KR102266114B1 - Method and apparatus for determining frequency usage ratio by using artificial intelligence in radio monitoring system - Google Patents

Abstract

The present invention relates to a method for determining a frequency usage ratio in a radio measurement system to efficiently determine the frequency usage rate. According to the present invention, a method for operating the radio measurement system comprises the following steps: generating feature information from spectrum data for a plurality of bands; classifying the bands into a plurality of types from the feature information by using a first machine learning model; verifying at least a portion of a classification result into the plurality of types by using a second machine learning model; storing spectrum data of at least one band in which a signal is determined to exist on the basis of classification and verification result using the first machine learning model and the second machine learning model; and transmitting analysis data generated on the basis of the spectrum data of the at least one band to a management device.

Description

METHOD AND APPARATUS FOR DETERMINING FREQUENCY USAGE RATIO BY USING ARTIFICIAL INTELLIGENCE IN RADIO MONITORING SYSTEM

The present invention relates to a radio wave measurement system, and to a method and apparatus for determining a frequency utilization rate using artificial intelligence (AI) technology in the radio wave measurement system.

Radio waves are also called radio waves, and are a kind of electromagnetic waves generated according to the flow of electric and magnetic fields. Radio waves are widely used as a means for signal transmission in broadcasting systems such as radio and television, radar, navigation, computer networks, communication systems such as mobile phones, medical devices, industrial control devices, and the like. Moreover, due to the rapid development of wireless communication technology in recent years, a very large number of radio waves are used in real life. For example, cellular communication base stations, wireless LAN access points (APs), etc. that can be easily found in the vicinity perform wireless communication using radio waves.

As the number of various systems using radio waves increases, the radio wave use environment becomes more complex, and interference problems between the systems may occur. To counter this problem, radio wave measurements are needed. For example, when constructing a new wireless network or installing a new base station, a network designer may design a cell or set a base station to minimize interference based on data obtained through radio wave measurement. In addition, even when a wireless network is being operated, radio wave measurement may be performed to monitor the use state of the frequency band.

An object of the present invention is to provide a method and apparatus for effectively determining a frequency utilization rate in a radio wave measurement system.

An object of the present invention is to provide a method and apparatus for determining a frequency utilization rate using artificial intelligence (AI) in a radio wave measurement system.

An object of the present invention is to provide a method and apparatus for providing information on frequency utilization and electromagnetic wave strength, which are important services in radio wave management, using a radio wave big data platform and machine learning in a radio wave measurement system.

The technical problems to be achieved in the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention belongs from the following description. will be able

According to an embodiment of the present invention, a method of operating a radio wave measurement system includes generating feature information from spectral data for a plurality of bands, and selecting the bands from the feature information using a first machine learning model. Classifying into types, verifying at least a portion of the classification results into the plurality of types using a second machine learning model, classification and verification using the first machine learning model and the second machine learning model Based on the result, it may include storing the spectrum data of at least one band in which the signal is determined to exist, and transmitting the analysis data generated based on the spectrum data of the at least one band to a management device. have.

According to an embodiment of the present invention, the types include a signal band in which a signal of a communication system to be detected is found, a noise band in which a signal of the communication system is not found, and a signal size in which a signal of the communication system is found. may include a weak signal band falling within a predefined range from the magnitude of the noise.

According to an embodiment of the present invention, the step of verifying at least a portion of the classification result into the plurality of types using the second machine learning model may include: The method may include determining whether the first band does not belong to the second type based on the spectrum data for the first band.

According to an embodiment of the present invention, in the method, when the first band is corrected to belong to the second type using the second machine learning model, the spectral data for the first band and the second The method may further include learning the first machine learning model by using the inference result of the machine learning model.

According to an embodiment of the present invention, a radio wave measurement system includes a measuring device that processes a signal received through an antenna, and a data server that analyzes data provided from the measuring device and provides the analyzed data. The measuring device generates characteristic information from spectral data for a plurality of bands, classifies the bands into a plurality of types from the characteristic information using a first machine learning model, and some of the plurality of types Spectral data of bands belonging to the types are transmitted to the data server. The data server verifies a classification result for at least one of the partial types using a second machine learning model, and applies the classification and verification results using the first machine learning model and the second machine learning model. It is possible to store spectrum data of at least one band in which a signal is determined to exist based on the data, and provide analysis data generated based on the spectrum data of the at least one band.

According to an embodiment of the present invention, the data server determines that the first band belongs to the second type based on the spectrum data for the first band classified as the first type by the first machine learning model. It can be determined whether or not

According to an embodiment of the present invention, when the first band is corrected to belong to the second type using the second machine learning model, spectral data for the first band and the second machine learning model The inference result may be used to train the first machine learning model.

According to an embodiment of the present invention, the characteristic information includes a difference between a maximum value and a minimum value of the spectral power standard deviation within a band, a difference between the maximum and minimum values of a maximum spectrum within a channel band, and a maximum value of an average spectrum within a band. and the minimum difference, the difference between the minimum of the maximum spectrum and the maximum of the minimum spectrum, the difference between the minimum of the average spectrum and the maximum of the minimum spectrum, the maximum of the standard deviation of the spectral power within the band, the minimum of the maximum spectrum and the maximum of the average spectrum at least one of the differences.

According to an embodiment of the present invention, the analysis data may include at least one of frequency utilization rate information for each band for the plurality of bands, and electromagnetic wave intensity information for each band.

According to the present invention, the frequency utilization rate can be determined more effectively in the radio wave measurement system.

The effects obtainable in the present invention are not limited to the above-mentioned effects, and other effects not mentioned may be clearly understood by those of ordinary skill in the art to which the present invention belongs from the following description. will be.

1 shows an example of frequency utilization measurement based on a threshold level.
2 shows a radio wave measurement system according to an embodiment of the present invention.
3 shows an implementation example of a radio wave measurement system according to an embodiment of the present invention.
4 shows a functional configuration of a radio wave sensor and radio wave big data platform in a radio wave measurement system according to an embodiment of the present invention.
5A illustrates an example of maximum/minimum/average of spectral power of a band in which a measurable signal does not exist in the radio wave measurement system according to an embodiment of the present invention.
5B illustrates an example of a standard deviation of spectral power and phase of a band in which a measurable signal does not exist in the radio wave measurement system according to an embodiment of the present invention.
5C illustrates an example of maximum/minimum/average of spectral power of a band in which a measurable signal exists in the radio wave measurement system according to an embodiment of the present invention.
5D illustrates an example of a standard deviation of spectral power and phase of a band in which a measurable signal exists in the radio wave measurement system according to an embodiment of the present invention.
6 illustrates an example of an interface for providing an analysis result in a radio wave measurement system according to an embodiment of the present invention.
7 illustrates a procedure for a measurement operation in a radio wave measurement system according to an embodiment of the present invention.

The following is merely illustrative of the principles of the invention. Therefore, those skilled in the art can devise various devices that, although not explicitly described or shown herein, embody the principles of the invention and are included in the spirit and scope of the invention. In addition, it should be understood that all conditional terms and examples listed herein are, in principle, expressly intended only for the purpose of understanding the inventive concept and are not limited to the specifically enumerated embodiments and states as such. .

The above-described objects, features and advantages will become more apparent through the following detailed description in relation to the accompanying drawings, and accordingly, those of ordinary skill in the art to which the invention pertains will be able to easily practice the technical idea of the invention. .

In the specification and claims, terms such as "first," "second," "third," and "fourth," are used, if any, to distinguish between like elements, and are not necessarily used in a specific sequence. or to describe the order of occurrence. It will be understood that the terms so used are interchangeable under appropriate circumstances to enable the embodiments of the invention described herein to operate, for example, in sequences other than those shown or described herein. Likewise, where methods are described herein as including a series of steps, the order of those steps presented herein is not necessarily the order in which those steps may be performed, and any described steps may be omitted and/or Any other steps not described may be added to the method.

Also, terms such as "left", "right", "front", "behind", "top", "bottom", "above", "below" in the specification and claims are used for descriptive purposes, It is not necessarily intended to describe an invariant relative position. It will be understood that the terms so used are interchangeable under appropriate circumstances to enable the embodiments of the invention described herein to operate otherwise than, for example, as shown or described herein. As used herein, the term “connected” is defined as being directly or indirectly connected in an electrical or non-electrical manner. Objects described herein as being "adjacent" to one another may be in physical contact with one another, in proximity to one another, or in the same general scope or area as appropriate for the context in which the phrase is used. The presence of the phrase “in one embodiment” herein refers to the same, but not necessarily, embodiment.

In addition, in the specification and claims, references to 'connected', 'connecting', 'fastened', 'fastening', 'coupled', 'coupled', etc., and various variations of these expressions, refer to other elements directly It is used in the sense of being connected or indirectly connected through other components.

In addition, the suffixes "module" and "part" for the components used in this specification are given or mixed in consideration of the ease of writing the specification, and do not have distinct meanings or roles by themselves.

In addition, the terms used herein are for the purpose of describing the embodiments and are not intended to limit the present invention. As used herein, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, 'comprise' and/or 'comprising' means that a referenced component, step, operation and/or element is the presence of one or more other components, steps, operations and/or elements. or addition is not excluded.

In addition, in the description of the invention, if it is determined that a detailed description of a known technology related to the invention may unnecessarily obscure the gist of the invention, the detailed description thereof will be omitted. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Hereinafter, the present invention proposes a method and an apparatus for determining a frequency utilization rate in a radio wave measurement system. In particular, the present invention proposes a technology for providing information on frequency utilization and electromagnetic wave strength derived from radio wave big data storage and analysis using a radio wave big data platform and machine learning.

The present era is a time when various services using big data and artificial intelligence (hereinafter referred to as 'AI') are being provided everywhere in society. From the radio wave management point of view, a new radio wave big data platform and AI technology are needed to automate existing tasks, measurement, and analysis. Measurement and analysis necessary for radio wave management were performed either by direct intervention of the existing operator or based on a scenario, and the result was composed of summarized measurement result parameters and result data for violations, rather than storing all data. became In the case of measurement-oriented work, similarity analysis that can be inferred from a lot of past data, case analysis, and time series analysis that can be analyzed from continuous data are impossible, and accordingly, the need for a new storage method and analysis method has been raised. .

According to various embodiments of the present invention, a radio wave big data platform and machine learning (ML) that can be stored and analyzed by collecting full-band (eg, 20 MHz to 7500 MHz) spectrum data at regular intervals using a radio wave sensor (eg, a receiver) ) to improve the automation and efficiency of radio wave management tasks and the reliability of analysis results. Here, the radio wave big data platform consists of a storage/analysis clustering server that stores and analyzes all of the full-band spectrum transmitted from spatially dispersed radio wave sensors that collect radio wave data (eg, spectrum data for each band).

Frequency utilization is an indicator of how much a licensed band (eg, channel) is used for radio wave management. Frequency utilization is necessary for radio wave management. For example, the frequency utilization rate may be utilized for frequency recovery and reallocation of the allocated and allocated service bands. Regarding the definition of frequency utilization rate in terms of radio wave management, measurement methods, and considerations during measurement, the International Telecommunication Union-Radiocommunication sector (ITU-R), an international standard organization, 'Recommendation ITU-R SM. 1880, Report ITU-R SM.2256' and other recommendations and reports were enacted, and each country is measuring and analyzing frequency resources according to the situation of each country based on the recommendations of ITU-R. Currently, in Korea, the frequency utilization rate is measured at the Central Radio Management Office under the Ministry of Science and ICT.

The frequency utilization rate is derived as the ratio of the time used during the total period (eg, the user's inquiry period) by measuring whether it is used by unit measurement time. In addition, in calculating the frequency utilization rate, the accuracy of the threshold level for determining whether an allocated band (eg, a channel) is used is very important. For example, a margin value (eg, 3dB, 6dB, 9dB, 12dB, etc.) above the noise level measured for each band is applied, or a margin level based on the operator's experience is applied. Accurate measurement of the noise level, which is a reference used as the threshold level, or setting the threshold level based on user experience has a great influence on the frequency utilization measurement accuracy. In addition to the threshold level, a method for calculating channel power, such as a spectrum measurement period or a spectrum measurement method according to the characteristics of a signal, may also affect the accuracy, but the present invention focuses on the criterion for determining the occupancy of a signal.

According to a general frequency utilization measurement technique, an operator arbitrarily sets a margin value above the measured noise level or sets it to a level based on experience, and then measures the frequency utilization ratio based on a threshold value reflecting the set margin value. When a signal greater than or equal to the determined threshold is found, it is determined that the corresponding frequency is used, and the frequency usage rate is calculated as a value for how many measurements exceed the threshold compared to the total number of measurements. That is, the above-described method relies on empirical values such as the inaccuracy of the noise level and the margin value of the threshold level set by the operator.

1 shows an example of frequency utilization measurement based on a threshold level. Referring to FIG. 1 , it is determined that the bands 101 and 102 in which the value of the channel power is greater than or equal to the threshold level are used as the channel. By counting the number of times above the threshold level several times during a certain measurement period, and calculating the ratio of the counted times, the frequency utilization rate is derived. In this case, since only the measurement result is stored and the spectrum data itself is not generally stored, in this case, there is a limit that cannot be re-analyzed using the modified threshold level. On the other hand, the technology proposed in the present invention can extract characteristics from spectrum data for each service channel and determine the existence of a signal itself using machine learning.

2 shows a radio wave measurement system according to an embodiment of the present invention.

Referring to FIG. 2 , the radio wave measurement system includes an antenna device 210 , a measurement device 220 , a data server 230 , and a management device 240 .

The antenna device 210 is installed at a site designated for radio wave measurement to detect a radio signal. The antenna device 210 may be disposed at a position where radio wave reception is easy. For example, the antenna device 210 may be installed outdoors. The antenna device 210 includes a structure for installation and a conductor pattern for signal sensing. For example, the antenna device 210 may be implemented as a dipole antenna, a monopole antenna, a patch antenna, a horn antenna, a reflector antenna, or the like.

The measurement device 220 generates measurement data by converting the radio waves sensed by the antenna device 210 into data, and provides the measurement data to the data server 230 . Here, the measurement data may vary according to conditions set by the system operator. For example, the measurement data may include a measurement result within a set frequency range, and may include a set measurement parameter. For example, the measurement data may include at least one of frequency spectrum data, metadata, and quality measurement data. Also, the measurement device 220 may perform time synchronization with another device using a global positioning system (GPS) and synchronize data based on the time synchronization. To this end, the measuring device 220 controls the overall control of the signal detecting unit 222 for processing the signal provided from the antenna device 210 , the communication unit 224 for generating and analyzing a signal for communication, and the measuring device 220 . and a processor 226 to perform The signal detection unit 222 processes a radio frequency (RF) signal, for example, an amplifier for amplifying the signal amplitude, a filter for passing a signal of a band to be measured, an oscillator and a mixer for frequency downconversion, At least one of an analog to digital converter (ADC) for converting an analog signal into a digital signal may be included. The communication unit 224 provides a wired or wireless interface, and performs a function for transmitting and receiving data according to a corresponding communication protocol. The processor 226 analyzes the radio wave data provided from the signal detector 222 and generates a data set for transmission to the data server 230 . In addition, the processor 226 may control various functions of the measurement device 220 .

The data server 230 manages and analyzes data provided from the measurement device 220 , and provides the analysis result to the management device 240 . To this end, the data server 230 includes a communication unit 232 for generating and analyzing a signal for communication, a storage unit 234 for storing data, and a processor 236 for overall control of the data server 230 . . The communication unit 232 provides a wired or wireless interface, and performs a function for transmitting and receiving data according to a corresponding communication protocol. The storage unit 234 stores data provided from the measurement device 220 and data representing the analysis result. The processor 236 analyzes the data provided from the measurement device 220 and processes it into a set form. In addition, the processor 236 may control various functions of the data server 230 . The data server 230 may be implemented in the form of a cloud server accessible through an external network (eg, an Internet network). The data server 230 may be implemented as a set of a plurality of hardware.

The management device 240 is a terminal device of a user or an operator that accesses the data server 230 . The management device 240 may request analysis data from the data server 230 , and receive and display the analysis data from the data server 230 . Also, the management device 240 may transmit a command for controlling settings or operations of the measurement device 220 and the data server 230 . To this end, the management device 240 includes a communication unit 242 that generates and interprets a signal for communication, an input/output unit 244 for interaction with a user, and a processor 246 that controls overall functions of the management device 240 . ) may be included. For example, the input/output unit 244 may include at least one of an input unit such as a button, a keyboard, a touch screen, and a microphone, and an output unit such as a screen, a light emitting diode (LED), and a speaker. According to a user's input input through the input/output unit 244 , the processor 246 generates a message requesting analysis data, and the communication unit 242 transmits the message to the data server 230 . When the analysis data is received from the data server 230 through the communication unit 242, the processor 246 generates output data for displaying the analysis data, and the input/output unit 244 outputs information based on the output data. can For example, the information may be output in the form of graphics or images.

In the embodiment of FIG. 2 , the measurement device 220 and the data server 230 have been described as separate devices. However, if necessary, the measurement device 220 and the data server 230 may be implemented as one piece of hardware. In this case, the communication between the communication unit 224 and the communication unit 232 is replaced with data exchange inside the device, and the analysis data request and response procedure between the measurement device 220 and the management device 240 may be performed.

The radio wave measurement system configured as shown in FIG. 2 may be composed of a plurality of spatially dispersed radio wave sensors and a clustered storage/analysis server, and may support radio wave big data collection/storage/analysis. A specific implementation example is described with reference to FIG. 3 .

3 shows an implementation example of a radio wave measurement system according to an embodiment of the present invention. Referring to FIG. 3 , the radio wave measurement system includes radio wave sensors 310 , a radio wave big data platform 320 , and an operation/analysis server 330 . Compared with FIG. 2 , the radio wave sensors 310 include the antenna 210 and the measurement device 220 of FIG. 2 , and the radio wave big data platform 320 includes the data server 230 of FIG. 2 , The operation/analysis server 330 may include the management device 240 of FIG. 2 .

The radio wave sensors 310 receive a radio signal, and obtain radio wave data (eg, spectrum data for each band) from the received radio signal. The spectral data collected by the radio wave sensors 310 is data of a spectral region having a constant bin interval, and includes spectral power and phase information. By repeatedly performing fast Fourier transform (FFT) within the same frequency band, the radio wave sensors 310 store statistical data such as maximum, minimum, average, and standard deviation of spectral power for each frequency bin. The radio wave sensors 310 generate new information capable of representing signal characteristics from data of a band channelized for each service from the stored data. The radio wave sensors 310 divide each band into a plurality of types, for example, a signal band, a noise band, and a weak signal band, using a machine learning model that determines the presence or absence of a signal for each band from radio wave data. Then, only the radio wave data for the signal band and the weak signal band are transmitted to the radio wave big data platform 320 . At this time, since information on some bands (eg, noise band) is not transmitted, the effect of data compression occurs.

The radio wave big data platform 320 may be implemented using a plurality of devices (eg, a server, a virtual machine, etc.) and various software modules executed in them. For example, the radio wave big data platform 320 includes an intermediate computer program module (eg, a message broker, an interface engine) for message protocol conversion, a database management system, a cluster computing framework, and streaming. It can be implemented using a data processing module. The radio wave big data platform 320 is connected to the radio wave sensors 310 through the external network 340 .

The method of storing all radio wave data enables various analyzes such as replaying past data, current signal patterns and similar cases, time series data analysis continuous from the past, and new analysis of past data by applying new parameters. However, the cost burden such as storage capacity for storing radio wave data and securing a data network may increase. Therefore, the system according to an embodiment of the present invention determines the presence or absence of a signal using in-band machine learning and then extracts only parameters such as noise level from the band determined to be a noise band, and stores only meaningful data such as a signal band or a weak signal band By doing so, it is possible to provide effects such as storage space reduction, analysis time reduction, and network maintenance. In this case, according to an embodiment, machine learning may follow a supervised learning method using labeling data indicating whether or not a channel in which a signal exists.

According to the above-described embodiment, the measured band is divided into a signal band, a weak signal band, and a noise band by the machine learning model. The signal band is a band in which a signal of the communication system to be detected is found, and the noise band is a band in which a signal of the communication system is not found. The weak signal band means a band having a spectral distribution pattern that is small enough to be distinguished from the signal band but has a spectrum distribution pattern different from that of the noise band. For example, the weak signal band may be a band in which a signal of a communication system exists, but its magnitude is similar to noise. Here, it can be understood that the magnitude similar to the noise is within a predefined range from the magnitude of the noise, or the difference between the magnitude of the corresponding signal and the magnitude of the noise is on average less than or equal to a threshold. Therefore, the training data for training the machine learning model is labeled data, and may include radio wave data of a signal band, radio wave data of a weak signal band, and radio wave data of a noise band.

Here, the labeled learning data may be obtained by various methods. According to an embodiment of the present invention, by clustering various radio wave data through unsupervised learning, it can be classified into signal band data, weak signal band data, and noise band data, and the classification result can be labeled. Alternatively, according to another embodiment of the present invention, results inferred by other machine learning models may be used as training data. According to another embodiment of the present invention, manually labeled data may be used as training data.

4 shows a functional configuration of a radio wave sensor and radio wave big data platform in a radio wave measurement system according to an embodiment of the present invention. 4 illustrates a data flow for determining the presence or absence of a signal using data collected from a radio wave sensor and deriving an analysis result from the radio wave big data platform.

Referring to FIG. 4 , the radio wave sensor 410 includes a feature extraction unit 411 and a signal determination unit 412 . The feature extractor 411 generates feature information for each band from the input spectrum data. The signal determination unit 412 divides the bands into a signal band, a weak signal band, and a noise band based on the characteristic information. According to an embodiment of the present invention, the signal determiner 412 may classify the bands using the machine learning model #1. That is, the machine learning model #1 classifies the bands into a signal band, a noise band, and a weak signal band by determining whether there is a signal. In addition, the signal determination unit 412 provides only radio wave data for the signal band and the weak signal band among the signal band, the noise band, and the weak signal band to the radio wave big data platform 420 .

The radio wave big data platform 420 includes a classification unit 421 , a signal determination unit 422 , a storage unit 423 , a frequency utilization analysis unit 424 , and an electromagnetic wave strength analysis unit 425 . The classification unit 421 classifies the radio wave data for the signal band and the radio data for the weak signal band in the data provided from the radio wave sensor 410, and divides the radio wave data for the weak signal band into the signal determining unit 422, The signal band data is provided to the storage unit 423 . The signal determining unit 422 determines again whether the signal band is not a signal band based on the radio wave data for the weak signal band. According to an embodiment of the present invention, the signal determination unit 422 may determine again whether the signal band using the machine learning model #2. If, as a result of inference using the machine learning model #2 in the signal determination unit 422, the band determined as a weak signal band by the signal determination unit 412 is corrected to a signal band, data is provided to the storage unit 423 . According to another embodiment of the present invention, the signal determination unit 412 may provide not only the data for the signal band but also the data for the weak signal band to the storage 423 . And, the data for the band corrected as the signal band and the inference result of the machine learning model #2 may be used as training data for improving the machine learning model #1 of the signal determining unit 412 . Here, the learning of the machine learning model #1 may be performed by the radio wave sensor 410 , the radio wave big data platform 420 , or may be performed by a third party device. The storage unit 423 stores radio wave data (eg, spectrum data) of a signal band and/or a band determined to be a weak signal band. The frequency utilization analyzer 424 and the electromagnetic wave strength analyzer 425 analyze the frequency utilization rate and the electromagnetic wave intensity by using the radio wave data stored in the storage 423 .

As described with reference to FIG. 4 , machine learning model #1 and machine learning model #2 are used to determine the signal band. When deep learning technology is adopted, machine learning model #1 includes a first artificial neural network for classifying a signal band, a noise band, and a weak signal band based on spectral data, and the machine learning model #2 includes a second artificial neural network for verification of a band determined as a weak signal band by the machine learning model #1. According to an embodiment, as in the example of FIG. 4 , the machine learning model #1 may be used by the radio wave sensor 410 . Since the radio wave sensor 410 may have lower hardware performance compared to the radio wave big data platform 420, the first artificial neural network may have a simpler structure than the second artificial neural network in consideration of the burden of computational complexity and computational complexity. have. For example, the first artificial neural network may have fewer layers or fewer nodes than the second artificial neural network. Alternatively, the first artificial neural network may have a simpler structure than the second artificial neural network. Specifically, the second artificial neural network may be composed of a combination of a plurality of sub artificial neural networks, and the first artificial neural network may be composed of a single artificial neural network or a smaller number of sub-artificial neural networks than the second artificial neural network. have.

In the embodiment described with reference to FIG. 4 , the machine learning model #1 has been described as a part of the radio wave sensor 410 . According to another embodiment, the machine learning model #1 may be included in a device other than the radio wave sensor 410, and the radio wave sensor 410 transmits input data necessary for an inference operation of the machine learning model #1 to another device. , and another device may provide a result inferred using the machine learning model #1 to the radio wave sensor 410 . That is, the machine learning model #1 exists outside the radio wave sensor 410 , and the radio wave sensor 410 may use the machine learning model #1 remotely. In this case, greater computational power may be provided compared to the case where the machine learning model #1 is a part of the radio wave sensor 410 .

5A to 5D below show examples of radio wave data. Specifically, FIG. 5A is an example of the maximum/minimum/average of spectral power of a band in which a measurable signal does not exist in the radio wave measurement system according to an embodiment of the present invention, and FIG. 5B is an embodiment of the present invention. An example of the standard deviation of spectral power and phase of a band in which a measurable signal does not exist in the radio wave measurement system, FIG. 5C is a diagram of spectral power of a band in which a measurable signal exists in the radio wave measurement system according to an embodiment of the present invention An example of maximum/minimum/average, FIG. 5D illustrates an example of standard deviation of spectral power and phase of a band in which a measurable signal exists in the radio wave measurement system according to an embodiment of the present invention. 5A and 5C, the maximum/minimum/average values of the spectral power do not deviate from a certain range when no signal is present, but the maximum/minimum/average values are relative in the corresponding frequency range when a signal is present. was observed to be large. In addition, comparing FIGS. 5B and 5D, the standard deviation of the spectral power value and the phase value does not deviate from a certain range when the signal is not present, but the spectral power value and the phase value in the corresponding frequency range when the signal is present It is confirmed that the standard deviation of

Based on the radio wave data as shown in FIGS. 5A to 5D , characteristic information such as at least one of items listed in Table 1 below may be extracted.

Item Difference with or without signal The difference between the maximum and minimum values of the spectral power standard deviation within the channel band When a signal is present, the standard deviation tends to decrease. When a signal is present, the maximum and minimum difference increases. The difference between the maximum and minimum values of the maximum spectrum within the channel band When a signal is present, the difference between the maximum and minimum values increases. If only noise is present, the difference between the maximum and minimum values is insignificant. The difference between the maximum and minimum values of the average spectrum within the channel band When a signal is present, the difference between the maximum and minimum values increases. If only noise is present, the difference between the maximum and minimum values is insignificant. The difference between the minimum value of the maximum spectrum and the maximum value of the minimum spectrum If a signal is present, it has a negative value. If only noise is present, it has a positive value. The difference between the minimum value of the average spectrum and the maximum value of the minimum spectrum If a signal is present, it has a negative value. If only noise is present, it has a positive value.

According to another embodiment, in addition to the items listed in [Table 1], at least one characteristic information of the difference between the maximum value of the standard deviation of the spectral power within the channel band, the minimum value of the maximum spectrum, and the maximum value of the average spectrum may be further used. have.

Through the radio wave measurement system according to the above-described embodiments, a user or an operator may obtain a radio wave measurement result including a frequency utilization rate. According to an embodiment of the present invention, the radio wave measurement result may be expressed as an interface as shown in FIG. 6 below.

6 illustrates an example of an interface for providing an analysis result in a radio wave measurement system according to an embodiment of the present invention. Referring to FIG. 6 , the interface includes at least one menu, at least one button, and at least one information window. For example, the usage space information window 611 indicating the usage rate within the specified frequency band in the specified area, the usage information window 612 within the period indicating usage information within the set period, and the usage rate within the specified frequency band by time zone A time series graph 613 to represent may be displayed. In addition, the interface includes a menu 621 for selecting an item to be displayed, a menu 622 for specifying display options, a menu 623 for specifying a period, and a menu 624 for selecting a database to be used. may include Also, the interface may include a screenshot button 631 for saving the measurement result as an image.

7 illustrates a procedure for a measurement operation in a radio wave measurement system according to an embodiment of the present invention. 7 illustrates a radio wave measurement method. In the following description, the subject of the operation is expressed as a 'system', but each operation is performed by one of the devices belonging to the radio wave measurement system (eg, measuring device, data server, radio wave sensor, radio wave big data platform). can also be understood as

Referring to FIG. 7 , in step 701 , the system generates characteristic information for a plurality of bands. For example, the system may detect signals received through an antenna in a frequency range including a plurality of bands, generate spectrum data from the sensed signal, and then generate characteristic information based on the spectrum data. For example, the characteristic information may include a difference between a maximum value and a minimum value of a spectral power standard deviation within a band, a difference between a maximum value and a minimum value of a maximum spectrum within a channel band, a difference between a maximum value and a minimum value of an average spectrum within a band, and a maximum spectrum includes at least one of the difference between the minimum and minimum spectral values, the difference between the average spectrum minimum and the minimum spectrum maximum, the maximum standard deviation of spectral power within a band, and the difference between the maximum spectrum minimum and the average spectrum maximum. can do.

In step 703, the system classifies the plurality of bands. That is, the system may classify a plurality of bands using the feature information. In this case, according to an embodiment, the first machine learning model is used. In other words, the system classifies the bands into a plurality of types from the feature information using the first machine learning model. Here, the plurality of types includes a signal band, a noise band, and a weak signal band. The system may input the feature information generated based on the spectral data into the first machine learning model, and determine whether each band belongs to a signal band, a noise band, or a weak signal based on the output of the first machine learning model. . The first machine learning model has been trained using the labeled data before this step.

In step 705, the system verifies the classification result. To this end, the system may use a second machine learning model. For verification of the classification result, the system may re-determine the type of the bands classified as at least one type among the classification results in step 703 . In other words, the system verifies at least a portion of the classification result into the plurality of types using the second machine learning model. According to an embodiment, the system may determine again whether the band determined as the weak signal band does not belong to the signal band using the second machine learning model. To this end, the system inputs at least one of spectral data and characteristic information of a band determined as a weak signal band to the second machine learning model, and verifies whether the corresponding band is a signal band based on the output of the second machine learning model can do.

In step 707, the system stores data for the signal bands. Here, the signal band includes at least one signal band determined by the first machine learning model, and may further include at least one signal band re-determined by the second machine learning model. The stored data may include spectral data and further include characteristic information. That is, based on the results of classification and verification using the first machine learning model and the second machine learning model, the system stores spectral data of at least one band in which a signal is determined to exist.

At 709 , the system generates analysis data. The system generates analysis data based on the stored spectral data. Here, the analysis data may be generated by a user's request and output in a form recognizable to the user. For example, the analysis data may indicate a frequency utilization rate, an electromagnetic wave strength of a signal band, and the like.

The radio wave measurement method according to the embodiment described above may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured according to the embodiment, or may be known and used by those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

Each of the drawings referenced in the description of the previous embodiment is merely an embodiment shown for convenience of description, and items, contents, and images of information displayed on each screen may be modified and displayed in various forms.

Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and those of ordinary skill in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Accordingly, the true technical protection scope of the present invention should be defined by the technical spirit of the appended claims.

Claims (10)

A method of operating a radio wave measurement system, comprising:
generating characteristic information from spectral data for a plurality of bands, wherein the characteristic information includes: a difference between a maximum value and a minimum value of a spectral power standard deviation within a band, a difference between a maximum spectrum value and a minimum value within a channel band, a band the difference between the maximum and minimum values of the average spectrum within, the difference between the minimum of the maximum spectrum and the maximum of the minimum spectrum, the difference between the minimum of the average spectrum and the maximum of the minimum spectrum, the maximum of the standard deviation of the spectral power within the band, of the maximum spectrum at least one of a difference between a minimum value and a maximum value of the average spectrum;
classifying the bands into a plurality of types from the characteristic information using a first machine learning model, wherein the types are a first type indicating a signal band in which a signal of a communication system to be sensed is found, the communication system a second type indicating a noise band in which the signal of is not found, and a third type indicating a weak signal band in which a signal of the communication system is found but the magnitude of the signal falls within a predefined range from the magnitude of the noise; ;
verifying at least a portion of the classification result into the plurality of types using a second machine learning model;
storing spectral data of at least one band determined as the first type based on a result of classification and verification using the first machine learning model and the second machine learning model; and
Transmitting analysis data including frequency utilization information generated based on the spectrum data of the at least one band to a management device,
The first machine learning model and the second machine learning model include an artificial neural network trained using radio wave data labeled as one of the first type, the second type, and the third type.
The method according to claim 1,
detecting a radio signal received through an antenna in a frequency range including the plurality of bands; and
and generating the spectral data by dataizing the radio signal.
The method according to claim 1,
The frequency utilization rate information is expressed as a ratio in which each band is determined to be the first type during a given period.
The method according to claim 1,
The step of verifying at least a portion of the classification result into the plurality of types using the second machine learning model,
and determining whether the first band does not belong to the first type based on spectral data for the first band classified as the third type by the first machine learning model.
5. The method according to claim 4,
When the first band is corrected to belong to the first type using the second machine learning model, the first machine using the spectral data for the first band and the inference result of the second machine learning model The method further comprising the step of training the learning model.
The method according to claim 1,
The first machine learning model includes a first artificial neural network,
The second machine learning model includes a second artificial neural network,
The first artificial neural network has a smaller number of layers than the second artificial neural network.
In the radio wave measurement system,
a measuring device for processing a signal received through an antenna; and
and a data server that analyzes the data provided from the measurement device and provides the analysis data,
The measuring device generates characteristic information from spectral data for a plurality of bands, classifies the bands into a plurality of types from the characteristic information using a first machine learning model, and some of the plurality of types transmit spectrum data of bands belonging to the types to the data server;
The data server verifies a classification result for at least one of the partial types using a second machine learning model, and applies the classification and verification results using the first machine learning model and the second machine learning model. Storing the spectrum data of at least one band determined as the first type based on the analysis data including frequency utilization information generated based on the spectrum data of the at least one band,
The characteristic information includes the difference between the maximum and minimum values of the spectral power standard deviation within the band, the difference between the maximum and minimum values of the maximum spectrum within the channel band, the difference between the maximum and minimum values of the average spectrum within the band, and the minimum value of the maximum spectrum. at least one of a difference between a maximum value of a minimum spectrum, a difference between a minimum value of an average spectrum and a maximum value of a minimum spectrum, a maximum value of a standard deviation of spectral power within a band, a difference between a minimum value of the maximum spectrum and a maximum value of the average spectrum,
The types are the first type indicating a signal band in which a signal of the communication system to be detected is found, the second type indicating a noise band in which a signal of the communication system is not found, and the signal of the communication system being detected. but includes a third type indicating a weak signal band in which the magnitude of the signal falls within a predefined range from the magnitude of the noise,
The first machine learning model and the second machine learning model may include an artificial neural network trained using radio wave data labeled as one of the first type, the second type, and the third type. system.
8. The method of claim 7,
The measuring device is a radio wave measuring system that detects a radio signal received through an antenna in a frequency range including the plurality of bands, and generates the spectrum data by converting the radio signal into data.
8. The method of claim 7,
The frequency utilization information is a radio wave measurement system in which each band is expressed in a ratio determined as the first type during a given period.
8. The method of claim 7,
The data server determines whether or not the first band belongs to the first type based on the spectrum data for the first band classified as the third type by the first machine learning model,
When the first band is corrected to belong to the first type using the second machine learning model, the spectral data for the first band and the inference result of the second machine learning model are, A radio measurement system used to train a model.

Images