional neural network, and is used for fusing the fourth feature vector and the fifth feature vector and outputting a second fused feature vector;
and the second output layer is used for outputting the traffic flow of the next time period according to the second fusion characteristic vector.
5. The intelligent traffic light control method according to claim 4, wherein an input terminal of the third convolutional neural network is connected to an input terminal of the cyclic neural network; and/or the presence of a gas in the gas,
and the output end of the third convolutional neural network is connected with the output end of the cyclic neural network.
6. The intelligent traffic light control method according to claim 4, wherein the recurrent neural network is an episodic memory neural network.
7. The intelligent traffic light control method according to claim 1, further comprising the steps of:
setting a third deep learning model, and training the third deep learning model by using historical road condition information of the current road section;
and inputting the current road condition information of the current road section into the third deep learning model for the third deep learning model to confirm whether a traffic accident occurs in the current road section, and sending an alarm signal through a wireless communication device when the traffic accident occurs in the current road section.
8. An intelligent traffic lamp, comprising:
a traffic light;
a control module connected to the traffic light, the control module comprising a memory, a processor, an intelligent traffic light control program stored on the memory and executable on the processor, the intelligent traffic light control program when executed implementing the intelligent traffic light control method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111317452.9A CN113781784B (en) | 2021-11-09 | 2021-11-09 | Intelligent traffic light and control method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111317452.9A CN113781784B (en) | 2021-11-09 | 2021-11-09 | Intelligent traffic light and control method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113781784A CN113781784A (en) | 2021-12-10 |
| CN113781784B true CN113781784B (en) | 2022-02-11 |
Family
ID=78956921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111317452.9A Active CN113781784B (en) | 2021-11-09 | 2021-11-09 | Intelligent traffic light and control method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113781784B (en) |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2013250854A (en) * | 2012-06-01 | 2013-12-12 | Mitsubishi Motors Corp | Emergency vehicle notification apparatus |
| CN205644972U (en) * | 2016-04-06 | 2016-10-12 | 兰州交通大学 | Intelligent urban traffic control system |
| CN106504548B (en) * | 2016-10-27 | 2019-04-09 | 广州子灵信息科技有限公司 | Traffic lights intelligent control method and system |
| CN106355885A (en) * | 2016-11-24 | 2017-01-25 | 深圳市永达电子信息股份有限公司 | Traffic signal dynamic control method and system based on big data analysis platform |
| CN107689158A (en) * | 2017-08-10 | 2018-02-13 | 五邑大学 | A kind of intellectual traffic control method based on image procossing |
| CN108831168B (en) * | 2018-06-01 | 2020-10-02 | 武汉泰坦智慧科技有限公司 | Traffic signal lamp control method and system based on visual identification of associated intersection |
| CN108629990A (en) * | 2018-06-14 | 2018-10-09 | 重庆同济同枥信息技术有限公司 | A kind of real-time dynamic timing method and system based on multi-source data |
| CN110969879B (en) * | 2018-09-29 | 2022-12-27 | 上海博泰悦臻网络技术服务有限公司 | Automatic control method and system of traffic signal lamp |
| AU2019250256A1 (en) * | 2018-10-19 | 2020-05-07 | Bradley James | An emergency vehicle proximity alert system |
| CN112419726B (en) * | 2020-11-20 | 2022-09-20 | 华南理工大学 | Urban traffic signal control system based on traffic flow prediction |
-
2021
- 2021-11-09 CN CN202111317452.9A patent/CN113781784B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113781784A (en) | 2021-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210334420A1 (en) | Driving simulation method and apparatus, electronic device, and computer storage medium | |
| CN109000668B (en) | Real-time intelligent navigation method based on Internet of vehicles | |
| CN111462484B (en) | Congestion state determination method, device, equipment and computer readable storage medium | |
| US20210138927A1 (en) | Device, system and method for predicting battery consumption of electric vehicle | |
| CN109064741B (en) | Method for reconstructing running track of main road vehicle based on multi-source data fusion | |
| CN109246364A (en) | Photographic device method of adjustment and relevant device | |
| US9834197B2 (en) | Section acquisition system, section acquisition method, and section acquisition program | |
| CN106971544B (en) | A kind of direct method that vehicle congestion is detected using still image | |
| KR20200040501A (en) | Method, apparatus and computer program for predicting traffic information | |
| US20210172749A1 (en) | Controlling platooning vehicle | |
| US10755564B2 (en) | System to optimize SCATS adaptive signal system using trajectory data | |
| CN111473789B (en) | Navigation route planning method, device and equipment | |
| Zhong et al. | Traffic flow characteristics and lane use strategies for connected and automated vehicles in mixed traffic conditions | |
| CN114743375A (en) | Traffic flow prediction method based on machine vision fusion road network space-time information | |
| CN113781784B (en) | Intelligent traffic light and control method thereof | |
| CN109166336B (en) | Real-time road condition information acquisition and pushing method based on block chain technology | |
| CN110189537B (en) | Parking guidance method, device and equipment based on space-time characteristics and storage medium | |
| CN113129609B (en) | Traffic signal lamp timing method and traffic signal equipment | |
| Helmus et al. | SEVA: A data driven model of electric vehicle charging behavior | |
| CN111047292A (en) | Intelligent transportation tool, intelligent equipment and intelligent travel reminding method | |
| CN108759850A (en) | New-energy automobile charging navigation system | |
| CN113795010B (en) | Method and device for testing automobile queue, electronic equipment and storage medium thereof | |
| Elbery et al. | Vehicular communication and mobility sustainability: the mutual impacts in large-scale smart cities | |
| CN108413979A (en) | New-energy automobile charging air navigation aid | |
| KR100805664B1 (en) | Server, system, and method for creating and providing traffic information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |