KR102643500B1 - Data collection apparatus for fire receiver based on communication signal photographing data and remote fire protection system comprising the same - Google Patents

Abstract

A data collection device for a firefighting receiver according to the present invention includes a signal collection unit that collects communication signals inside the firefighting receiver; a data generation unit that generates status data of the firefighting receiver using the communication signal collected by the signal collection unit; and a data output unit that outputs the status data generated in the data generation unit to the outside. In this way, since the present invention is configured to collect communication signals inside the fire-fighting receiver and generate and output status data of the fire-fighting receiver, as long as it corresponds to a fire-fighting receiver that monitors whether a fire has occurred through a communication signal, the fire-fighting receiver It can be used for general purposes regardless of type.

Description

Data acquisition device for firefighting receiver based on communication signal and remote firefighting control system including the same

The present invention relates to a device installed in a firefighting receiver to collect and control data from the firefighting receiver, and a remote firefighting control system including the device.

As is well known, a firefighting receiver is installed in the disaster prevention room of a building, which monitors whether a fire has occurred through a fire signal detected by a fire detector and operates an alarm generator when a fire occurs. In order for these firefighting receivers to operate normally, a person must reside in the disaster prevention room where the firefighting receivers are installed and continuously monitor the firefighting receivers. However, in reality, it is very difficult for a person to continuously monitor a firefighting receiver 24 hours a day.

If firefighting receivers are not continuously monitored, damage to life and property increases in the event of a dangerous situation such as a fire. And if the fire receiver malfunctions, it will result in a very large loss of manpower, such as unnecessary mobilization of fire trucks or ambulances and unnecessary evacuation of people in the building or nearby residents. Accordingly, there is a need to establish a remote fire control system that immediately informs the manager of the fire receiver of the status of the fire receiver and allows the manager of the fire receiver to remotely control the fire receiver.

There are various manufacturers of firefighting receivers, and in general, even if it is a single manufacturer, the models of firefighting receivers produced by that manufacturer are diverse. Accordingly, in order to build a remote fire control system, it is necessary to provide a data collection device for firefighting receivers that can be universally used in various types of firefighting receivers. Furthermore, there is a need to prepare a method for the manager of the fire receiver to immediately respond when the fire receiver detects a fire or malfunctions.

Meanwhile, the following patent document discloses a fire-fighting disaster prevention monitoring system and method that can check the connection status of each sensor and fire-fighting equipment and immediately respond to fire by immediately checking when a fire occurs.

1. Korean Patent Publication No. 10-2019-0091757 (published on August 7, 2019) 2. U.S. Patent Publication No. 2012-0072053 (published March 22, 2012)

The purpose of the present invention is to provide a data collection device for a firefighting receiver that can be universally used in various types of firefighting receivers and a remote firefighting control system including the same.

In addition, the present invention provides a data collection device for a firefighting receiver that allows the manager of the firefighting receiver to immediately respond to such abnormal situation when an abnormal situation such as a fire or malfunction occurs in the firefighting receiver, and a remote firefighting control system including the same. There is a purpose to that.

However, the technical problem to be solved by the present invention is not limited to the above-mentioned problems, and other problems not mentioned can be clearly understood by those skilled in the art from the description of the invention described below.

In order to achieve the above object, a data collection device for a firefighting receiver according to the present invention includes a signal collection unit that collects communication signals inside the firefighting receiver; a data generation unit that generates status data of the firefighting receiver using the communication signal collected by the signal collection unit; and a data output unit that outputs the status data generated in the data generation unit to the outside.

Here, the data generator performs machine learning to predict the state of the fire-fighting receiver using the communication signal collected by the signal collection unit, and generates state data of the fire-fighting receiver using the results of the machine learning. can do.

The data collection device for a firefighting receiver according to the present invention may further include a signal pre-processing unit that performs a labeling operation on the communication signal collected by the signal collection unit, and the data generating unit performs the labeling operation by the signal pre-processing unit. Machine learning can be performed to predict the state of the firefighting receiver using communication signals.

Here, the signal preprocessor may perform different labeling tasks on a communication signal when the firefighting receiver detects a fire and a communication signal when the firefighting receiver does not detect a fire.

Here, the signal collection unit may further collect at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and data on presence or absence of oil stains on the fire-fighting receiver, The data generator, together with the communication signal collected by the signal collection unit, illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and data on whether or not grease is attached to the fire-fighting receiver. Machine learning can be performed to predict the state of the firefighting receiver using at least one of the following.

Here, the data generator may perform machine learning to predict the state of the firefighting receiver using a GAN learning model.

The data collection device for a firefighting receiver according to the present invention initializes the operation of the firefighting receiver according to the recovery signal when receiving a recovery signal transmitted from the outside in response to the status data output to the outside by the data output unit. It may further include a recovery signal collection unit.

Meanwhile, in order to achieve the above object, a remote fire control system according to the present invention includes a firefighting receiver that monitors whether a fire has occurred; A data collection device for a firefighting receiver that collects communication signals inside the firefighting receiver and generates status data of the firefighting receiver; a server that stores status data generated by the data collection device for the firefighting receiver; And it may include an administrator terminal that receives status data from the server and displays the status data on a screen so that the manager of the fire protection receiver can monitor the fire protection receiver.

Here, the data collection device for the firefighting receiver includes a signal collection unit that collects communication signals inside the firefighting receiver; A data generation unit that performs machine learning to predict the state of the fire-fighting receiver using the communication signal collected by the signal collection unit, and generates state data of the fire-fighting receiver using the results of the machine learning; and a data output unit that outputs the status data generated in the data generation unit to the outside.

Here, the data collection device for the firefighting receiver may further include a signal pre-processing unit that performs a labeling operation on the communication signal collected by the signal collection unit, and the data generating unit may perform a communication signal on which the labeling operation is performed by the signal pre-processing unit. Machine learning can be performed to predict the state of the firefighting receiver using the signal.

Here, the signal preprocessor may perform different labeling tasks on a communication signal when the firefighting receiver detects a fire and a communication signal when the firefighting receiver does not detect a fire.

Here, the signal collection unit may further collect at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and data on presence or absence of oil stains on the fire-fighting receiver, The data generator, together with the communication signal collected by the signal collection unit, illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and data on whether or not grease is attached to the fire-fighting receiver. Machine learning can be performed to predict the state of the firefighting receiver using at least one of the following.

Here, the data generator may perform machine learning to predict the state of the firefighting receiver using a GAN learning model.

Here, the data collection device for the firefighting receiver initializes the operation of the firefighting receiver according to the recovery signal when receiving a recovery signal transmitted from the outside in response to the status data output to the outside by the data output unit. It may further include a recovery signal collection unit.

The remote fire control system according to the present invention may further include a CCTV that provides an image of a target monitored by the fire receiver to the manager terminal when the manager terminal is connected to communication.

Since the present invention is configured to collect communication signals inside the firefighting receiver to generate and output status data of the firefighting receiver, the type of firefighting receiver can be classified as long as it corresponds to a firefighting receiver that monitors whether a fire has occurred through a communication signal. Regardless, it can be used for general purposes. In addition, since the present invention has no effect on the operation of the fire-fighting receiver, it does not require design changes to the fire-fighting receiver, and therefore has great practicality and wide application.

In addition, since the present invention is configured so that the status data generated by the data collection device for the firefighting receiver is not only stored in the server, but is also transmitted to the administrator terminal through the server, it is necessary for the manager of the firefighting receiver to continuously monitor the firefighting receiver. There is no In addition, according to the present invention, the manager of the fire receiver can connect to the CCTV through the manager terminal and check the image of the target monitored by the fire receiver in real time, so that the fire receiver can detect a fire or When the receiver malfunctions, the manager of the fire receiver can immediately respond to the fire or malfunction.

However, the effects that can be obtained from the present invention are not limited to the effects mentioned above, and other effects not mentioned can be clearly understood by those skilled in the art from other descriptions in the specification. It could be.

Figure 1 is a diagram schematically showing a remote fire control system according to an embodiment of the present invention.
FIG. 2 is a block diagram of a data collection device for a firefighting receiver shown in FIG. 1.
Figure 3 is an example of status data indicating that the firefighting receiver has detected a fire displayed on the screen of the administrator terminal.
Figure 4 is an example diagram of a CCTV linkage screen transmitted through an administrator terminal.
Figure 5 is an example of a recovery button that can initialize the operation of the firefighting receiver on the screen of the administrator terminal.

Hereinafter, a data collection device for a firefighting receiver and a remote firefighting control system including the same according to the present invention will be described in detail with reference to the attached drawings. The attached drawings are provided as examples in order to sufficiently convey the technical idea of the present invention to those skilled in the art, and the present invention is not limited to the drawings presented below and can be embodied in many other forms. there is.

Figure 1 is a diagram schematically showing a remote fire control system 1000 according to an embodiment of the present invention. As shown in FIG. 1, the remote fire control system 1000 according to an embodiment of the present invention includes a firefighting receiver 100, a data collection device for the firefighting receiver 200, a server 300, and an administrator terminal 400. may include.

The firefighting receiver 100 may be installed in a disaster prevention room of a building such as an apartment, supermarket, factory, or officetel, and monitors whether a fire has occurred through a fire signal detected by a fire detector, and sends an alarm if a fire occurs in the building. It plays a role such as operating the generator. Here, the fire signal detected by the fire detector is generally a communication signal, and the interior of the firefighting receiver 100 is provided with a plurality of circuits that receive communication signals such as the fire signal. In the present invention, a data collection device 200 for a fire-fighting receiver that can be used universally regardless of the type of fire-fighting receiver as long as it corresponds to a fire-fighting receiver 100 that monitors whether a fire has occurred through a communication signal, and includes the same. We propose a remote fire control system (1000) that

One or more data collection devices 200 for firefighting receivers may be installed outside or inside the firefighting receiver 100. As described below, the data collection device 200 for a firefighting receiver collects communication signals inside the firefighting receiver 100, generates status data of the firefighting receiver 100, and transmits the generated status data to an external (e.g. For example, output to the server).

FIG. 2 is a block diagram of the data collection device 200 for a firefighting receiver shown in FIG. 1. The data collection device 200 for a firefighting receiver basically includes a signal collection unit 210, a data generation unit 230, and a data output. Includes part 240.

The signal collection unit 210 is connected to a number of circuits provided inside the firefighting receiver 100, and algorithms such as sniffing or Man In The Middle are stored therein. The signal collection unit 210 collects communication signals transmitted to multiple circuits by executing algorithms such as sniffing or Man In The Middle stored therein. Here, the communication signal may be a fire signal or a fire non-occurrence signal (i.e., a normal signal). Additionally, the signal collection unit 210 may collect location information of an object monitored by the firefighting receiver 100 (for example, a room in a building, etc.) from the firefighting receiver 100 . Here, the location information of the object monitored by the firefighting receiver 100 may be previously stored in the firefighting receiver 100.

The signal collection unit 210 as described above can capture and receive signals transmitted and received by the firefighting receiver 100, by remotely capturing electromagnetic waves generated from signals transmitted and received by the firefighting receiver 100 for a preset time. Electromagnetic signal flows can be generated in chronological order, and record sets selected within the signal flow can be clustered according to preset criteria.

Here, the preset standard may be the signal strength level of electromagnetic waves measured when electromagnetic waves are captured remotely.

The signal collection unit 210 as described above can perform monitoring of the firefighting receiver 100 through multiple signal flow collection in order to safely collect signals from the firefighting receiver 100, and monitors the firefighting receiver 100 through multiple signal flows for a preset time. Data can not only be generated, but also output an alarm if the signal strength level value of the signal flow collected when clustering the recordset selected from the multiple signal flow does not meet (less than) the preset standard (i.e., the reference signal strength level value). An alarm signal can be transmitted to the manager terminal 400 through the output unit 240.

Meanwhile, the signal collection unit 210 may include a plurality of buffers for storing recordsets selected within the clustered signal flow. When storing the recordsets selected within the clustered signal flow, each buffer is stored according to the signal strength level of the electromagnetic wave. Data switching can be performed so that recordsets of different clusters can be stored in different buffers.

In addition, when selecting a record set within a clustered signal flow, the signal collection unit 210 individually allocates packet identifiers and individually stores each assigned packet identifier in a plurality of buffers to prevent interference with other record sets. Selected recordsets can be separated and safely stored within the clustered signal flow.

The data generation unit 230 generates status data of the firefighting receiver 100 using the communication signal collected by the signal collection unit 210.

For example, a communication signal analysis module may be stored in the data generator 230. In this case, the data generator 230 can analyze the communication signal collected by the signal collection unit 210 through the communication signal analysis module to determine whether the communication signal is a fire signal, and the determination result Accordingly, status data of the firefighting receiver 100 can be generated. Here, the status data generated by the data generator 330 may include data regarding whether the firefighting receiver 100 has detected a fire.

Alternatively, a machine learning model may be stored in the data generator 230. In this case, the data generator 230 performs machine learning to predict the state of the firefighting receiver 100 using the machine learning model and the communication signal collected by the signal collection unit 210. In addition, when another communication signal is collected by the signal collection unit 210 after the machine learning is performed, the data generator 230 uses the results of the machine learning to generate status data of the firefighting receiver 100. can be created. Here too, the status data generated by the data generator 330 may include data regarding whether the firefighting receiver 100 has detected a fire.

In order to improve the accuracy of the state data generated by the data generator 230, the data collection device 200 for a firefighting receiver may further include a signal preprocessor 220. The signal pre-processing unit 220 performs a labeling operation on the communication signal collected by the signal collection unit 210. In this case, the data generating unit 230 uses the communication signal for which the labeling operation has been performed by the signal pre-processing unit 220. Then, machine learning is performed to predict the state of the firefighting receiver 100.

More specifically, the signal pre-processing unit 220 performs different labeling operations on a communication signal when the firefighting receiver 100 detects a fire and a communication signal when the firefighting receiver 100 does not detect a fire. can do. For example, the state of the firefighting receiver 100 can be classified and labeled by binary classification according to the presence or absence of fire detection. In this case, the signal preprocessor 220 is configured to operate when the firefighting receiver 100 detects a fire. The communication signal of can be labeled as 0, and the communication signal when the firefighting receiver 100 does not detect a fire can be labeled as 1. However, the signal pre-processing unit 220 can also perform labeling for more diverse states of the firefighting receiver 100, and may also perform labeling using methods other than 0 and 1.

In this way, when the signal pre-processing unit 220 performs different labeling tasks depending on whether the fire protection receiver 100 detects a fire, the data generation unit 230 performs machine learning to predict the state of the fire protection receiver 100, In particular, machine learning on whether the firefighting receiver 100 has detected a fire can be performed with much higher accuracy.

In addition, the signal preprocessor 220 may extract record set data from different clusters stored in different buffers provided in the signal collection unit 210 and perform a labeling task for clustering processing.

Meanwhile, the signal collection unit 210 includes illuminance data around the firefighting receiver 100, smoke amount data around the firefighting receiver 100, dust amount data around the firefighting receiver 100, and data on whether or not grease is attached to the firefighting receiver 100. You can collect at least one more of them. Here, illuminance data, smoke amount data, and dust amount data may be collected from an illuminance sensor, a smoke sensor, and a dust sensor installed in or near the firefighting receiver 100, respectively. In addition, data on whether grease is attached to the firefighting receiver 100 can be collected through data input by the manager of the firefighting receiver 100, etc.

In this case, the data generator 230 generates 'illuminance data around the firefighting receiver 100, smoke amount data around the firefighting receiver 100, and firefighting receiver 100 along with the communication signal collected by the signal collection unit 210. Predicting the state of the firefighting receiver 100 using at least one of the surrounding dust amount data and the oil stain adhesion data on the firefighting receiver 100 (hereinafter referred to as 'illuminance data around the firefighting receiver 100, etc.') Machine learning can be performed for The reason why the data generator 230 uses the illuminance data around the firefighting receiver 100 as learning data for machine learning is to take into account the conditions in which the firefighting receiver 100 is placed in various environments, and the firefighting receiver 100 This is to generate state data as accurately as possible.

The data generator 230 may have a GAN (Generative Adversarial Network) learning model stored in advance. Communication signals collected by the signal collection unit 210 (or communication signals labeled by the signal pre-processing unit 220) and illuminance data around the firefighting receiver 100 are sent to the data generating unit 230 as learning data. When input, the data generator 230 may perform machine learning to predict the state of the firefighting receiver 100 by applying the learning data to the input layer of the GAN learning model.

Here, the GAN learning model generates a learning model with characteristics very similar to the actual firefighting receiver 100 by competing with a generator that learns the probability distribution and a discriminator that distinguishes different sets. More specifically, the communication signal collected by the signal collection unit 210 (or the communication signal labeled by the signal preprocessor 220), and the illuminance data around the firefighting receiver 100 are input layers of the GAN learning model. When applied to the GAN learning model, the generator of the GAN learning model creates fake examples of the characteristics of the actual firefighting receiver (100) and trains them to deceive the discriminator as much as possible, and the discriminator uses the fake examples presented by the generator to match the characteristics of the actual firefighting receiver (100). Train to be able to distinguish them as accurately as possible. Through this training process of the generator and discriminator, the GAN learning model generates a learning model with characteristics very similar to the actual firefighting receiver 100.

When the data generator 230 performs machine learning using a GAN learning model, machine learning is performed with variously transformed learning data, so the status data of the firefighting receiver 100 generated subsequently has high accuracy. You can have it.

After the data generator 230 performs machine learning to predict the state of the firefighting receiver 100, the results of the machine learning (i.e., learning model) can be stored. Afterwards, when a communication signal (or a communication signal labeled by the signal pre-processing unit 220) and illuminance data around the firefighting receiver 100 are input to the data generating unit 230, the data generating unit 230 Status data of the firefighting receiver 100 is generated using the stored results of machine learning. At this time, the data generator 230 generates data in different formats (i.e., data indicating that the firefighting receiver 100 has detected a fire, and data indicating that the firefighting receiver 100 has detected a fire) depending on whether the firefighting receiver 100 has detected a fire. The data indicating that the fire has not been detected may have a different format.

The data generator 230 as described above may include a learning module including a data input layer, a plurality of weight addition layers, and an inferred value output layer. This learning module is provided through a signal collection unit in the data input layer. Collected signal data is input, weights are added to each signal data through a weight addition layer, and weights are added through an inferred value output layer to output the calculated value.

Here, the learning module uses the data value input to the data input layer as the input value, the output value as the result value, and learns by using the operating state of the firefighting receiver 100 as the target value while collecting signal data. By performing this, an operating state data generation algorithm for the firefighting receiver 100 can be created.

In addition, the data generator 240 may generate real-time operating state data of the firefighting receiver 100 by inputting the real-time signals collected by the signal collection unit 210 into an operating state data generation algorithm.

In addition, the data generator 230 calculates a loss score by comparing the output value output to the inference value output layer with the target value. If the loss score is less than a preset standard score, the learning of the learning module cannot be completed. And, if the loss score exceeds the preset standard score, the weight value of the weighting layer can be repeatedly modified until the loss score becomes less than the preset standard score.

Here, in modifying the weight value of the weight addition layer, the data generator 230 adjusts the weight value of the weight addition layer according to a preset weight correction function in the direction from the inferred value output layer to the data input layer based on the loss score. After correction, the loss score can be recalculated by performing a weighting operation from the data input layer to the inferred value output layer.

This data generator 240 can learn to improve inference accuracy by inputting the algorithm generated through the learning module back into the general-purpose inference engine. The algorithm generated through learning of the learning module can be converted into an independent high-level language. It can be converted into an executable code, and the executable code is separated into general purpose language (GPL: general purpose language) code and domain specific language (DSL: dmmain specific language) code depending on whether accelerated operations are needed, and the separated general purpose language code. And domain-specific language codes can be generated as target codes optimized for firefighting receivers.

When generating a target code, this general-purpose inference engine is executed in the central processing unit (CPU) or computational accelerator of the fire-fighting receiver 100 based on the result of analyzing the general-purpose language code or the configuration state of the computational accelerator of the fire-fighting receiver 100. It can be created with target code.

The data output unit 240 outputs the status data generated in the data generation unit 230 to the outside. As an example, the data output unit 240 may output the status data to the server 300 located outside the firefighting receiver data collection device 200. For this purpose, it is preferable that the firefighting receiver data collection device 200 and the server 300 are connected to each other.

The data output unit 240 as described above can transmit and receive data by performing quantum-resistant encryption using public key cryptography, and transmits the data generated in the data generation unit 230 to the manager terminal 400. , When receiving a control signal transmitted from the administrator terminal 400, McEliece, Modern McElice, Niederreiter, MCPC-McElice, Wild McEliece, McBits, NTRU, SS- Data packets applying at least one algorithm selected from NTRU, BLISS, NewHope, W-OTS+, SPHINCS, and HORS can be transmitted or received.

Such quantum-resistant encryption can be classified into multivariate-based, code-based, grid-based, isogeny-based, hash-based, etc. Multivariate-based encryption is encryption based on the difficulty of the multivariate function problem calculated on a finite field, To improve safety and computational efficiency, quadratic functions can be mainly used, and since encryption and decryption are polynomial calculations, it has the advantage of being safe as it is resistant to side-channel attacks of power analysis. For example, HFE, ZHFE, UOV, etc. It can be included.

Code-based encryption is an encryption based on the difficulty of decoding general linear codes. It can improve operation speed through matrix operations, and intentionally injects errors so that only users who are aware of the errors can restore the message. This may include, for example, McEllis, Modern McEllis, Niederator, MCPC-McEllis, Wild McEllis, McBits, etc.

In addition, lattice-based encryption is an encryption based on the difficulty of calculating on a lattice, and is designed based on the difficulty of calculating problems such as LWE. It is not only strong in stability because it is based on safety in a mathematical problem called NP-hard. However, when implemented in a programming language such as C language or C++ language, it has the advantage of high computational efficiency because all operations are multiplication and addition of matrices. For example, NTRU, SS-NTRU, BLISS, New Hop, and NTRU-Prime. , LWE-Prodo, etc.

Meanwhile, isogeny-based encryption is an encryption based on the difficulty of finding the isogeny that exists between two elliptic curves of the same order, and may include, for example, protocols such as Diffie-Hellman, SIDH, etc.

In addition, hash-based encryption is an electronic signature system based on the safety of the hash function. An electronic signature scheme is constructed using a cryptographic hash function, and safety can be guaranteed by the collision resistance of the hash function used, e.g. For example, it may include W-OTS, W-OTS+, HORS, SPHINCS, etc.

The server 300 can receive status data generated by the data collection device 200 for a firefighting receiver and output from the data output unit 240 and store the status data. Additionally, when there is a video request from the manager terminal 400, the server 300 may transmit the stored state data to the manager terminal 400. For this purpose, it is desirable that the server 300 and the administrator terminal 400 are also connected to communication.

The administrator terminal 400 can connect to the server 300 and receive status data from the server 300. In this case, the manager terminal 400 may display status data on the screen 410 of the manager terminal 400 so that the manager of the firefighting receiver 100 can remotely monitor the firefighting receiver 100.

In addition, when status data of the data collection device 200, server 300, and administrator terminal 400 is transmitted and received, the status data includes sensitive data related to the status or control of the firefighting receiver at a specific location. , if this state data is exposed or manipulated by external hacking, etc., large social costs may occur. To solve this problem, it is desirable to apply an appropriate encryption scheme when transmitting and receiving the status data. More specifically, it is desirable for this encryption scheme to use an on way function, such as a hash function.

In the case of a hash function, it usually performs the function of converting various data inputs into fixed-length numeric strings (strings in the form of numbers), and has the characteristic of not producing the same output for different inputs, preventing forgery and falsification of data. Provides an easily recognizable effect.

Therefore, a hash function can be used when transmitting and receiving status data of the data collection device 200, server 300, and administrator terminal 400, and in one embodiment, MD5 (Message Digest algorithm 5) or SHA-1 (Secure Hash Algorithm) algorithm can be used.

In another embodiment, considering the frequent hacking attempts on the integrity of the MD5 or SHA-1 algorithms, applying the ANN (Aritificial Neural Network) hash algorithm can also be considered. That is, among the multi-layered layers of ANN, the input value of the state data (in the form of numbers and letters) as well as the output value of the hash function are learned in the input layer, so that the output value of the hash function is reflected when calculating the hidden layer, By reflecting this in the result value (message digest) of the output layer, the unique characteristics of each ANN (differences in the activation function of each neuron/layer) are reflected in the result value, and external hacking, etc., is more effective than the conventional hash function. It can provide an effect that makes it even more difficult.

Figure 3 is an example diagram showing status data indicating that the firefighting receiver 100 has detected a fire displayed on the screen 410 of the manager terminal 400. As shown in Figure 3, as status data related to fire detection is displayed on the screen 410 of the manager terminal 400, the manager can immediately check whether a fire has occurred remotely.

When the data generator 230 generates status data of the firefighting receiver 100, location information of the object monitored by the firefighting receiver 100 may be received from the signal collection unit 210 and included in the status data. For example, as shown in Figure 3, when a fire occurs in 'Vidia Apartment Building 101, Room 1317', the data generator 230 includes the status data related to fire detection, the target monitored by the firefighting receiver 100. Location information called 'Vidia Apartment Building 101, Room 1317' can be received from the signal collection unit 210 and included in the status data. Accordingly, the screen 410 of the manager terminal 400 displays the status of fire detection and fire information. Information on the location where the event occurred may be displayed together.

Meanwhile, as the firefighting receiver 100 malfunctions, status data related to fire detection may be displayed on the manager terminal 400 even though no fire has occurred. Accordingly, the remote fire control system 1000 may further include a closed-circuit television (CCTV) 500.

The CCTV 500 may be installed in an object monitored by the firefighting receiver 100 (for example, a room in a building, etc.), and is preferably connected to the administrator terminal 400 for communication. After the manager confirms that a fire has occurred through the screen 410 of the manager terminal 400, communication can be connected to the CCTV 500 through the manager terminal 400, and at this time, the CCTV 500 is connected to the fire receiver 100. ) can transmit the image of the object being monitored to the manager terminal 400.

Figure 4 is an example diagram of a CCTV (500) linkage screen transmitted through the administrator terminal (400). As shown in FIG. 4, the manager can remotely check the CCTV (500) linked screen through the manager terminal 400, and as a result of the confirmation, it is confirmed whether a fire has truly occurred in the target monitored by the firefighting receiver 100. You can check it in real time.

As a result of the manager checking the CCTV (500) linkage screen, if a fire truly occurs in an object monitored by the firefighting receiver (100), the manager quickly takes necessary measures, such as calling the fire department and moving immediately to the fire scene. You can take it.

On the other hand, as a result of the manager checking the CCTV (500) linkage screen, if a fire does not occur in the target monitored by the firefighting receiver (100), status data related to fire detection is displayed on the manager terminal (400). Since it is caused by a malfunction of the firefighting receiver 100, it is desirable for the manager to initialize the operation of the firefighting receiver 100 so that the firefighting receiver 100 can operate normally again.

Figure 5 is an example of a recovery button that can initialize the operation of the firefighting receiver 100 provided on the screen of the administrator terminal 400. The manager who confirms on the CCTV (500) linkage screen that a fire has not occurred in the target monitored by the fire receiver (100) can click the "Recovery" button shown in FIG. 5, and in this case, the manager terminal (400) A recovery signal is generated in response to the click.

The recovery signal generated by the manager terminal 400 may be transmitted to the firefighting receiver data collection device 200 via the server 300. Here, the recovery signal is output to the outside (e.g., the server 300 and the manager terminal 400) in response to the status data output by the data output unit 240 to the outside (e.g., the manager terminal 400 and the Corresponds to a signal transmitted from the server 300). The data collection device 200 for a firefighting receiver may further include a recovery signal receiver 350, where the recovery signal receiver 350 receives data transmitted from the outside (e.g., the administrator terminal 400 and the server 300). When a recovery signal is input, the operation of the firefighting receiver 100 is initialized according to the recovery signal so that the firefighting receiver 100 can operate normally again.

As described above, although the present invention has been described with reference to limited embodiments and drawings, the present invention is not limited to the above embodiments, and those skilled in the art can make various modifications and modifications from these descriptions. Transformation is possible. For example, above, only machine learning to predict the state of the firefighting receiver 100 using the GAN learning model was described, but CNN (Convolution Neural Network), DCNN (Deep Convolution Neural Network), and DNN (Deep Neural Network) , machine learning models other than GAN, such as RNN (Recurrent Neural Network), RBM (Restricted Boltzmann Machine), DBN (Deep Belief Network), SSD (SingleShot Detector), and YOLO (You Only Look Once), can be used. Therefore, the technical idea of the present invention should be understood only by the claims, and all equivalent or equivalent modifications thereof shall fall within the scope of the technical idea of the present invention.

100: Fire receiver
200: Data collection device for firefighting receiver
210: signal collection unit
220: signal preprocessing unit
230: Data generation unit
240: data output unit
250: recovery signal receiver
300: server
400: Administrator terminal
410: Screen (of administrator terminal)
500: CCTV
1000: Remote fire control system

Claims (15)

A signal collection unit that collects communication signals inside the firefighting receiver;
a data generation unit that generates status data of the firefighting receiver using the communication signal collected by the signal collection unit; and
It includes a data output unit that outputs the status data generated in the data generation unit to the outside,
The data generator,
Perform machine learning to predict the state of the fire-fighting receiver using the communication signal collected by the signal collection unit, and generate state data of the fire-fighting receiver using the results of the machine learning,
The signal collection unit,
The firefighting receiver captures and receives signals transmitted and received, but when electromagnetic waves are captured remotely, the electromagnetic waves generated from the signals transmitted and received by the firefighting receiver are remotely captured for a preset time according to the signal strength level of the measured electromagnetic waves. Generates electromagnetic signal flows in order, clusters selected record sets within the signal flow according to preset criteria,
It includes a plurality of buffers for storing recordsets selected within a clustered signal flow, and when storing the recordsets, data switching is performed so that recordsets of different clusters can be stored in different buffers according to the signal strength level of the electromagnetic wave. When a record set is selected within the clustered signal flow, packet identifiers are individually assigned, and each assigned packet identifier is assigned to a plurality of buffers and individually stored,
The data generator,
A learning module including a data input layer, a plurality of weighting layers, and an inferred value output layer, wherein signal data collected through the signal collection unit is input to the data input layer, and each weighting layer is input to the data input layer. Weights are added to the signal data, and learning is performed in such a way that the weights are added through the inferred value output layer and the calculated value is output to generate an operating state data generation algorithm for the firefighting receiver,
The operating state data generation algorithm generated through the learning module is input back into the general-purpose inference engine to learn to improve inference accuracy, and the operating state data generation algorithm generated through learning of the learning module is input to an independent higher level inference engine. It is converted into a level language executable code, and the executable code is separated into general purpose language (GPL: general purpose language) code and domain specific language (DSL: dmmain specific language) code depending on whether accelerated operations are needed, and the separated general-purpose language Generate a code and a domain-specific language code as a target code corresponding to the firefighting receiver,
The data output unit,
Quantum-resistant encryption using public key encryption is performed to transmit and receive data, but when transmitting and receiving the state data, a hash function, which is a one-way function, is used, and MD5 (Message Digest algorithm 5) is used. , applying at least one hash algorithm selected from SHA-1 (Secure Hash Algorithm) and ANN (Aritificial Neural Network)
Data acquisition device for firefighting receivers.
delete According to paragraph 1,
It further includes a signal pre-processing unit that performs a labeling operation on the communication signal collected by the signal collection unit,
The data generation unit is characterized in that it performs machine learning to predict the state of the fire protection receiver using a communication signal for which labeling has been performed by the signal preprocessor.
According to paragraph 3,
The signal preprocessor,
A data collection device for a fire-fighting receiver, characterized in that different labeling operations are performed on a communication signal when the fire-fighting receiver detects a fire and a communication signal when the fire-fighting receiver does not detect a fire.
According to paragraph 1,
The signal collection unit,
Further collect at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and oil stain presence data on the fire-fighting receiver,
The data generator,
In addition to the communication signal collected by the signal collection unit, at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and oil stain presence data on the fire-fighting receiver is used. A data collection device for a firefighting receiver, characterized in that it performs machine learning to predict the state of the firefighting receiver.
According to paragraph 1,
The data generator,
A data collection device for a firefighting receiver, characterized in that it performs machine learning to predict the state of the firefighting receiver using a GAN learning model. According to paragraph 1,
When receiving a recovery signal transmitted from the outside in response to the status data output to the outside by the data output unit, further comprising a recovery signal collection unit for initializing the operation of the fire protection receiver according to the recovery signal, for a fire protection receiver. Data collection device.
Firefighting receivers that monitor whether a fire is occurring;
A data collection device for a firefighting receiver that collects communication signals inside the firefighting receiver and generates status data of the firefighting receiver;
a server that stores status data generated by the data collection device for the firefighting receiver; and
It includes an administrator terminal that receives status data from the server and displays the status data on a screen so that the manager of the fire receiver can monitor the fire receiver,
The data collection device for the firefighting receiver is,
A signal collection unit that collects communication signals inside the firefighting receiver;
A data generation unit that performs machine learning to predict the state of the fire-fighting receiver using the communication signal collected by the signal collection unit, and generates state data of the fire-fighting receiver using the results of the machine learning; and
It includes a data output unit that outputs the status data generated in the data generation unit to the outside,
The signal collection unit,
The firefighting receiver captures and receives signals transmitted and received, but when electromagnetic waves are captured remotely, the electromagnetic waves generated from the signals transmitted and received by the firefighting receiver are remotely captured for a preset time according to the signal strength level of the measured electromagnetic waves. Generates electromagnetic signal flows in order, clusters selected record sets within the signal flow according to preset criteria,
It includes a plurality of buffers for storing recordsets selected within a clustered signal flow, and when storing the recordsets, data switching is performed so that recordsets of different clusters can be stored in different buffers according to the signal strength level of the electromagnetic wave. When a record set is selected within the clustered signal flow, packet identifiers are individually assigned, and each assigned packet identifier is assigned to a plurality of buffers and individually stored,
The data generator,
A learning module including a data input layer, a plurality of weighting layers, and an inferred value output layer, wherein signal data collected through the signal collection unit is input to the data input layer, and each weighting layer is input to the data input layer. Weights are added to the signal data, and learning is performed in such a way that the weights are added through the inferred value output layer and the calculated value is output to generate an operating state data generation algorithm for the firefighting receiver,
The operating state data generation algorithm generated through the learning module is input back into the general-purpose inference engine to learn to improve inference accuracy, and the operating state data generation algorithm generated through learning of the learning module is input to an independent higher level inference engine. It is converted into a level language executable code, and the executable code is separated into general purpose language (GPL: general purpose language) code and domain specific language (DSL: dmmain specific language) code depending on whether accelerated operations are needed, and the separated general-purpose language Generate a code and a domain-specific language code as a target code corresponding to the firefighting receiver,
The data output unit,
Quantum-resistant encryption using public key cryptography is performed to transmit and receive data, but when transmitting and receiving the state data, a hash function, which is a one-way function (on way function), is used, and MD5 (Message Digest algorithm 5) is used. , applying at least one hash algorithm selected from SHA-1 (Secure Hash Algorithm) and ANN (Aritificial Neural Network)
Remote fire control system.
delete According to clause 8,
The data collection device for the firefighting receiver is,
It further includes a signal pre-processing unit that performs a labeling operation on the communication signal collected by the signal collection unit,
A remote fire control system, characterized in that the data generator performs machine learning to predict the state of the firefighting receiver using a communication signal for which labeling has been performed by the signal preprocessor.
According to clause 10,
The signal preprocessor,
A remote fire control system, characterized in that different labeling operations are performed on a communication signal when the firefighting receiver detects a fire and a communication signal when the firefighting receiver does not detect a fire.
According to clause 8,
The signal collection unit,
Further collect at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and oil stain presence data on the fire-fighting receiver,
The data generator,
In addition to the communication signal collected by the signal collection unit, at least one of illuminance data around the fire-fighting receiver, smoke amount data around the fire-fighting receiver, dust amount data around the fire-fighting receiver, and oil stain presence data on the fire-fighting receiver is used. A remote firefighting control system, characterized in that it performs machine learning to predict the state of the firefighting receiver.
According to clause 8,
The data generator,
A remote firefighting control system, characterized in that it performs machine learning to predict the state of the firefighting receiver using a GAN learning model.
According to clause 8,
The data collection device for the firefighting receiver is,
When receiving a recovery signal transmitted from the outside in response to the status data output externally by the data output unit, the remote fire control control further includes a recovery signal collection unit that initializes the operation of the firefighting receiver according to the recovery signal. system.
According to clause 8,
The remote fire control system is,
When the manager terminal is connected to communication, a remote fire control system further comprising a CCTV that provides an image of a target monitored by the fire receiver to the manager terminal.