JP2020009299A - Traffic signal control system, traffic management device, and traffic signal control method - Google Patents

Abstract

To make it possible to perform a signal control equivalent to a predetermined signal control method in real time even by a device having a not so high processing capability.SOLUTION: A control unit 12 of a traffic management device 3 causes a deep learning model to perform learning using learning information obtained by combining traffic condition information for a predetermined period and a signal control parameter generated by a predetermined signal control method based on traffic condition information, and constructs a deep learning model in which a signal control method is learned. Then the control unit executes the deep learning model that has learned the signal control method while using latest traffic condition information of a target area collected by a vehicle sensor 1 as input information, and the control unit acquires a signal control parameter as output information. In particular, the control unit constructs a deep learning model that has learned batch optimization control as a signal control method.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a traffic signal control system and a traffic signal control method for controlling a traffic light installed in a target area.

  The traffic management system controls traffic signals installed at intersections based on the traffic conditions of the target road network for the purpose of ensuring smooth traffic flow, ensuring traffic safety, and suppressing traffic pollution. Has been done. In this traffic signal control, a sub-area configuration is determined based on information collected by a vehicle detector installed on a road, and signal control parameters (cycle length, offset, split) are generated. .

  In such traffic signal control, pattern control and MODERAT (Management by Origin-Destination Related Adaptation for Traffic Optimization) control are adopted as signal control methods. , Split, and sub-area configuration are calculated separately, and the control index (such as demand rate) is indirect as an evaluation index for signal control. For this reason, it is desirable to perform control that can simultaneously optimize the cycle length, offset, split, and subarea configuration based on a direct and quantitative index, and actual operation has been started as batch optimization control (Non-Patent Document 1). reference).

  However, such batch optimization control has a problem that a processing load is large and a device having a high processing capability is required to perform real-time control, and the cost of the system is greatly increased. Therefore, it is conceivable to construct a machine learning model using learning information in which input information and output information of the collective optimization control are combined, and make the machine learning model simulate the collective optimization control.

  As a technology that applies such machine learning to traffic management, conventionally, a machine learning model (neural network) is used to convert future traffic volume and signal display design information as output information from traffic volume as input information. Techniques for acquiring the information are known (see Patent Documents 1 and 2).

JP-A-7-6292 Japanese Patent Application Laid-Open No. H8-1224085

Koichiro Iwaoka et al., "Application of batch optimization control using uplink collected from optical beacons", Transactions of Transportation Engineering, vol. 2, no. 2 (B), p. B_1-B_9, 2016.2

  However, in the above-described conventional technology, it is possible to predict a change in traffic demand or to make an apparatus perform a presenting design performed by an expert based on the experience and technology. It does not generate control parameters, and cannot solve the problem of the processing load of control for collectively optimizing signal control parameters.

  Accordingly, an object of the present invention is to provide a traffic signal control system and a traffic signal control method that can perform signal control equivalent to a predetermined signal control method in real time even with a device having a not so high processing capability. I do.

  A traffic signal control system according to the present invention is a traffic signal control system that controls a traffic light installed in a target area, and generates signal control information based on the latest traffic condition information of the target area collected by an information collecting device. A traffic signal control device, the traffic signal control device executing the deep learning model learning a predetermined signal control method using the latest traffic condition information as input information, and acquiring the signal control information as output information And a control unit that performs the control.

  Further, the traffic signal control method of the present invention is a traffic signal control method for controlling a traffic light installed in a target area, and in the traffic signal control device, the latest traffic condition information of the target area collected by the information collecting device is provided. The configuration is such that a deep learning model that has learned a predetermined signal control method is executed as input information, and signal control information is obtained as output information.

  According to the present invention, signal control simulating a predetermined signal control method can be performed by a deep learning model that has learned a predetermined signal control method. For this reason, even a device having not so high processing capability can perform signal control equivalent to a predetermined signal control method in real time. Thereby, the cost of the system can be reduced. In addition, signal control for a large-scale road network, which cannot be realized by a device having a high processing capability in the conventional control using a predetermined signal control method, can be performed.

Overall configuration diagram of the traffic signal control system according to the present embodiment The block diagram which shows the schematic structure of the traffic management apparatus 3. Explanatory drawing showing the outline of the processing performed by the learning information generation unit 21 Explanatory diagram showing random numbers that make traffic volume minutely change Explanatory drawing showing the outline of the processing performed by the learning information generation unit 21 The flowchart which shows the procedure of the process performed in the learning information generation part 21 The flowchart which shows the procedure of the process performed in the learning processing part 22 Explanatory diagram showing the outline of the deep learning model The flowchart which shows the procedure of the process performed by the deep learning model construction part 23 and the signal control information generation part 24.

  A first invention made to solve the problem is a traffic signal control system for controlling a traffic light installed in a target area, based on the latest traffic condition information of the target area collected by an information collecting device. A traffic signal control device that generates signal control information, the traffic signal control device executes the deep learning model that has learned a predetermined signal control method using the latest traffic condition information as input information, and outputs the information as output information. It is configured to include a control unit that acquires the signal control information.

  According to this, the signal control simulating the predetermined signal control method can be performed by the deep learning model that has learned the predetermined signal control method. For this reason, even a device having not so high processing capability can perform signal control equivalent to a predetermined signal control method in real time. Thereby, the cost of the system can be reduced. In addition, signal control for a large-scale road network, which cannot be realized by a device having a high processing capability in the conventional control using a predetermined signal control method, can be performed.

  In a second aspect, the control unit is configured to execute a deep learning model that has learned batch optimization control as the signal control method.

  According to this, the deep learning model simulates the batch optimization control, and the signal control equivalent to the batch optimization control can be performed without executing the batch optimization control algorithm having a large processing load.

  Also, a third invention is a traffic signal control system for controlling a traffic light installed in a target area, comprising an information processing device, the information processing device comprising: traffic condition information for a predetermined period; Using a learning information in which the signal control information generated by a predetermined signal control method based on the combination is used, a deep learning model is trained, and a control unit that constructs a deep learning model that has learned the signal control method is provided. A configuration is provided.

  According to this, a deep learning model that has learned a predetermined signal control method can be constructed.

  A fourth invention is configured such that the control unit constructs a deep learning model that has learned batch optimization control as the signal control method.

  According to this, the deep learning model simulates the batch optimization control, and the signal control equivalent to the batch optimization control can be performed without executing the batch optimization control algorithm having a large processing load.

  Further, in the fifth invention, the control unit executes a simulation by the signal control method using the traffic condition information of the predetermined period as input information, acquires the signal control information as output information, It is configured to generate information.

  According to this, it is possible to efficiently generate learning information by the signal control method.

  In a sixth aspect, the control unit generates virtual traffic condition information by changing the actually measured traffic condition information by using a random number, and generates the virtual traffic condition information for a predetermined period including the virtual traffic condition information. It is configured to acquire information.

  According to this, the trouble of actually measuring the necessary amount of traffic condition information can be omitted, and the necessary amount of learning information can be efficiently acquired.

  A seventh invention is a traffic signal control method for controlling a traffic light installed in a target area, wherein the traffic signal control device uses the latest traffic condition information of the target area collected by the information collecting device as input information. , A deep learning model that has learned a predetermined signal control method is executed to acquire signal control information as output information.

  According to this, similarly to the first invention, signal control equivalent to the predetermined signal control method can be performed in real time even by an apparatus having a not so high processing capability.

  An eighth invention is a traffic signal control method for controlling a traffic light installed in a target area, wherein the information processing device includes a traffic condition information for a predetermined period and a predetermined signal control based on the traffic condition information. Using the learning information in combination with the signal control information generated by the method, the deep learning model is made to learn, and a deep learning model that learns the signal control method is constructed.

  According to this, similarly to the second invention, a deep learning model that has learned a predetermined signal control method can be constructed.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is an overall configuration diagram of a traffic signal control system according to the present embodiment.

  This traffic signal control system (traffic control system) controls a traffic signal installed in a target area, and includes a vehicle detector 1 (information collection device), a signal controller 2, and a traffic management device (traffic signal control). Device, information processing device) 3.

  The vehicle sensor 1 is, for example, an ultrasonic vehicle sensor and an optical vehicle sensor (optical beacon), and detects vehicle on a road to obtain vehicle sensor information (traffic information) indicating a traffic condition of the vehicle on the road. Status information). This vehicle sensor information is transmitted from the vehicle sensor 1 to the traffic management device 3.

  The signal controller 2 controls the operation of the signal (signal lamp) based on the signal control parameter (signal control information) transmitted from the traffic management device 3.

  The traffic management device 3 is a central device of the traffic control center, generates a signal control parameter (signal control information) based on the vehicle sensor information transmitted from the vehicle sensor 1, and transmits the signal control parameter to a signal. Transmit to controller 2.

  In the present embodiment, the traffic control device 3 performs signal control using vehicle sensor information. However, the traffic control device 3 uses an ITS spot, an optical beacon, and a radio beacon (information collection device) installed on a road. Alternatively, signal control may be performed using probe information (traffic condition information) collected from an on-vehicle device mounted on a running vehicle.

  Next, a schematic configuration of the traffic management device 3 will be described. FIG. 2 is a block diagram illustrating a schematic configuration of the traffic management device 3.

  The traffic management device 3 includes a communication unit 11, a control unit 12, and a storage unit 13.

  The communication unit 11 communicates with the vehicle sensor 1 and the signal controller 2, receives the vehicle sensor information transmitted from the vehicle sensor 1, and transmits a signal control parameter (signal control information) to the signal controller 2. Send to

  The storage unit 13 stores vehicle sensor arrangement information, signal control setting information, predetermined signal control method setting information, vehicle sensor information, event regulation information, weather information, signal control parameters (signal control information), and the like. I do. Further, the storage unit 13 stores a program executed by a processor constituting the control unit 12. Here, the collective optimization control will be described as an example of the predetermined signal control method, and the storage unit 13 stores the setting information for the collective optimization control as the predetermined signal control method setting information.

  Here, the vehicle sensor placement information is information on the position of the vehicle sensor 1 installed on the road network in the target area, and specifically, the number of the link where the vehicle sensor 1 is installed and the end of the link From the vehicle detector 1 to the vehicle detector 1.

  The signal control setting information is setting information relating to signal control, such as the current configuration of each intersection in the target area. The collective optimization control setting information is setting information relating to the collective optimization control.

  The vehicle sensor information is information indicating the traffic situation collected by the vehicle sensor 1, and more specifically, the traffic volume, that is, the vehicle sensor information passes through the position of the vehicle sensor 1 every unit time (for example, every 5 minutes). The number of vehicles and the occupancy rate (time occupancy rate), that is, the ratio of the time during which the vehicle was present at the position of the vehicle sensor 1 within a unit time, and the like.

  The event regulation information is information related to lane regulation, closed roads, and the like in the road network in the target area. The weather information is information on the weather (rainfall or the like) in the target area. This weather information may be obtained from the weather information providing server, but the manager may input the weather of the day.

  The signal control parameter (signal control information) is information serving as an element that determines a display timing of a traffic signal at a traffic light, and specifically, is a cycle length, a split, and an offset.

  The vehicle sensor information, event regulation information, weather information, and signal control parameters include past information and the latest information, and the past information becomes learning information for building a deep learning model. Are related to the current signal control.

  The control unit 12 includes a learning information generation unit 21, a learning processing unit 22, a deep learning model construction unit 23, and a signal control information generation unit 24. The control unit 12 is configured by a processor, and each unit of the control unit 12 is realized by executing a program stored in the storage unit 13 by the processor.

  The learning information generation unit 21 generates learning information for causing the deep learning model to learn appropriate signal control parameters obtained by a predetermined signal control method. The learning information includes vehicle sensor information, event regulation information when the vehicle sensor information is acquired, and weather information as learning information input information, and is generated by a predetermined signal control method based on the input information. The obtained signal control parameters are associated as output information of learning information.

  The learning information generation unit 21 executes a predetermined signal control using a vehicle flow sensor simulator, event regulation information, and weather information of a predetermined number of days as input information using a traffic flow simulator based on a predetermined signal control method, and outputs output information. To obtain the signal control parameter.

  The learning processing unit 22 executes a learning process on the deep learning model using the learning information acquired by the learning information generating unit 21, and acquires a model parameter (neural network parameter) related to the deep learning model as a learning result.

  The deep learning model construction unit 23 applies the model parameters acquired by the learning processing unit 22 to the deep learning model, and constructs a deep learning model on which the learning result is reflected.

  The signal control information generation unit 24 executes the deep learning model constructed by the deep learning model construction unit 23, and generates signal control parameters (signal control information) from the vehicle sensor information. In this embodiment, in addition to the vehicle sensor information, event regulation information and weather information are input to the deep learning model.

  As described above, in the present embodiment, the signal control simulating the predetermined signal control is performed by the deep learning model that has learned the predetermined signal control. For this reason, even a device having a not so high processing capability can perform signal control equivalent to advanced signal control processing in real time, that is, in time for a control cycle (for example, 5 minutes).

  When there are a plurality of target areas, the traffic management device 3 constructs a deep learning model for each target area using the learning information for each target area.

  In the present embodiment, the traffic management device 3 is provided with the learning information generation unit 21, the learning processing unit 22, the deep learning model construction unit 23, and the signal control information generation unit 24. Only the signal control information generator 24 is operated. Further, each unit except the signal control information generation unit 24 is configured by another information processing device, and only the signal control information generation unit 24 is provided in the traffic management device 3, and the learned deep learning generated by another information processing device is performed. The model may be introduced into the traffic management device 3.

  Further, in the present embodiment, the signal control parameter serving as learning information (output information) will be described as being generated by the batch optimization control. May be generated. In this case, it is possible to construct a deep learning model that simulates pattern control and MODERATO control.

  Next, a process performed by the learning information generating unit 21 will be described. Here, a case where batch optimization control is adopted as a predetermined signal control method will be described as an example. FIG. 3 is an explanatory diagram illustrating an outline of a process performed by the learning information generation unit 21. FIG. 4 is an explanatory diagram showing random numbers that slightly change the traffic volume.

  The learning information generation unit 21 generates learning information for causing the deep learning model to learn a predetermined signal control method, specifically, the batch optimization control. This learning information is a combination of vehicle sensor information (input information) and signal control parameters (output information) generated by batch optimization control based on the vehicle sensor information.

  Here, since a deep learning model (multilayer neural network) has a plurality of intermediate layers, a large amount (for example, about 50 days) of learning information is required to construct a deep learning model with sufficient accuracy.

  Therefore, in the present embodiment, as shown in FIG. 3, virtual vehicle sensor information for a necessary number of days is generated based on vehicle sensor information (traffic volume) for one day actually measured in the target area. I do. Note that the vehicle sensor information is acquired for each unit period (for example, 5 minutes) that is a signal control parameter generation interval (control cycle).

  Specifically, the actually measured traffic amount for one day is minutely changed by a random number, and the traffic amount for a necessary number of days is generated. In the example shown in FIG. 4, a random number (normal random number) whose occurrence frequency indicates a normal distribution is generated in the range of -3 to +3, and the traffic volume at each measured time is in the range of -3 to +3. To obtain a virtual traffic volume at each time.

  In addition, the vehicle sensor information (traffic volume and occupancy) for a predetermined period is generated as learning information (input information). It may be included in the learning information. In this case, since the branching rate is collected by visual measurement, it takes a great deal of time to collect the branching rate for the required number of days. If the branching rate of the number of days is generated, the trouble of the actual measurement can be omitted.

  When the required number of days of vehicle sensor information is generated in this way, a traffic flow simulator is used to generate signal control parameters suitable for the traffic conditions represented by the vehicle sensor information of each day. In the present embodiment, the learning information generation unit 21 generates a signal control parameter corresponding to the vehicle sensor information of each day using a traffic flow simulator that performs batch optimization control. Then, learning information is generated by combining the vehicle sensor information that is the input information of the collective optimization control and the signal control parameter that is the output information of the collective optimization control. This makes it possible to efficiently acquire the necessary number of days of learning information.

  Next, an outline of a process performed by the learning information generation unit 21 will be described. FIG. 5 is an explanatory diagram illustrating an outline of a process performed by the learning information generation unit 21.

The learning information generation unit 21 includes a simulator for performing a traffic flow model process and a batch optimization control process. In this simulator, traffic condition information C _{1 to} C _N (vehicle sensor information) for a predetermined number of days, an event A simulation based on a predetermined signal control method, specifically, a collective optimization control is executed using the regulation information and the weather information as input information, and signal control parameters P _{1 to PN} are obtained as output information.

  As described in Non-Patent Document 1, the collective optimization control includes a sub-area configuration of an intersection where a traffic signal is installed in a road network in a target area, and a signal control parameter (cycle length, offset, Split) is optimized collectively, a PI (performance index) for evaluating the traffic condition of the entire road network is estimated, and control parameters for minimizing the PI of the entire road network are calculated.

  Specifically, first, a numerical sequence model in which the subarea configuration and the control parameters are represented by a numerical sequence is introduced. In this numerical sequence model, the sub-area configuration (sub-area coupling) is defined by a link between sub-areas (inter-area link).

  Then, based on the vehicle sensor information (traffic volume), the control parameters, and the initial values of the sub-area configuration, the traffic flow model calculates the delay time and the number of stops of the entire road network. Is estimated as a linear sum of Next, based on the acquired PI, a search for a numerical sequence model is performed by meta-heuristics, and a new control parameter and a candidate for a sub-area configuration are obtained. Further, PI is estimated by a traffic flow model based on the control parameters and the candidates for the sub-area configuration. By repeating the PI estimation and the numerical sequence model search, the optimal control parameters and the sub-area configuration are determined.

  Next, a procedure of a process performed by the learning information generating unit 21 will be described. FIG. 6 is a flowchart illustrating the procedure of the process performed by the learning information generation unit 21.

  First, the learning information generating unit 21 acquires from the storage unit 13 vehicle sensor arrangement information, signal control setting information, batch optimization control setting information, event regulation information, and weather information (ST101). Then, a preparation process for the batch optimization control is performed based on the information (ST102).

  Specifically, the form of the target road network is determined based on the vehicle sensor arrangement information. Further, based on the signal control setting information and the batch optimization control setting information, processing necessary for optimizing the signal control parameters, such as associating a variable related to the signal control parameters with a setting related to meta-heuristics, is performed.

  Next, the learning information generation unit 21 acquires vehicle sensor information for a predetermined number of days (ST103). At this time, virtual vehicle sensor information for the required number of days is generated from the reference vehicle sensor information for one day.

  Next, the learning information generating unit 21 performs collective optimization control using the vehicle sensor information, the event regulation information at the time of acquiring the vehicle sensor information, and the weather information as input information, and performs output optimization information as output information. A signal control parameter is obtained (ST104).

  Next, the learning information generating unit 21 combines the vehicle sensor information, the event regulation information, and the weather information, which are the input information of the collective optimization control, with the signal control parameter, which is the output information of the collective optimization control, to perform learning. Information is generated (ST105). At this time, the input information and the output information are combined for each unit period (for example, 5 minutes) which is a signal control parameter generation interval (control cycle).

  Next, processing performed by the learning processing unit 22 will be described. FIG. 7 is a flowchart illustrating the procedure of the process performed by the learning processing unit 22.

  The learning processing unit 22 uses the learning information acquired by the learning information generation unit 21, that is, the vehicle sensor information (input information) for a predetermined period and the signal control parameters (output information) corresponding to the learning information to perform deep learning. A learning process is performed on the model, and a model parameter (neural network parameter) is obtained as a learning result. For this learning processing, a known learning algorithm, for example, an error back propagation method or the like may be used.

  Specifically, first, the learning information generated by the learning information generating unit 21 is obtained from the storage unit 13 (ST201). Next, a learning process is performed using the learning information according to a learning algorithm such as an error back propagation method (ST202). Then, the model parameters acquired in the learning process are stored in the storage unit 13 (ST203).

  Next, processing performed by the deep learning model construction unit 23 and the signal control information generation unit 24 will be described. FIG. 8 is an explanatory diagram illustrating an outline of the deep learning model.

  The deep learning model construction unit 23 sets the model parameters acquired by the learning processing unit 22 in the deep learning model, and constructs a deep learning model in which the learning result is reflected. The signal control information generation unit 24 generates a signal control parameter (signal control information) from the vehicle sensor information (traffic condition information) using a deep learning model. In this embodiment, in addition to the vehicle sensor information, event regulation information and weather information are input to the deep learning model.

The deep learning model is configured by a multilayer neural network, and includes an input layer, an intermediate layer, and an output layer. The input layer, traffic information (traffic volume of each link) C ₁ ... C _N, event regulation information and weather information are inputted, the output layer, the signal control parameter P ₁ ... P _N is output .

  In the present embodiment, since the deep learning model is constructed using the learning information generated by the collective optimization control, the learned deep learning model simulates the collective optimization control.

  Next, a procedure of processing performed by the deep learning model construction unit 23 and the signal control information generation unit 24 will be described. FIG. 9 is a flowchart illustrating a procedure of processing performed by the deep learning model construction unit 23 and the signal control information generation unit 24.

  The deep learning model construction unit 23 first acquires the model parameters generated by the learning processing unit 22, as shown in FIG. 9A (ST301). Then, the model parameters are set to the unlearned deep learning model, and a deep learning model reflecting the learning result is constructed (ST302).

  As shown in FIG. 9B, the signal control information generation unit 24 acquires vehicle sensor information, event regulation information, and weather information from the storage unit 13 (ST401). Next, vehicle sensor information, event regulation information, and weather information are input to the learned deep learning model constructed by the deep learning model construction unit 23 (ST402). Next, the signal control parameters output from the deep learning model are stored in the storage unit 13 and transmitted to the signal controller 2 (ST403).

  As described above, the embodiments have been described as examples of the technology disclosed in the present application. However, the technology in the present disclosure is not limited to this, and can be applied to embodiments in which changes, replacements, additions, omissions, and the like are made. Further, it is also possible to form a new embodiment by combining the components described in the above embodiment.

  The traffic signal control system and the traffic signal control method according to the present invention have an effect that signal control equivalent to a predetermined signal control method can be performed in real time even by a device having a not so high processing capability, and is installed in a target area. The present invention is useful as a traffic signal control system and a traffic signal control method for controlling a traffic signal.

1 vehicle detector (information collection device)
2 signal controller 3 traffic management device (traffic signal control device, information processing device)
Reference Signs List 11 Communication unit 12 Control unit 13 Storage unit 21 Learning information generation unit 22 Learning processing unit 23 Deep learning model construction unit 24 Signal control information generation unit

A first invention for solving the above-mentioned problem is a traffic signal control system that controls a traffic light installed in a target area, and detects a vehicle on a road and indicates a traffic situation of the vehicle on the road. a vehicle detector for generating a vehicle detector information, and traffic management device that generates a signal control information based on the latest traffic conditions information of the target area generated by the vehicle detectors, signal control this traffic management system has generated and a signal controller for controlling the operation of the traffic based on the information, the traffic control device, the as input information the latest traffic conditions information, running learning models trained a predetermined signal control method And a control unit that acquires the signal control information as output information.

According to this, the learning models trained a predetermined signal control method, it is possible to perform a simulated signal controlling a predetermined signal control method. For this reason, even a device having not so high processing capability can perform signal control equivalent to a predetermined signal control method in real time. Thereby, the cost of the system can be reduced. Further, signal control for a large-scale road network, which cannot be realized by a device having a high processing capability by the conventional control using a predetermined signal control method, can be performed.

In a second aspect, a vehicle detector in the traffic signal control system is configured.

In a third aspect, a signal controller in the traffic signal control system is configured.

According to a fourth aspect of the present invention, an input unit for inputting the latest traffic condition information on a road, and a learning model that learns a predetermined signal control method with respect to the latest traffic condition information input from the input unit is executed. The traffic management device includes a control unit and an output unit that outputs the signal control information generated by the control unit to a signal controller .

According to a fifth aspect of the present invention , learning is performed on a learning model using learning information obtained by combining traffic condition information for a predetermined period and signal control information generated by a predetermined signal control method based on the traffic condition information. In addition, a traffic management device including a control unit that constructs a learning model that has learned the signal control method is configured.

According to a sixth aspect of the present invention, there is provided a traffic signal control method for controlling a traffic light installed in a target area. In the traffic signal management device, latest traffic condition information of the target area collected by a vehicle detector is used as input information. A traffic signal control method characterized by executing a learning model that has learned a predetermined signal control method and acquiring signal control information for controlling the operation of the traffic signal as output information .

Further, a seventh invention is a traffic signal control method for controlling a traffic light installed in a target area, wherein the traffic signal management device includes traffic condition information for a predetermined period and a predetermined signal based on the traffic condition information. A traffic signal control method is characterized in that a learning model is trained using learning information obtained by combining signal control information generated by a control method and a learning model learning the signal control method is constructed. And

A first invention for solving the above-mentioned problem is a traffic signal control system for controlling a traffic light installed in a target area, which detects a vehicle on a road and indicates a traffic condition on the road. Sensor that generates sensor information, a traffic management device that generates signal control information based on the latest vehicle sensor information of the target area generated by the vehicle sensor, and signal control information that is generated by the traffic management device A traffic controller that controls the operation of the traffic signal based on the latest vehicle sensor information obtained from the vehicle sensor as input information, at predetermined intervals from the vehicle sensor. run the vehicle detectors information acquired, a learning model that learns the predetermined signal control method by coupling the signal control information generated at predetermined intervals, and output information A configuration including a control unit for acquiring the signal control information.

A second invention is a traffic signal control system that controls a traffic light installed in a target area, and detects a vehicle on a road and generates vehicle sensor information indicating a traffic condition on the road. Device, a traffic management device that generates signal control information based on the latest vehicle sensor information of the target area generated by the vehicle sensor, and an operation of the traffic signal based on the signal control information generated by the traffic management device. A signal controller for controlling the traffic control device, the traffic management device uses the latest vehicle sensor information as input information, and fluctuates vehicle sensor information actually measured by the vehicle sensor using a random number. A learning model in which a predetermined signal control method is learned using the vehicle sensor information for a predetermined period including the vehicle sensor information and the corresponding signal control information is executed, and signal control is performed as output information. A configuration including a control unit for acquiring information.

Further, a third invention provides an input unit for inputting the latest vehicle sensor information on a road, and using the latest vehicle sensor information input from the input unit as input information at predetermined intervals from the vehicle sensor. A control unit that executes a learning model that has learned a predetermined signal control method by combining the acquired vehicle sensor information and the signal control information generated at the predetermined intervals, and generates signal control information ; And a signal output device for outputting the signal control information generated by the control unit to the signal controller.

In a fourth aspect , the vehicle sensor information acquired at each of the predetermined intervals from the vehicle sensor includes a virtual sensor generated by changing the actually measured vehicle sensor information at each of the predetermined intervals by using a random number. The traffic management device is characterized in that the traffic management device includes important vehicle sensor information .

A fifth aspect of the present invention is a traffic signal control method for controlling a traffic signal installed in the target area, the traffic signal control apparatus, as input information the latest vehicle detectors information Target area, the vehicle sensing The vehicle control system executes a learning model in which a predetermined signal control method is learned by combining the vehicle sensor information acquired at predetermined intervals from the sensor and the signal control information generated at each predetermined interval , and outputs the signal control information. It generates, as a constituent of the traffic signal control method characterized by obtaining the signal control information for controlling the operation of the traffic signal as output information.

According to a sixth aspect of the present invention, there is provided a traffic signal control method for controlling a traffic light installed in a target area, wherein the traffic signal management device inputs latest vehicle sensor information acquired from a vehicle sensor in the target area. As the information, vehicle sensor information for a predetermined period including virtual vehicle sensor information generated by varying the vehicle sensor information actually measured by the vehicle sensor with a random number, and signal control information corresponding thereto. A traffic signal control method comprising: executing a learning model that has learned a predetermined signal control method using the control unit; generating signal control information; and obtaining signal control information for controlling operation of the traffic signal as output information. And

Claims (8)

A traffic signal control system that controls a traffic light installed in a target area,
A traffic signal control device that generates signal control information based on the latest traffic condition information of the target area collected by the information collection device,
This traffic signal control device
A traffic signal control system, comprising: a control unit that executes a deep learning model that has learned a predetermined signal control method using the latest traffic condition information as input information and acquires the signal control information as output information. . The control unit includes:
The traffic signal control system according to claim 1, wherein a deep learning model that has learned batch optimization control is executed as the signal control method. A traffic signal control system that controls a traffic light installed in a target area,
Equipped with an information processing device,
This information processing device
Using a learning information in which traffic condition information of a predetermined period is combined with signal control information generated by a predetermined signal control method based on the traffic condition information, a deep learning model is made to perform learning, and the signal control method is performed. A traffic signal control system, comprising: a control unit that constructs a deep learning model that has learned the learning. The control unit includes:
The traffic signal control system according to claim 3, wherein a deep learning model that learns batch optimization control as the signal control method is constructed. The control unit includes:
The simulation according to the signal control method is executed by using the traffic condition information of the predetermined period as input information, the signal control information is obtained as output information, and the learning information is generated. The traffic signal control system as described. The control unit includes:
The method according to claim 3, wherein the measured traffic condition information is varied by random numbers to generate virtual traffic condition information, and the learning information for a predetermined period including the virtual traffic condition information is acquired. The traffic signal control system as described. A traffic signal control method for controlling a traffic light installed in a target area,
In the traffic signal control device,
The latest traffic condition information of the target area collected by the information collection device is used as input information, a deep learning model that has learned a predetermined signal control method is executed, and signal control information is obtained as output information. Signal control method. A traffic signal control method for controlling a traffic light installed in a target area,
In the information processing device,
Using a learning information in which traffic condition information of a predetermined period is combined with signal control information generated by a predetermined signal control method based on the traffic condition information, a deep learning model is made to perform learning, and the signal control method is performed. A traffic signal control method characterized by constructing a deep learning model that has learned the language.

Images