CN109697867B - Deep learning-based traffic control method and system - Google Patents
Deep learning-based traffic control method and system Download PDFInfo
- Publication number
- CN109697867B CN109697867B CN201910078907.2A CN201910078907A CN109697867B CN 109697867 B CN109697867 B CN 109697867B CN 201910078907 A CN201910078907 A CN 201910078907A CN 109697867 B CN109697867 B CN 109697867B
- Authority
- CN
- China
- Prior art keywords
- traffic
- data
- local processor
- signal lamp
- central server
- 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
- G08G1/07—Controlling traffic signals
- G08G1/08—Controlling traffic signals according to detected number or speed of vehicles
Abstract
The invention relates to the field of artificial intelligence and traffic control, and discloses a traffic control method and system based on deep learning, wherein the method comprises the following steps: the local processor obtains video data of the traffic intersection, a deep neural network is adopted for analysis and calculation to obtain traffic data, and the local processor under each mode selects a self-control signal lamp or optimizes the setting of the signal lamp by combining with a city central server according to the traffic data. Through the mode, the embodiment of the invention can solve the technical problem that the traffic road condition cannot be accurately monitored at present, realizes the balanced dispatching of the signal lamps and improves the traffic strain capacity.
Description
Technical Field
The embodiment of the invention relates to the field of artificial intelligence and traffic control, in particular to a traffic control method and system based on deep learning.
Background
With the high-speed urbanization of society and the improvement of the living standard of people, the number of vehicles is more and more, and the phenomenon of traffic jam is more and more serious. The traditional video monitoring technology cannot give accurate data such as vehicle density, average speed, traffic flow, pedestrian flow and the like. Due to the lack of accurate data, it is difficult to provide an accurate real-time traffic control strategy. Even if a large amount of manpower is spent, the signal lamp scheduling of a plurality of intersections cannot be balanced, and the 24-hour management cannot be achieved. In the traffic road condition prediction, the time of the peak period cannot be accurately judged due to the fact that accurate data are not available. And macroscopic judgment of the traffic condition of the whole city cannot be given.
At present, a video monitoring mode is generally adopted, road traffic conditions are obtained through a camera arranged on a road and fed back to a background server, but real-time data analysis of the traffic conditions is lacked, and the traffic road conditions cannot be predicted.
Based on the above, the invention provides a traffic control method and system based on deep learning, which solve the technical problem that the traffic road condition cannot be accurately monitored at present, realize the balanced dispatching of signal lamps and improve the traffic strain capacity.
Disclosure of Invention
In order to solve the technical problems, embodiments of the present invention provide a traffic control method and system based on deep learning, which solve the technical problem that the traffic road condition cannot be accurately monitored at present, implement balanced scheduling of signal lamps, and improve traffic strain capacity.
In order to solve the above technical problem, an embodiment of the present invention provides the following technical solutions:
in a first aspect, an embodiment of the present invention provides a deep learning-based traffic control method, which is applied to a deep learning-based traffic control system, where the traffic control system includes: the high-definition camera and the signal lamp are respectively connected with the local processor, the local processor is connected with the urban traffic central server, and the method comprises the following steps:
the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase;
determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode;
if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data;
if not, the local processor sends a signal lamp adjustment request to the urban traffic central server;
and the urban traffic central server receives the signal lamp adjusting request and judges whether the signal lamp adjusting request is reasonable, if so, the local processor is controlled to adjust the current setting value of the signal lamp based on the signal lamp adjusting request.
In some embodiments, the method further comprises: and if the local processor is in an autonomous control mode, adjusting the signal lamp in real time according to the real-time retention and the real-time retention of each traffic intersection.
In some embodiments, said processing said video data through a deep neural network to determine traffic data comprises:
recording the number of vehicles waiting for red lights on each road of the traffic intersection as real-time parking amount according to the video data of the traffic intersection; recording the number of vehicles which are at green lights but do not pass through a stop line on each road of the traffic intersection as real time lag allowance;
recording the approximate time required by all vehicles in a certain traffic flow direction to pass through the traffic intersection as the passing time; the average value of the ratio of the multiple groups of traffic time of a certain phase to the corresponding green light time represents the traffic jam index of the phase;
determining the number of vehicles passing through a road section corresponding to a certain phase in unit time as the traffic flow of the phase;
determining the maximum number of vehicles which continuously pass through a stop line corresponding to a certain phase within one continuous green light time as the traffic capacity of the phase; and taking the ratio of the number of passing vehicles in a unit time of a certain phase to the passing capacity of the phase as the passing index of the phase.
In some embodiments, the method further comprises:
the urban traffic central server receives traffic data sent by the local processor;
the urban traffic central server establishes and initializes a plurality of performance tables of a plurality of traffic intersections according to the traffic data, wherein each traffic intersection corresponds to each performance table one by one, the performance tables are used for correspondingly storing the data of the traffic intersections, and the data comprises the week number, the time period, the holidays, the traffic light setting, the traffic jam index and the traffic capacity.
In some embodiments, the method further comprises:
and the urban traffic central server updates the performance table in real time according to the traffic data sent by the local processor in real time.
In some embodiments, the signal light adjustment request comprises: the judging whether the signal lamp adjustment request is reasonable or not includes:
determining the traffic jam index of N pieces of data close to the current set value of the signal lamp according to the performance table, wherein N is a positive integer and is more than or equal to 2;
determining a traffic jam index of N pieces of data close to the adjustment value of the signal lamp according to the week number, the time period and the holiday type corresponding to the signal lamp adjustment request, wherein N is a positive integer and is more than or equal to 2;
respectively screening K pieces of data from the N pieces of data with the approximate current setting value and the N pieces of data with the approximate adjustment value, wherein K is a positive integer and is more than or equal to 2;
respectively calculating a first congestion index corresponding to the current setting value and a second congestion index corresponding to the adjustment value according to the K pieces of data corresponding to the current setting value and the K pieces of data corresponding to the adjustment value;
and judging whether the first congestion index is larger than the second congestion index, if so, determining that the signal lamp adjustment request is reasonable, and if not, determining that the signal lamp adjustment request is unreasonable.
In some embodiments, the method further comprises:
the local processor sends the adjusted current setting value of the signal lamp of the traffic intersection to the urban traffic central server;
and the urban traffic central server updates the performance table according to the adjusted current setting value sent by the local server.
In some embodiments, the method further comprises:
and presetting a data threshold, and if the data volume of the performance table is larger than the data threshold, deleting data with larger difference with the current traffic data so that the data volume of the performance table does not exceed the data threshold.
In some embodiments, the method further comprises:
and determining data with larger difference with the current traffic data according to a K-Means clustering algorithm.
In a second aspect, an embodiment of the present invention provides a deep learning-based traffic control system, where the deep learning-based traffic control method is applied, and includes: the system comprises a high-definition camera, a signal lamp, a local processor and an urban traffic central server;
the high-definition camera is connected with the local processor and used for acquiring video data of a traffic intersection and sending the video data to the local processor;
the signal lamp is arranged at a traffic intersection, is connected with the local processor and is used for receiving the instruction sent by the local processor;
the local processor is connected with the high-definition camera, the signal lamp and the urban traffic central server and is used for receiving the video data acquired by the high-definition camera and controlling the current setting value of the signal lamp;
and the urban traffic central server is connected with the local processor and is used for receiving the traffic data sent by the local processor.
The beneficial effects of the embodiment of the invention are as follows: in contrast to the prior art, an embodiment of the present invention provides a deep learning-based traffic control method, which is applied to a deep learning-based traffic control system, where the traffic control system includes: the high-definition camera and the signal lamp are respectively connected with the local processor, the local processor is connected with the urban traffic central server, and the method comprises the following steps: the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase; determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode; if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data; if not, the local processor sends a signal lamp adjustment request to the urban traffic central server; and the urban traffic central server receives the signal lamp adjusting request and judges whether the signal lamp adjusting request is reasonable, if so, the local processor is controlled to adjust the current setting value of the signal lamp based on the signal lamp adjusting request. Through the mode, the traffic signal lamp monitoring system and the traffic signal lamp monitoring method can solve the technical problem that the traffic road condition cannot be accurately monitored at present, achieve balanced dispatching of the signal lamp and improve traffic strain capacity.
Drawings
One or more embodiments are illustrated in drawings corresponding to, and not limiting to, the embodiments, in which elements having the same reference number designation may be represented as similar elements, unless specifically noted, the drawings in the figures are not to scale.
Fig. 1 is a schematic view of vehicle identification of road conditions at a traffic intersection according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a vehicle trajectory identification provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a detailed manner of installing a camera according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a connection of a local processor at a traffic intersection according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a connection of a central server of urban traffic at a traffic intersection according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a four-phase system according to an embodiment of the present invention;
fig. 7 is a schematic flow chart of a deep learning-based traffic control method according to an embodiment of the present invention;
FIG. 8 is a schematic diagram of a performance table provided by an embodiment of the present invention;
fig. 9 is a detailed flowchart of a deep learning-based traffic control method according to an embodiment of the present invention;
fig. 10 is a schematic structural diagram of a deep learning-based traffic control system according to an embodiment of the present invention;
fig. 11 is a schematic structural diagram of a city traffic central server according to an embodiment of the present invention.
Detailed Description
In order to facilitate an understanding of the invention, the invention is described in more detail below with reference to the accompanying drawings and detailed description. It will be understood that when an element is referred to as being "secured to" another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may be present. The terms "vertical," "horizontal," "left," "right," and the like as used herein are for descriptive purposes only.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
In recent years, deeper and larger-scale neural networks are better and better represented in the field of image recognition, and the technology can be applied to accurate recognition of traffic road conditions. In addition, no one has proposed a better method to obtain the optimal strategy for controlling the signal lamps of multiple intersections.
Based on this, the embodiment of the present invention provides a traffic control method based on deep learning, which is applied to a traffic control system based on deep learning, and the traffic control system includes: the high-definition camera and the signal lamp are respectively connected with the local processor, and the local processor is connected with the urban traffic central server.
The high-definition camera is arranged at a traffic intersection and used for acquiring video data or image data of the traffic intersection and sending the video data or the image data to the local processor.
The signal lamp, namely a traffic lamp, is used for indicating the driving direction and the driving time of vehicles at a traffic intersection, and is connected with the local processor, and the local processor can control the setting time of the signal lamp.
Wherein, the local processor is connected with at least one signal lamp and used for controlling the set time of the signal lamp, such as: the duration of the red light, the duration of the green light, and the duration of each turn, and so on.
The urban traffic central server is connected with at least one local processor and used for receiving a signal lamp adjusting request of the local processor, and the urban traffic central server can send a signal lamp adjusting command to the local processor so that the local processor controls the setting time of the signal lamp.
In the embodiment of the invention, the urban traffic central server is used for macroscopically regulating and controlling the traffic condition of the city, the high-definition cameras of a plurality of traffic intersections acquire video data or image data of the traffic intersections and send the video data or the image data to the local processor corresponding to the traffic intersections, the local processor analyzes the video data or the image data according to the video data or the image data acquired by the high-definition cameras to obtain the traffic data of the traffic intersections, and the local processor also judges whether the traffic intersections send congestion according to the traffic data, determines whether the traffic intersections need to adjust the setting of signal lamps according to the congestion condition, and sends a signal lamp adjusting request to the urban traffic central server so that the urban traffic central server determines whether the setting of the signal lamps needs to be adjusted, the urban traffic central server can also directly send an adjusting command to the local processor so as to enable the local processor to adjust the setting time of the signal lamp.
In the embodiment of the invention, the local processor can be in an autonomous control mode or a multi-intersection combined control mode, the autonomous control mode is a mode that the local processor autonomously controls the traffic lights of the corresponding traffic intersection in real time, and the multi-intersection combined control mode is a mode that the local processor macroscopically regulates and controls the traffic intersection through the urban traffic central server.
Example one
Referring to fig. 1, fig. 1 is a schematic view illustrating vehicle identification of road conditions at a traffic intersection according to an embodiment of the present invention;
as shown in fig. 1, after the high definition camera acquires video data or image data of a traffic intersection, the local processor acquires a current image of the traffic intersection according to the video data or the image data, and identifies vehicles in the current image.
Referring to fig. 2 again, fig. 2 is a schematic view illustrating a vehicle driving track recognition according to an embodiment of the present invention;
as shown in fig. 2, after the high definition camera acquires video data or image data of a traffic intersection, the local processor acquires a current image of the traffic intersection according to the video data or the image data, identifies vehicles in the current image, and determines a driving track of each vehicle at the traffic intersection according to the video data or the image data. By tracking the track of each vehicle, the running speed and running direction of the vehicle and the number of passing vehicles are calculated.
Referring to fig. 3, fig. 3 is a schematic diagram illustrating a detailed manner of mounting a camera according to an embodiment of the present invention;
as shown in fig. 3, arrows are used to indicate the direction of flow of the phase, and dashed lines are used to indicate the stop line, for example: the green long arrows indicate the traffic direction in the first phase, the yellow long arrows indicate the traffic direction in the second phase, and the black dashed lines indicate the stop lines. The high-definition camera comprises a short-focus snapshot camera and a long-focus camera, wherein the short-focus snapshot camera is suitable for short-distance snapshot of vehicles running red light, and the long-focus camera is used for calculating traffic conditions such as the staying amount, the staying amount and the passing vehicles of each traffic intersection.
Referring to fig. 4 again, fig. 4 is a schematic connection diagram of a local processor at a traffic intersection according to an embodiment of the present invention;
as shown in fig. 4, the traffic intersection is provided with a plurality of high-definition cameras and a signal lamp, the high-definition cameras are respectively connected with a local processor, and the local processor is connected with the signal lamp.
The local processor is configured to receive video data or image data sent by the high-definition camera, and it can be understood that the high-definition camera and the local processor may communicate in a wired connection or a wireless connection manner. The high-definition camera can be a rotatable camera, and the local processor can also control the high-definition camera to rotate according to the traffic condition of the traffic intersection, so that the high-definition camera can better acquire video data or image data of the traffic intersection.
Wherein, the signal lamp is a traffic signal lamp, the signal lamp is connected with the local processor, and the local processor can control the setting of the signal lamp, such as: duration of red light, flashing time, duration of green light, flashing time, and duration of each turn.
Referring to fig. 5 again, fig. 5 is a schematic connection diagram of an urban traffic central server at a traffic intersection according to an embodiment of the present invention;
as shown in fig. 5, a plurality of local processors are respectively connected to the urban traffic central server, and the local processors identify video data or image data acquired by the high-definition cameras and then acquire corresponding traffic data, where the traffic data includes: and the traffic data is sent to the urban traffic central server, wherein the urban traffic central server receives and stores the traffic data, and correspondingly stores the traffic data and the corresponding traffic intersection so that the urban traffic central server can determine whether the signal lamp setting of the traffic intersection is reasonable or not according to the historical data of the traffic intersection.
It will be appreciated that each traffic intersection corresponds to a unique local processor, and different local processors are distributed in different places of a city, so that the local processors and the city traffic central server are connected via a network, such as: the system comprises a 4G network, optical fibers and a WIFI network, wherein the local processors are connected with cameras or signal lamps through WIFI, and also can be connected through networks such as the 4G network and the optical fibers, and the local processors respectively send traffic data of each traffic intersection acquired by the local processors to the urban traffic central server through wireless signals.
Each traffic intersection can correspond to a plurality of phases, that is, each signal lamp corresponds to a plurality of phases, and the phases of the signal lamps are a plurality of directions given to corresponding passing aiming at traffic flows in different directions. There are as many phases as there are variations of the signal light. Referring to fig. 6, fig. 6 is a schematic diagram of a four-phase system according to an embodiment of the present invention; as shown in fig. 6, the four-phase system includes a first phase, a second phase, a third phase, and a fourth phase, each of which indicates that the vehicle is drivable in the direction indicated by the arrow in the figure.
It will be appreciated that the phase system in embodiments of the present invention may also be other multi-phase systems, such as: two-phase systems, six-phase systems, eight-phase systems, and so on.
Referring to fig. 7, fig. 7 is a schematic flow chart illustrating a deep learning-based traffic control method according to an embodiment of the present invention;
as shown in fig. 7, the deep learning-based traffic control method is applied to a deep learning-based traffic control system, and the traffic control system includes: the high-definition camera and the signal lamp are respectively connected with the local processor, the local processor is connected with the urban traffic central server, and the method comprises the following steps:
step S10: the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase;
specifically, the high-definition cameras are arranged at traffic intersections, each traffic intersection is provided with a plurality of high-definition cameras for acquiring video data of the traffic intersections in different directions and at different angles, the high-definition cameras are connected with a local processor, and all the high-definition cameras at each traffic intersection are connected with the same local processor, so that the local processor can perform corresponding processing according to the video data acquired by the high-definition cameras. The high-definition camera sends video data of the traffic intersection to a corresponding local processor, and the local processor processes the video data through a deep neural network and determines traffic data according to the video data. Wherein the deep neural network model may employ yolov3, SSD, and so on. Detecting a vehicle in a video through the deep neural network model, and then tracking by using an LK optical flow method to further obtain traffic data, wherein the traffic data comprises: the amount of the traffic flow, the traffic jam index, the traffic capacity, the traffic flow, the red light running amount and the like of each phase. In an embodiment of the present invention, the traffic data further includes: hold up, traffic flow, congestion index, and traffic capacity.
Specifically, the processing the video data through the deep neural network to determine the traffic data includes:
determining the number of vehicles passing through a road section corresponding to a certain phase in unit time as the traffic flow of the phase; the traffic flow is determined by calculating the number of vehicles passing through a certain road section in unit time, and the unit of the traffic flow is PCU/h, wherein PCU (passsenger car unit) is called standard passenger car equivalent. It can be understood that the large-volume vehicle and the small passenger car have different volumes, and therefore, the conversion is performed through a corresponding conversion standard, and the large-volume vehicle is converted into a standard small passenger car equivalent corresponding to the small passenger car. For example: because the volume of the truck and the bus is large, the large-volume vehicles such as the truck and the bus can be calculated according to 2 to 6 passenger cars.
And recording the number of vehicles with red lights on each road of the traffic intersection according to the video data processing calculation of the traffic intersection as real-time parking amount, wherein the data is obviously increased continuously under the red light. The number of vehicles which are in green light on each road of the intersection but do not pass through the stop line is recorded as real time lag, and the data is continuously reduced under the condition that the green light does not block the vehicle.
The transit time refers to the approximate time that all vehicles in a certain traffic flow direction take to pass through the intersection. For example: intersections in the west-east direction have many cars at the beginning of the green light:
if t1After second, the lamp is switched to red, when n is available1Vehicle passing, but still n2Vehicle detention crossing (n)2>M), the transit time in the west-east direction can be determined at this point as t1+t1*(n2/n1) And second. When the number of parked vehicles is less than a certain value m, n can be considered to be20. Empirically, m can be made 0.05 n2. There may be new vehicles in the parked vehicle, but it is negligible;
if t2(t2<Green time) of seconds, the time of staying of the vehicle is less than m, and the passing time in the west-east direction is t2；
And (4) representing the traffic jam index of a certain phase by the average value of the ratios of the multiple groups of traffic time of the phase to the corresponding green light time. It is understood that a congestion index of less than 1 indicates that vehicles with a green time sufficiently waiting substantially pass, and a congestion index of greater than 1 indicates that some vehicles and the like are in the stop line after the end of the green time. For example, a traffic jam index of 2 indicates that the green time needs to be doubled to allow all vehicles in the stop line to pass.
And determining the maximum number of vehicles which continuously pass through the stop line corresponding to a certain phase in one continuous green light time as the traffic capacity of the phase. In the embodiment of the invention, the unit is PCU/h, and the maximum number of vehicles which can continuously pass through a stop line by a fleet on an intersection approach in one continuous green light time is referred to. Preferentially, the traffic capacity of the phase can be better determined by acquiring the maximum value in a period of time when the traffic jam condition occurs in the period of time and taking the maximum value in the period of time as the traffic capacity of the phase.
Step S20: determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode;
specifically, the local processor comprises two working modes, namely an autonomous control mode and a multi-interface joint control mode; the automatic control mode is a mode that the local processor automatically controls the traffic lights of the corresponding traffic intersection in real time, and the automatic real-time control means that the local processor analyzes the self-regulation signal lights without applying; the multi-intersection combined control mode is a mode that the local processor macroscopically regulates and controls the traffic intersection through the urban traffic central server, the multi-intersection combined control is to locally provide a strategy, the strategy is uploaded to the urban central server for auditing, and the strategy is allowed to be executed after the auditing is passed. The working mode of the local processor can be set by the urban traffic central server or can be set manually. The urban traffic central server can switch the working mode of the local processor according to the traffic condition of the traffic intersection corresponding to the local processor.
And the local processor switches the signal lamps according to the real-time retention amount and the real-time retention amount of each phase and the duration time of the traffic lights if the local processor is in the autonomous real-time control mode. It should be noted that such signal lights only begin to count down the first few seconds of the switch, as such signal lights do not change periodically, but rather change depending on the number of vehicles waiting on each road. For example, fig. 3 shows a two-phase system with four roads, three roads, north and south, east and south, having stop lines, which require red lights. The south and north roads belong to phase one, and the east and west roads belong to phase two. The green time of phase one at this time lasts for t1Second, real-time hold-up of n1Red light time of phase two lasts for t2Second, real time hold-up of n2(ii) a If w1*n1–w2w*t1t1<w3w3*n2n2+w44*t2And t1Greater than a certain minimum value T1Then switch from green to red, w is the weight parameter, T1These several values need to be set empirically for the green minimum. The strategy of how phase two is switched from green to red is the same as phase one. When there are more than two phases, n2And t2Is (w)33*n22+w44*t2) The group having the largest value.
Step S30: if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data;
in the embodiment of the present invention, the current setting value of the signal lamp refers to a traffic light setting, the current setting value of the signal lamp is vector data, that is, the current setting value of the signal lamp is a vector with a plurality of parameters, specifically, the length of the vector is determined according to the phase number of the traffic intersection, for example: if n phases exist at the traffic intersection, the vector length of the vector is 2n +1, n is a positive integer and n is larger than or equal to 2, wherein the first parameter of the vector indicates that the flashing starts several seconds after a green light, or the yellow light is cut off, each 2 parameters at the back are in a group, the former parameter indicates the offset time of the start of a certain phase, the latter parameter indicates the duration of the phase, and the like, so that the 2n parameters can indicate the offset time and the duration of the start of all the phases.
The urban traffic central server verifies whether the signal lamp adjustment request is reasonable or not according to historical traffic data, and intervenes in the adjustment of the local processor; if the local processor is in the automatic control mode, the signal lamp is automatically controlled according to the retention of each road junction, and the vehicle is released in real time.
In an embodiment of the present invention, the traffic data includes: the local processor analyzes whether the current set value of the signal lamp is reasonable or not according to the traffic data, and the method comprises the following steps:
and judging whether the traffic jam index of a certain phase of the traffic intersection is far larger than the traffic jam indexes of other phases, if so, indicating that the current setting value of a signal lamp of the traffic intersection is unreasonable, and at the moment, sending a signal lamp adjustment request to the urban traffic central server by the local processor.
Step S40: if not, the local processor sends a signal lamp adjustment request to the urban traffic central server;
specifically, the local processor only sends statistical data to the urban traffic central server, and the signal lamp adjustment request includes: an adjustment value for a signal light, the adjustment value for the signal light representing a suggested setting value for the signal light. Wherein, the local processor is connected with the urban traffic central server through a wireless or wired network, such as: 4G network, 5G network, fiber, and so on, the local processor interacting with the urban traffic central server based on a wireless communication protocol. The signal lamp adjustment request is a data frame, the data frame includes a position of a signal lamp, a serial number, a position of a local processor, an ID of the local processor, a communication address, and an adjustment value of the signal lamp, and the communication address of the local processor may be an IP address of the local processor.
Step S50: and the urban traffic central server receives the signal lamp adjusting request and judges whether the signal lamp adjusting request is reasonable, if so, the local processor is controlled to adjust the current setting value of the signal lamp based on the signal lamp adjusting request.
In an embodiment of the present invention, the local processor periodically sends the traffic data to the urban traffic central server, and the urban traffic central server receives the traffic data sent by the local processor, where the traffic data includes: the congestion index and the traffic capacity of each phase.
The urban traffic central server establishes and initializes a plurality of performance tables of a plurality of traffic intersections according to the traffic data, wherein each traffic intersection corresponds to each performance table one by one, the performance tables are used for correspondingly storing the data of the traffic intersections, and the data comprises the week number, the time period, the holidays, the traffic light setting, the traffic jam index and the traffic index. Specifically, the urban traffic central server establishes a performance table for each traffic intersection of the city, each traffic intersection corresponds to a unique performance table, the local processor sends traffic data of the traffic intersection acquired by the local processor to the urban traffic central server, and the urban traffic central server establishes the performance tables of the traffic intersections according to the traffic data of the traffic intersections. It can be understood that each local processor corresponds to at least one traffic intersection, the local processor sends the traffic data of a plurality of traffic intersections corresponding to the local processor to the urban traffic central server, and the urban traffic central server establishes a performance table corresponding to each traffic intersection according to the traffic data of the plurality of traffic intersections.
Referring to fig. 8 again, fig. 8 is a schematic diagram of a performance table according to an embodiment of the invention;
as shown in fig. 8, the performance table includes: the traffic jam monitoring system comprises a plurality of parameters such as the week number, the time period, the holidays, the traffic light setting, the traffic jam index and the traffic capacity, wherein the parameters are correspondingly stored. Where the first term represents the day of the week and the second term is the time period. There are several time periods for how many times a day the signal lamp of a certain traffic intersection is switched. The third item indicates whether the day is a holiday, which holiday is a holiday, for example, a holiday is represented by 0, a holiday is represented by a letter plus a number, for example, gq-1 is represented by the first day of the national festival, and gq-7 is represented by the 7 th day of the national festival. The first day of Qingming is denoted by qm-1. The fourth item is signal lamp setting of the current intersection, which is vector data, and if the intersection has n phases, the vector length is 2n + 1. The first element of the vector indicates that a few seconds after a green light starts flashing, or a yellow light is cut. The next 2 elements are grouped, the first representing the offset time of the start of the phase, the next representing the duration of the phase, and so on. The fifth item represents the congestion condition of the intersection, and is also vector data, and each element represents the traffic congestion index of the corresponding phase. The sixth term represents the traffic capacity of each phase of the intersection, and is vector data, and each element represents the maximum number of vehicles which can continuously pass through a stop line in an intersection entrance lane within one continuous green light time corresponding to the phase, and the unit is PCU/h, wherein the PCU is called standard passenger car equivalent.
Wherein the signal lamp adjustment request sent by the local processor to the urban traffic central server comprises: the judging whether the signal lamp adjustment request is reasonable or not includes:
determining the traffic jam index of N pieces of data close to the current set value of the signal lamp according to the performance table, wherein N is a positive integer and is more than or equal to 2;
determining a traffic jam index of N pieces of data close to the adjustment value of the signal lamp according to the week number, the time period and the holiday type corresponding to the signal lamp adjustment request, wherein N is a positive integer and is more than or equal to 2;
respectively screening K pieces of data from the N pieces of data with the approximate current setting value and the N pieces of data with the approximate adjustment value, wherein K is a positive integer and is more than or equal to 2;
respectively calculating a first congestion index corresponding to the current setting value and a second congestion index corresponding to the adjustment value according to the K pieces of data corresponding to the current setting value and the K pieces of data corresponding to the adjustment value;
and calculating the first congestion index and the second congestion index of other intersections jointly controlled by the intersection by adopting the same method to obtain the average value of the first congestion index and the second congestion index. And judging whether the average value of the first congestion index is larger than the average value of the second congestion index, if so, determining that the signal lamp adjustment request is reasonable, and if not, determining that the signal lamp adjustment request is unreasonable.
Specifically, the calculating the first congestion index corresponding to the current setting value includes: calculating a weighted average (traffic capacity as weight) of the traffic congestion indexes of the K pieces of data corresponding to the current setting value, taking the weighted average as a first congestion index, and calculating a second congestion index corresponding to the adjustment value, including: and calculating a weighted average value of the traffic congestion indexes of the K pieces of data corresponding to the adjustment value, and taking the weighted average value as a second congestion index.
Specifically, two types of data need to be extracted from the performance table of the traffic intersection, the first type of data is data with the same or similar data of the current setting value, and specifically, the first type of data is N pieces of data with the same or similar types of the current time period, the current week number and the current holiday. The N data sets are further required to be screened for K data sets, and the K data sets are obtained through screening by a corresponding algorithm, for example, at 10 am on saturday, intersection a is very congested from 9 am to 10 am, but in the last week, the time point of the last week is not congested. It is clear that the current traffic situation should not be described according to the statistics of this time. Therefore, K pieces of data (K is a positive integer and K is more than or equal to 2) which are the same as the current situation need to be found to calculate the average value, and the current traffic situation can be described through the average value, so that the method can be more objective.
The second type of data is N pieces of data corresponding to an adjustment value, where the data corresponding to the adjustment value refers to data of a time period in the performance table after a period of time elapses after the setting value of the signal lamp is adjusted, for example: and adjusting the signal lamp in a certain time period, and taking the data corresponding to the next time period of the time period as the data corresponding to the adjustment value. Similarly, the data corresponding to the adjustment value also needs to use the average value to characterize the traffic condition. For example, at 10 o' clock on saturday, it is known that 9 am to 10 am on saturday is very blocked, if the signal light is adjusted, it is obvious that the signal light is effective after 10 am, and the historical data after 10 am needs to be taken as reference, that is, K pieces of data are needed to calculate the average value. From this average, we can get how the traffic conditions change under the new signal light setting. For example, in the last week, the 10 o 'clock to 11 o' clock signal settings of the last week are close to the new settings proposed by the local processor, and the average of these several data can be used to predict the traffic performance in the new configuration of Saturday. The traffic condition is represented by the average value of the data corresponding to the adjustment value, and the method is more objective.
In an embodiment of the present invention, the method further comprises: and the urban traffic central server updates the performance table in real time according to the traffic data sent by the local processor in real time. Specifically, the local processor sends the traffic data of at least one traffic intersection corresponding to the local processor to the urban traffic central server at regular intervals, and the urban traffic central server updates the performance table of the at least one traffic intersection in real time according to the traffic data of the at least one traffic intersection sent by the local processor. Specifically, the updating the performance table of the traffic intersection includes: and updating the corresponding relation in the performance table of the traffic intersection, such as: and updating the week number, the time period, the traffic light setting corresponding to the holidays, the traffic jam index and the traffic capacity in the performance table.
In an embodiment of the present invention, the method further comprises: the local processor sends the adjusted current setting value of the signal lamp of the traffic intersection to the urban traffic central server; the urban traffic central server updates the performance table according to the adjusted current setting value sent by the local server, for example: and updating the traffic light setting in the performance table.
And the urban traffic central server receives the signal lamp adjusting request, judges whether the signal lamp adjusting request is reasonable or not, and controls the local processor to adjust the current setting value of the signal lamp based on the signal lamp adjusting request if the signal lamp adjusting request is reasonable. In an embodiment of the present invention, the method further comprises: and the local processor receives an adjusting command sent by the urban traffic central server and adjusts the current setting value of the signal lamp of the traffic intersection according to the adjusting command. In the embodiment of the invention, as a city corresponds to a plurality of traffic intersections, as the local processors send more and more traffic data to the urban traffic central server, the data of the performance table corresponding to the traffic intersections will be more and more, the control basis of the urban traffic central server will be more and more, a large amount of memory will be occupied, and the processing speed of the urban traffic central server will be correspondingly slow, and the signal lamp adjustment requests sent by the local processors cannot be responded in time, so that the performance table needs to be maintained, and the data amount of the performance table is controlled. Specifically, by presetting a data threshold, if the data volume of the performance table is greater than the data threshold, deleting the data with a larger difference from the current traffic data, so that the data volume of the performance table does not exceed the data threshold. For example: presetting the data threshold value as 1000000, and if the data quantity of the performance table is larger than 1000000, deleting the data with larger difference with the current traffic data so as to enable the data quantity of the performance table not to exceed the data threshold value. Specifically, the deleting the data having a larger difference from the current traffic data includes: and determining data with larger difference with the current traffic data according to a K-Means clustering algorithm, and deleting the data with larger difference with the current traffic data. It will be appreciated that the data in the table will grow larger and larger over time, and that the basis of control by the urban traffic central server will increase. But the importance of each piece of data has also changed, and data that is closest to current traffic data will be accessed frequently, but data that differs greatly from current traffic will be accessed rarely. At this time, the K-mean method can be used for clustering, finding out data with large differences and deleting the data, so as to maintain the quantity of the performance table. By maintaining the quantity of the performance table, the memory occupation of the urban traffic central server is reduced, and the response speed of the urban traffic central server is improved.
Referring to fig. 9, fig. 9 is a detailed flowchart of a deep learning-based traffic control method according to an embodiment of the present invention;
as shown in fig. 9, the deep learning-based traffic control method includes:
step S10: the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase;
specifically, the number of vehicles with red lights on each road of the traffic intersection is recorded as the real-time parking amount, and the data is obviously increased continuously under the red light; and recording the number of vehicles which are in green light but do not pass through the stop line on each road of the traffic intersection as real time lag, wherein the data is continuously reduced under the condition that the green light does not block the vehicle.
The transit time refers to the approximate time that all vehicles in a certain traffic flow direction take to pass through the intersection. For example, a west-east intersection has many cars at the beginning of a green light:
if t1After second, the lamp is switched to red, when n is available1Vehicle passing, but still n2Vehicle detention crossing (n)2>M), the transit time in the west-east direction can be determined at this point as t1+t1*(n2/n1) And second. When the number of parked vehicles is less than a certain value m, n can be considered to be20. Empirically, m can be made 0.05 n2. There may be new vehicles in the parked vehicle, but it is negligible;
if t2(t2<Green time) of seconds, the time of staying of the vehicle is less than m, and the passing time in the west-east direction is t2；
And (4) representing the traffic jam index of a certain phase by the average value of the ratios of the multiple groups of traffic time of the phase to the corresponding green light time.
And determining the maximum number of vehicles which continuously pass through the stop line corresponding to a certain phase in one continuous green light time as the traffic capacity of the phase. In the embodiment of the invention, the unit is PCU/h, and the maximum number of vehicles which can continuously pass through a stop line by a fleet on an intersection approach in one continuous green light time is referred to. The maximum value in a period of time is required to be obtained, and the period of time is counted only when the traffic jam occurs.
And taking the ratio of the passing vehicles in a certain phase unit time to the passing capacity of the phase as the passing index of the phase.
Step S20: determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode;
step S30: if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data; if yes, go to step S31; if not, go to step S32;
step S31: maintaining a current setting value of the signal lamp;
step S40: the local processor sends a signal lamp adjustment request to the urban traffic central server;
step S50: the urban traffic central server receives the signal lamp adjustment request and judges whether the signal lamp adjustment request is reasonable or not; if yes, go to step S51; if not, go to step S52;
step S51: controlling the local processor to adjust the current setting value of the signal lamp based on the signal lamp adjustment request;
step S52: maintaining a current setting value of the signal lamp;
in an embodiment of the present invention, the method further comprises: and determining the current setting value of the signal lamp in a manual intervention mode.
In an embodiment of the present invention, a deep learning-based traffic control method is provided, which is applied to a deep learning-based traffic control system, and the traffic control system includes: the high-definition camera and the signal lamp are respectively connected with the local processor, the local processor is connected with the urban traffic central server, and the method comprises the following steps: the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase; determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode; if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data; if not, the local processor sends a signal lamp adjustment request to the urban traffic central server; and the urban traffic central server receives the signal lamp adjusting request and judges whether the signal lamp adjusting request is reasonable, if so, the local processor is controlled to adjust the current setting value of the signal lamp based on the signal lamp adjusting request. Through the mode, the traffic signal lamp monitoring system and the traffic signal lamp monitoring method can solve the technical problem that the traffic road condition cannot be accurately monitored at present, achieve balanced dispatching of the signal lamp and improve traffic strain capacity.
Example two
Referring to fig. 10, fig. 10 is a schematic structural diagram of a deep learning-based traffic control system according to an embodiment of the present invention;
as shown in fig. 10, the deep learning based traffic control system, to which the deep learning based traffic control method described in the first embodiment is applied, includes: the system comprises a high-definition camera 91, a signal lamp 92, a local processor 93 and an urban traffic central server 94, wherein the high-definition camera 91 is connected with the local processor 93, the signal lamp 92 is connected with the local processor 93, and the local processor 93 is connected with the urban traffic central server 94;
specifically, the high-definition camera 91 is connected to the local processor 93, and is configured to acquire video data or image data of a traffic intersection and send the video data or image data to the local processor 93; it can be understood that the high definition cameras 91 are disposed at traffic intersections in a city, each of the traffic intersections may be provided with a plurality of high definition cameras 91, and the plurality of high definition cameras 91 are used for omni-directionally acquiring road condition information of the traffic intersection, such as: vehicles, pedestrians, and so forth.
Specifically, the signal lamp 92 is disposed at a traffic intersection, connected to the local processor 93, and configured to receive an instruction sent by the local processor 93; it will be appreciated that the signal lamp includes a plurality of phases, each phase being used to indicate the direction of travel of a vehicle in a different direction so that the vehicle travels correctly according to the phase of the signal lamp. Wherein the signal lamp 92 communicates with the local processor 93 by means of wired connection or wireless connection.
Specifically, the local processor 93 is connected to the high-definition camera 91, the signal lamp 92 and the urban traffic central server 94, and is configured to receive video data or image data acquired by the high-definition camera and control a current setting value of the signal lamp; the local processor 93 may send a signal light adjustment command to the signal lights 92 to adjust the current settings of the signal lights. In the embodiment of the present invention, the local processor 93 is further connected to the urban traffic central server 94, and configured to send a signal lamp adjustment request to the urban traffic central server 94, so that the urban traffic central server 94 receives the signal lamp adjustment request, and determines whether the signal lamp adjustment request is reasonable according to the signal lamp adjustment request, and if so, controls the local processor to adjust the current setting value of the signal lamp based on the signal lamp adjustment request. Wherein the local processor 93 and the urban traffic central server 94 communicate wirelessly, for example: 4G, WIFI, etc. The local processor 93 includes: and the deep neural network unit is used for identifying vehicles, pedestrians and motion tracks of the vehicles and the pedestrians in all directions of the road. The local processor 93 further comprises: and the visual database is used for storing the video data or the image data acquired by the high-definition camera 91.
Specifically, the urban traffic central server 94 is connected to the local processor 93, and is configured to receive the traffic data sent by the local processor 93. And, the urban traffic central server 94 is further configured to receive the signal lamp adjustment request sent by the local processor 93.
In the embodiment of the present invention, the urban traffic central server 94 includes: a first wireless communication module, the local processor 93 comprising: a second wireless communication module, wherein the first wireless communication module is in communication connection with the second wireless communication module and is used for data transmission between the urban traffic central server 94 and the local processor 93. The first wireless communication module and the second wireless communication module are both Bluetooth modules or WIFI modules or 4G modules.
In the present embodiment, the urban traffic central server 94 exists in various forms, including but not limited to:
(1) tower server
The general tower server chassis is almost as large as the commonly used PC chassis, while the large tower chassis is much larger, and the overall dimension is not a fixed standard.
(2) Rack-mounted server
Rack-mounted servers are a type of server that has a standard width of 19 inch racks, with a height of from 1U to several U, due to the dense deployment of the enterprise. Placing servers on racks not only facilitates routine maintenance and management, but also may avoid unexpected failures. First, placing the server does not take up too much space. The rack servers are arranged in the rack in order, and no space is wasted. Secondly, the connecting wires and the like can be neatly stored in the rack. The power line, the LAN line and the like can be distributed in the cabinet, so that the connection lines accumulated on the ground can be reduced, and the accidents such as the electric wire kicking off by feet can be prevented. The specified dimensions are the width (48.26cm ═ 19 inches) and height (multiples of 4.445 cm) of the server. Because of its 19 inch width, a rack that meets this specification is sometimes referred to as a "19 inch rack".
(3) Blade server
A blade server is a HAHD (High Availability High Density) low cost server platform designed specifically for the application specific industry and High Density computer environment, where each "blade" is actually a system motherboard, similar to an individual server. In this mode, each motherboard runs its own system, serving a designated group of different users, without any relationship to each other. Although system software may be used to group these motherboards into a server cluster. In the cluster mode, all motherboards can be connected to provide a high-speed network environment, and resources can be shared to serve the same user group.
Referring to fig. 11 again, fig. 11 is a schematic structural diagram of an urban traffic central server according to an embodiment of the present invention;
as shown in fig. 11, the urban traffic central server 94 includes one or more processors 941 and memory 942. In fig. 11, one processor 941 is taken as an example.
The processor 941 and the memory 942 may be connected by a bus or other means, such as a bus connection in fig. 11.
The memory 942, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The processor 941 performs various functional applications and data processing by executing nonvolatile software programs, instructions, and modules stored in the memory 942. The processor 941 is configured to determine whether the signal lamp adjustment request is reasonable.
The memory 942 may include high speed random access memory and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some embodiments, the memory 942 optionally includes memory located remotely from the processor 941, which may be connected to the processor 941 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof. The memory 942 is configured to store traffic data of each traffic intersection of the city and a performance table corresponding to each traffic intersection.
In the embodiment of the present invention, the urban traffic central server 94 further includes: a first wireless communication module (not shown) for communicatively coupling the local processor, such as: receiving traffic data sent by the local processor, receiving a semaphore adjustment request sent by the local processor, or sending a semaphore adjustment command to the local processor, and so on.
In an embodiment of the present invention, by providing a deep learning-based traffic control system, applying the deep learning-based traffic control method according to the first embodiment, the deep learning-based traffic control system includes: the system comprises a high-definition camera, a signal lamp, a local processor and an urban traffic central server; the high-definition camera is connected with the local processor and is used for acquiring video data of a traffic intersection and sending the video data to the local processor; the signal lamp is arranged at a traffic intersection, is connected with the local processor and is used for receiving the instruction sent by the local processor; the local processor is connected with the high-definition camera, the signal lamp and the urban traffic central server and is used for receiving the video data acquired by the high-definition camera and controlling the current setting value of the signal lamp; and the urban traffic central server is connected with the local processor and is used for receiving the traffic data sent by the local processor. Through the mode, the embodiment of the invention can solve the technical problem that the traffic road condition cannot be accurately monitored at present, realizes the balanced dispatching of the signal lamps and improves the traffic strain capacity.
It should be noted that the description of the present invention and the accompanying drawings illustrate preferred embodiments of the present invention, but the present invention may be embodied in many different forms and is not limited to the embodiments described in the present specification, which are provided as additional limitations to the present invention, and the present invention is provided for understanding the present disclosure more fully. Furthermore, the above-mentioned technical features are combined with each other to form various embodiments which are not listed above, and all of them are regarded as the scope of the present invention described in the specification; further, modifications and variations will occur to those skilled in the art in light of the foregoing description, and it is intended to cover all such modifications and variations as fall within the true spirit and scope of the invention as defined by the appended claims.
Claims (8)
1. A deep learning-based traffic control method is applied to a deep learning-based traffic control system, and the traffic control system comprises: the high-definition camera and the signal lamp are respectively connected with the local processor, the local processor is connected with the urban traffic central server, and the method comprises the following steps:
the local processor acquires video data of a traffic intersection, processes the video data through a deep neural network and determines traffic data, wherein the traffic data comprises: traffic flow, retention, traffic jam index and traffic capacity of each phase;
determining an operating mode of the local processor, the operating mode comprising: an autonomous control mode and a multi-interface joint control mode;
if the local processor is in a multi-interface combined control mode, the local processor analyzes whether the current setting value of the signal lamp is reasonable or not according to the traffic data;
if not, the local processor sends the traffic data and a signal lamp adjustment request to the urban traffic central server;
the urban traffic central server receives traffic data sent by the local processor;
the urban traffic central server establishes and initializes a plurality of performance tables of a plurality of traffic intersections according to the traffic data, wherein each traffic intersection corresponds to each performance table one by one, the performance tables are used for correspondingly storing data of the traffic intersections, and the data comprises the week number, the time period, the holidays, the traffic light setting, the traffic jam index and the traffic capacity;
determining the traffic jam index of N pieces of data close to the current set value of the signal lamp according to the performance table, wherein N is a positive integer and is more than or equal to 2;
determining a traffic jam index of N pieces of data close to the adjustment value of the signal lamp according to the week number, the time period and the holiday type corresponding to the signal lamp adjustment request, wherein N is a positive integer and is more than or equal to 2;
respectively screening K pieces of data from the N pieces of data with the approximate current setting value and the N pieces of data with the approximate adjustment value, wherein K is a positive integer and is more than or equal to 2;
respectively calculating a first congestion index corresponding to the current setting value and a second congestion index corresponding to the adjustment value according to the K pieces of data corresponding to the current setting value and the K pieces of data corresponding to the adjustment value;
and judging whether the first congestion index is larger than the second congestion index, if so, determining that the signal lamp adjusting request is reasonable, controlling the local processor to adjust the current setting value of the signal lamp based on the signal lamp adjusting request, and if not, determining that the signal lamp adjusting request is unreasonable.
2. The traffic control method according to claim 1, characterized in that the method further comprises:
and if the local processor is in the autonomous control mode, adjusting the signal lamps in real time according to the real-time reserves and the real-time reserves of each traffic intersection.
3. The traffic control method of claim 1, wherein the processing the video data through a deep neural network to determine traffic data comprises:
recording the number of vehicles waiting for red lights on each road of the traffic intersection as real-time parking amount according to the video data of the traffic intersection; recording the number of vehicles which are at green lights but do not pass through a stop line on each road of the traffic intersection as real time lag allowance;
recording the approximate time required by all vehicles in a certain traffic flow direction to pass through the traffic intersection as the passing time; the average value of the ratio of the multiple groups of traffic time of a certain phase to the corresponding green light time represents the traffic jam index of the phase;
determining the number of vehicles passing through a road section corresponding to a certain phase in unit time as the traffic flow of the phase;
determining the maximum number of vehicles which continuously pass through a stop line corresponding to a certain phase within one continuous green light time as the traffic capacity of the phase; and taking the ratio of the number of passing vehicles in a unit time of a certain phase to the passing capacity of the phase as the passing index of the phase.
4. The traffic control method according to claim 1, characterized in that the method further comprises:
and the urban traffic central server updates the performance table in real time according to the traffic data sent by the local processor in real time.
5. The traffic control method according to claim 1, characterized in that the method further comprises:
the local processor sends the adjusted current setting value of the signal lamp of the traffic intersection to the urban traffic central server;
and the urban traffic central server updates the performance table according to the adjusted current setting value sent by the local server.
6. The traffic control method according to claim 5, characterized in that the method further comprises:
and presetting a data threshold, and if the data volume of the performance table is larger than the data threshold, deleting data with larger difference with the current traffic data so that the data volume of the performance table does not exceed the data threshold.
7. The traffic control method according to claim 6, characterized in that the method further comprises:
and determining data with larger difference with the current traffic data according to a K-Means clustering algorithm.
8. A deep learning based traffic control system, characterized in that the deep learning based traffic control method according to any one of claims 1-7 is applied, comprising: the system comprises a high-definition camera, a signal lamp, a local processor and an urban traffic central server;
the high-definition camera is connected with the local processor and used for acquiring video data of a traffic intersection and sending the video data to the local processor;
the signal lamp is arranged at a traffic intersection, is connected with the local processor and is used for receiving the instruction sent by the local processor;
the local processor is connected with the high-definition camera, the signal lamp and the urban traffic central server and is used for receiving the video data acquired by the high-definition camera and controlling the current setting value of the signal lamp;
and the urban traffic central server is connected with the local processor and is used for receiving the traffic data sent by the local processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910078907.2A CN109697867B (en) | 2019-01-28 | 2019-01-28 | Deep learning-based traffic control method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910078907.2A CN109697867B (en) | 2019-01-28 | 2019-01-28 | Deep learning-based traffic control method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109697867A CN109697867A (en) | 2019-04-30 |
| CN109697867B true CN109697867B (en) | 2020-10-20 |
Family
ID=66234396
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910078907.2A Active CN109697867B (en) | 2019-01-28 | 2019-01-28 | Deep learning-based traffic control method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109697867B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109558969A (en) * | 2018-11-07 | 2019-04-02 | 南京邮电大学 | A kind of VANETs car accident risk forecast model based on AdaBoost-SO |
| CN110491146B (en) * | 2019-08-21 | 2020-08-21 | 浙江工业大学 | Deep learning-based traffic signal control scheme real-time recommendation method |
| CN110473412A (en) * | 2019-09-12 | 2019-11-19 | 杭州京安交通工程设施有限公司 | A kind of controlling system of traffic light |
| CN112185139B (en) * | 2020-09-23 | 2022-03-22 | 上海勤电信息科技有限公司 | City management method and system based on intelligent comprehensive pole |
| CN113095220B (en) * | 2021-04-12 | 2023-04-18 | 浙江吉利控股集团有限公司 | Traffic signal lamp digital identification method and digital identification system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101281685A (en) * | 2008-01-30 | 2008-10-08 | 吉林大学 | Coordination control method for area mixed traffic self-adaption signal |
| JP2009146138A (en) * | 2007-12-13 | 2009-07-02 | Sumitomo Electric Ind Ltd | Traffic signal controller, program, and method |
| CN102542818A (en) * | 2012-01-13 | 2012-07-04 | 吉林大学 | Organic-calculation-based coordination control method for traffic signal of zone boundary |
| JP5104344B2 (en) * | 2008-01-28 | 2012-12-19 | 住友電気工業株式会社 | Traffic signal control device, computer program, and traffic signal control method |
| CN102867409A (en) * | 2012-09-17 | 2013-01-09 | 李伟 | Road traffic cooperative control method for urban central area |
| US8830086B2 (en) * | 2011-10-28 | 2014-09-09 | International Business Machines Corporation | Adjusting traffic lights |
| CN105513379A (en) * | 2015-12-18 | 2016-04-20 | 天津通翔智能交通系统有限公司 | Multi-target traffic signal networked joint control method |
| CN108564786A (en) * | 2018-05-14 | 2018-09-21 | 苏州思在科技有限公司 | A kind of method and system of the urban traffic signal control based on cloud computing |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6539300B2 (en) * | 2001-07-10 | 2003-03-25 | Makor Issues And Rights Ltd. | Method for regional system wide optimal signal timing for traffic control based on wireless phone networks |
| KR20130015802A (en) * | 2011-08-05 | 2013-02-14 | 이흥수 | Traffic signal controlling device and method using identification information with hierarchy structure |
| CN104732768B (en) * | 2015-04-08 | 2016-05-04 | 浙江传媒学院 | Pass through collecting method at crossing, signal lamp place |
| CN105761517B (en) * | 2016-05-18 | 2017-11-17 | 杭州智诚惠通科技有限公司 | A kind of traffic signal timing method |
| CN106448193A (en) * | 2016-10-28 | 2017-02-22 | 杭州利艾智能科技有限公司 | Urban traffic control system and method |
| CN106875702B (en) * | 2017-04-11 | 2017-09-29 | 冀嘉澍 | A kind of crossroad access lamp control method based on Internet of Things |
-
2019
- 2019-01-28 CN CN201910078907.2A patent/CN109697867B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009146138A (en) * | 2007-12-13 | 2009-07-02 | Sumitomo Electric Ind Ltd | Traffic signal controller, program, and method |
| JP5104344B2 (en) * | 2008-01-28 | 2012-12-19 | 住友電気工業株式会社 | Traffic signal control device, computer program, and traffic signal control method |
| CN101281685A (en) * | 2008-01-30 | 2008-10-08 | 吉林大学 | Coordination control method for area mixed traffic self-adaption signal |
| US8830086B2 (en) * | 2011-10-28 | 2014-09-09 | International Business Machines Corporation | Adjusting traffic lights |
| CN102542818A (en) * | 2012-01-13 | 2012-07-04 | 吉林大学 | Organic-calculation-based coordination control method for traffic signal of zone boundary |
| CN102867409A (en) * | 2012-09-17 | 2013-01-09 | 李伟 | Road traffic cooperative control method for urban central area |
| CN105513379A (en) * | 2015-12-18 | 2016-04-20 | 天津通翔智能交通系统有限公司 | Multi-target traffic signal networked joint control method |
| CN108564786A (en) * | 2018-05-14 | 2018-09-21 | 苏州思在科技有限公司 | A kind of method and system of the urban traffic signal control based on cloud computing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109697867A (en) | 2019-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109697867B (en) | Deep learning-based traffic control method and system | |
| CN114418468B (en) | Smart city traffic scheduling strategy control method and Internet of things system | |
| CN111371904A (en) | Cloud-side-end-coordinated highway cloud control system and control method | |
| CN110895884B (en) | Traffic signal lamp control method and device | |
| CN105894831B (en) | Intelligent traffic control device | |
| CN106504583A (en) | The cloud monitoring method of parking position state, cloud server and cloud monitoring system | |
| CN210155839U (en) | Green wave passing system based on 5G Internet of vehicles | |
| US20190049264A1 (en) | Adaptive optimization of navigational routes using traffic data | |
| US20190051167A1 (en) | System and method of adaptive traffic management at an intersection | |
| CN111882894A (en) | Intelligent traffic signal lamp control method and system | |
| CN108922205B (en) | Traffic light switching time control system and method for plane intersection congestion situation | |
| CN111353375A (en) | System and method for determining priority of data processing | |
| CN105844925B (en) | Urban traffic control device, system and method based on mobile phone wireless net | |
| CN105844926B (en) | A kind of traffic lamp control method, apparatus and system | |
| CN112258865B (en) | Intelligent red and green signal lamp control system based on Internet of vehicles V2X | |
| CN114419904B (en) | Traffic light control system and control method based on vehicle Lu Yun cooperation | |
| CN104008659B (en) | A kind of system and method precisely monitoring intersection signal control device control effects | |
| CN113327415A (en) | Method, apparatus, device, storage medium and system for distributing traffic information | |
| CN110728850A (en) | Road traffic signal self-adaptive optimization system | |
| CN208781404U (en) | A kind of traffic light control system | |
| CN205670385U (en) | Urban traffic control device based on mobile phone wireless net | |
| CN113971883A (en) | Vehicle-road cooperative automatic driving method and efficient transportation system | |
| CN107230376A (en) | Parking management system and the information presentation system with its both-way communication | |
| CN113538943A (en) | Traffic light control method and system | |
| CN108109400A (en) | A kind of Intelligent traffic signal lamp control device that can be independently switched according to zebra linear flow rate |
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 | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Patentee after: Shenzhen Huier intelligent Co.,Ltd. Address before: 518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Patentee before: SHENZHEN OUDEKE TECHNOLOGY Co.,Ltd. |