Information processing device, information processing method, and information processing program

Abstract

To realize an information processing device for accurately identifying interference wave sources. [Solution] The information processing device comprises a feature calculation unit that calculates the characteristic quantities of radio waves, a determination unit that determines whether or not abnormal waves are present by inputting the feature quantities into an abnormality determination model, an interference wave extraction unit that extracts interference waves if it is determined that abnormal waves are present, and an interference wave classification unit that classifies the interference waves into one or more groups according to their similarity.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and an information processing program.

Background Art

[0002] Techniques related to the removal of interference waves in wireless communication are known. For example, Patent Document 1 discloses a wireless communication system including a transmission device and a reception device. In this wireless communication system, the transmission device generates a replica signal of an output signal, and the reception device generates a replica signal of an interference wave using the replica signal as a reference signal.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the wireless communication system described in Patent Document 1, when there are a plurality of interference waves, each interference wave cannot be identified. Therefore, there is a problem that the interference wave source cannot be accurately identified in the wireless communication system described in Patent Document 1.

[0005] The present disclosure has been made in view of the above problems, and an exemplary object thereof is to provide a technique for accurately identifying an interference wave source.

Means for Solving the Problems

[0006] An information processing device relating to an exemplary aspect of this disclosure includes: a feature calculation means for calculating feature quantities of radio waves received by a receiver; a determination means for determining whether or not abnormal waves, which are radio waves outside a predetermined range, are included in the radio waves received by the receiver by inputting the feature quantities calculated by the feature calculation means into an abnormality determination model that takes the feature quantities of radio waves as input and determines whether or not the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range; an interference wave extraction means for extracting interference waves, which are radio waves outside the predetermined range, if the determination means determines that abnormal waves are included; and an interference wave classification means for classifying the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity.

[0007] An information processing method relating to an exemplary aspect of this disclosure includes: a feature calculation process in which at least one processor calculates feature quantities of radio waves received by a receiver; a determination process in which the at least one processor takes the feature quantities of radio waves as input to an anomaly determination model that determines whether the feature quantities of radio waves include feature quantities of radio waves outside a predetermined range, thereby determining whether the radio waves received by the receiver include an anomaly wave which is a radio wave outside the predetermined range; an interference wave extraction process in which, if the determination process determines that the anomaly wave is included, the at least one processor extracts an interference wave which is a radio wave outside the predetermined range; and an interference wave classification process in which the at least one processor classifies the interference waves extracted in the interference wave extraction process into one or more groups according to their similarity.

[0008] An illustrative aspect of the present disclosure relates to an information processing program that causes a computer to function as an information processing device, wherein the computer functions as: a feature calculation means for calculating feature quantities of radio waves received by a receiver; a determination means for determining whether or not the radio waves received by the receiver contain abnormal waves that are radio waves outside the predetermined range, by inputting the feature quantities calculated by the feature calculation means into an abnormality determination model that takes the feature quantities of the radio waves as input and determines whether or not the feature quantities of the radio waves contain feature quantities of radio waves outside the predetermined range; an interference wave extraction means for extracting interference waves that are radio waves outside the predetermined range if the determination means determines that abnormal waves are included; and an interference wave classification means for classifying the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity. [Effects of the Invention]

[0009] One exemplary effect of this disclosure is that it can provide a technique for accurately identifying interference sources. [Brief explanation of the drawing]

[0010] [Figure 1] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 2] This is a flowchart showing the flow of the information processing method related to this disclosure. [Figure 3] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 4] This figure shows an example of the radio wave feature quantities related to this disclosure. [Figure 5] This figure shows a map of the vectors of multiple radio wave feature quantities related to this disclosure onto a vector space. [Figure 6] This flowchart shows the flow of the learning method related to this disclosure. [Figure 7] This is a flowchart showing the flow of the generation method related to this disclosure. [Figure 8] This is a flowchart showing the flow of the information processing method related to this disclosure. [Figure 9]This is a block diagram showing the configuration of a computer that functions as an information processing device related to this disclosure. [Modes for carrying out the invention]

[0011] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining some or all of the technologies (things or methods) employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. In addition, the effects mentioned in each of the exemplary embodiments shown below are examples of effects that can be expected in that exemplary embodiment and do not define the scope of the present invention. That is, embodiments that do not produce the effects mentioned in each of the exemplary embodiments shown below may also be included in the scope of the present invention.

[0012] [First Exemplary Embodiment] A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is the basic form for each of the exemplary embodiments described later. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur. Furthermore, each technology shown in the drawings referenced to explain this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur.

[0013] (Configuration of Information Processing Device 1) The configuration of the information processing apparatus 1 will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the information processing apparatus 1. As shown in FIG. 1, the information processing apparatus 1 includes a feature quantity calculation unit 11, a determination unit 12, an interference wave extraction unit 13, and an interference wave classification unit 14. The feature quantity calculation unit 11, the determination unit 12, the interference wave extraction unit 13, and the interference wave classification unit 14 respectively implement feature quantity calculation means, determination means, interference wave extraction means, and interference wave classification means in this exemplary embodiment.

[0014] (Feature quantity calculation unit 11) The feature quantity calculation unit 11 calculates the feature quantity of the radio wave received by the receiver. The feature quantity calculation unit 11 supplies the calculated feature quantity to the determination unit 12 and the interference wave extraction unit 13.

[0015] (Determination unit 12) The determination unit 12 inputs the feature quantity of the radio wave calculated by the feature quantity calculation unit 11 into an abnormality determination model that determines whether the feature quantity of the radio wave includes the feature quantity of a radio wave outside a predetermined range. By doing so, the determination unit 12 determines whether the radio wave received by the receiver includes an abnormal wave that is a radio wave outside the predetermined range. The determination unit 12 supplies the determination result to the interference wave extraction unit 13.

[0016] (Interference wave extraction unit 13) When the interference wave extraction unit 13 determines that an abnormal wave is included by the determination unit 12, the interference wave extraction unit 13 extracts an interference wave that is a radio wave outside a predetermined range. The interference wave extraction unit 13 supplies the extracted interference wave to the interference wave classification unit 14.

[0017] (Interference wave classification unit 14) The interference wave classification unit 14 classifies the interference wave extracted by the interference wave extraction unit 13 into one or more groups according to the similarity.

[0018] (Effect of the information processing apparatus 1) As described above, the information processing device 1 employs a configuration comprising: a feature calculation unit 11 that calculates the feature quantities of radio waves received by a receiver; a determination unit 12 that takes the feature quantities of radio waves as input and inputs the feature quantities calculated by the feature calculation unit 11 to an anomaly determination model that determines whether or not the feature quantities of radio waves include feature quantities of radio waves outside a predetermined range, thereby determining whether or not the radio waves received by the receiver contain an anomaly wave that is outside a predetermined range; an interference wave extraction unit 13 that extracts interference waves that are outside a predetermined range if the determination unit 12 determines that an anomaly wave is present; and an interference wave classification unit 14 that classifies the interference waves extracted by the interference wave extraction unit 13 into one or more groups according to their similarity.

[0019] Therefore, the information processing device 1 has the effect of providing a technique for accurately identifying interference wave sources.

[0020] (Information processing method S1 flow) The flow of the information processing method S1 will be explained with reference to Figure 2. Figure 2 is a flowchart showing the flow of the information processing method S1. As shown in Figure 2, the information processing method S1 includes a feature calculation process S11, a judgment process S12, an interference wave extraction process S13, and an interference wave classification process S14.

[0021] (Feature calculation process S11) In the feature calculation process S11, the feature calculation unit 11 calculates the feature quantities of the radio waves received by the receiver. The feature calculation unit 11 supplies the calculated feature quantities to the determination unit 12 and the interference wave extraction unit 13.

[0022] (Decision process S12) In the determination process S12, the determination unit 12 takes the characteristic quantities of the radio waves as input and inputs the characteristic quantities calculated by the characteristic quantity calculation unit 11 into an anomaly determination model that determines whether the characteristic quantities of the radio waves include characteristic quantities of radio waves outside a predetermined range. The determination unit 12 then determines whether the radio waves received by the receiver contain an anomaly wave which is a radio wave outside a predetermined range. The determination unit 12 supplies the determination result to the interference wave extraction unit 13.

[0023] (Interference wave extraction process S13) In the interference wave extraction process S13, if the determination unit 12 determines that abnormal waves are present, the interference wave extraction unit 13 extracts interference waves that are outside a predetermined range. The interference wave extraction unit 13 supplies the extracted interference waves to the interference wave classification unit 14.

[0024] (Interference wave classification processing S14) In the interference wave classification process S14, the interference wave classification unit 14 classifies the interference waves extracted by the interference wave extraction unit 13 into one or more groups according to their similarity.

[0025] (Effects of information processing method S1) As described above, the information processing method S1 employs a configuration that includes: a feature calculation unit 11 which calculates the feature quantities of radio waves received by the receiver in a feature calculation process S11; a determination unit 12 which takes the feature quantities of radio waves as input and inputs the feature quantities calculated by the feature calculation unit 11 into an anomaly determination model which determines whether the feature quantities of radio waves include features outside a predetermined range in a determination process S12; an interference wave extraction unit 13 which, if the determination unit 12 determines that an abnormal wave is included, extracts the interference wave which is outside a predetermined range in an interference wave extraction process S13; and an interference wave classification unit 14 which classifies the interference waves extracted by the interference wave extraction unit 13 into one or more groups according to their similarity in an interference wave classification process S14. Therefore, the same effects as the information processing device 1 described above can be obtained with the information processing method S1.

[0026] [Second exemplary embodiment] A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same function as those described in the above-described exemplary embodiment are denoted by the same reference numerals, and their descriptions are omitted as appropriate. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise. Furthermore, each technology shown in the drawings referenced to describe this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise.

[0027] (Overview of Information Processing Device 2) The information processing device 2 is a device that, if the received radio waves (observed wave OW) contain interference waves, extracts those interference waves and classifies the extracted interference waves.

[0028] More specifically, the information processing device 2 first takes the characteristic quantities of the radio waves as input and uses an anomaly determination model AJM, which determines whether or not the characteristic quantities of the radio waves include characteristic quantities of radio waves outside a predetermined range, to determine whether or not the received radio waves contain an anomaly.

[0029] The characteristics of radio waves will be explained with reference to Figure 4. Figure 4 is a diagram showing examples of radio wave characteristics. Examples of radio wave characteristics include the spectrum, which is represented by frequency and intensity, as shown in the upper part of Figure 4; the spectrogram, which represents the frequency analyzed over time using color (grayscale in Figure 4), as shown in the center of Figure 4; and the graph of statistically analyzed information, as shown in the lower part of Figure 4.

[0030] The characteristics of radio waves will be explained with reference to Figure 4. Figure 4 is a diagram showing examples of radio wave characteristics. Examples of radio wave characteristics include the spectrum, which is represented by frequency and intensity, as shown in the upper part of Figure 4; the spectrogram, which represents the frequency analyzed over time using color (grayscale in Figure 4), as shown in the center of Figure 4; and the graph of statistically analyzed information, as shown in the lower part of Figure 4.

[0031] The features may be extracted directly from the time-domain received data, or they may be extracted for each frequency after being converted to frequency-domain data by spectralization or spectrogramming. The information processing device 2 handles these features as numerical data in vector form. The information processing device 2 may also map the spectrum or spectrogram using the kernel trick, or it may be treated as a vector with dimensionality reduction by principal component analysis.

[0032] The Anomaly Detection Model (AJM) is a model generated by machine learning using radio wave features within a predetermined range. The process by which the Anomaly Detection Model (AJM) determines whether or not the input radio wave features include features outside the predetermined range will be explained with reference to Figure 5. Figure 5 is a diagram showing the mapping of multiple radio wave feature vectors into a vector space.

[0033] The anomaly detection model AJM is trained to perform a process that obtains the boundary line PR between the features of normal waves and the features of abnormal waves in the vector space shown in Figure 5. With this configuration, the anomaly detection model AJM can determine that the feature IW shown in Figure 5 is a feature of an abnormal wave outside the boundary line PR (outside a predetermined range).

[0034] Examples of methods for training an anomaly detection model (AJM) include One-class SVM (One-class Support Vector Machine), Hotelling's theory, LoF method, and k-nearest neighbors. Alternatively, machine learning of models such as deep anomaly detection models using deep learning may also be used.

[0035] Another example of a method for training the Anomaly Detection Model (AJM) is to learn a threshold for statistically detecting anomalies from the distribution of high-dimensional (multi-dimensional) feature vectors, such as the Mahalanobis distance or variance (standard deviation), as described above.

[0036] Furthermore, the similarity between features is calculated according to the training method used to train the anomaly detection model AJM. For example, when the anomaly detection model AJM is trained using a one-class SVM, the similarity between features is calculated based on the coordinates of the features in the high-dimensional space mapped by the kernel trick, their distance from each other, and their distance from the classification surface. Alternatively, when the anomaly detection model AJM is trained using the k-nearest neighbors method, the similarity between features is calculated based on the Euclidean distance from the centroid of each feature group.

[0037] Another example of a method for training the Anomaly Detection Model (AJM) is to use state filters such as Kalman filters or particle filters to predict trends in change, and then train the model to detect deviations from those predicted trends.

[0038] If it is determined that an abnormal wave is present, the information processing device 2 extracts the interference wave using a normal wave representative model NWM that is similar to a radio wave that does not contain an abnormal wave.

[0039] The normal wave representative model (NWM) models a representative of one or more groups obtained by classifying multiple radio waves that do not contain abnormal waves according to their similarity. For example, in the diagram shown in Figure 5, if the features of normal waves inside the boundary PR are classified into groups for each similar sample (hereinafter also referred to as "normal wave sample NWS"), they can be classified into normal wave group NWG1, normal wave group NGW2, and normal wave group NWG3. In this case, the normal wave model NWM is generated by modeling the representative of each of the normal wave group NWG1, normal wave group NGW2, and normal wave group NWG3. As an example, the normal wave representative model NWM is a model of the arithmetic mean or median of each normal wave group. Also, when using the arithmetic mean in the normal wave representative model NWM, outliers may be excluded beforehand.

[0040] The information processing device 2 extracts interference waves and classifies the extracted interference waves into one or more groups according to their similarity. The information processing device 2 also identifies the type of interference wave using an interference wave representative model IWM similar to the extracted interference wave. The interference wave representative model IWM, like the normal wave representative model NWM described above, models a representative of each of the one or more interference wave groups classified according to similarity. The interference wave representative model IWM is a model of the arithmetic mean or median of each interference wave group. Alternatively, the interference wave representative model IWM can be generated using supervised learning methods in machine learning, where labels for each interference wave group are assigned to samples of that group for training, and then classified by machine learning.

[0041] (Configuration of Information Processing Device 2) The configuration of the information processing device 2 will be explained with reference to Figure 3. Figure 3 is a block diagram showing the configuration of the information processing device 2. As shown in Figure 3, the information processing device 2 includes a control unit 20, a storage unit 40, an input / output unit 41, a communication unit 42, and a receiver 43.

[0042] (Storage unit 40) The memory unit 40 stores data that the control unit 20 references. Examples of data stored in the memory unit 40 include one or more normal wave samples NWS, one or more normal wave representative models NWM, one or more observed waves OW, one or more interfering wave samples IWS, one or more interfering wave representative models IWM, labeled interfering waves LIW, and anomaly detection model AJM. When we say that the anomaly detection model AJM is stored in the memory unit 40, it means that the parameters that define the anomaly detection model AJM are stored in the memory unit 40.

[0043] (Input / output section 41) The input / output unit 41 is an interface to input devices that accept data input and output devices that output data. Examples of input devices include, but are not limited to, microphones, cameras, eye-tracking devices, keyboards, and touchpads. Examples of output devices include, but are not limited to, speakers and liquid crystal displays.

[0044] (Communications Section 42) The communication unit 42 is an interface for sending and receiving data over a network. Examples of the communication unit 42 include, but are not limited to, communication chips in various communication standards such as Ethernet®, Wi-Fi®, and wireless communication standards for mobile data communication networks, as well as USB-compliant connectors.

[0045] (Receiver 43) The receiver 43 is equipped with an antenna for receiving radio waves and uses the antenna to receive radio waves. The receiver 43 also measures time-series data of radio wave intensity in a set frequency band from the radio waves received by the antenna, converts it into formats such as IQ (In-Phase Quadrature-Phase) data and spectra, and supplies it to the control unit 20. The data that the receiver 43 supplies to the control unit 20 is also called the observed wave OW.

[0046] (Control unit 20) The control unit 20 controls each component of the information processing device 2. As shown in Figure 3, the control unit 20 also includes an acquisition unit 21, a feature quantity calculation unit 11, a normal wave classification unit 22, a normal wave representative model generation unit 23, an anomaly determination model learning unit 24, a determination unit 12, a normal wave identification unit 25, an interference wave extraction unit 13, an interference wave classification unit 14, an interference wave representative model generation unit 26, an interference wave type identification unit 27, and an output unit 28. In this exemplary embodiment, the feature quantity calculation unit 11, determination unit 12, interference wave extraction unit 13, interference wave classification unit 14, interference wave representative model generation unit 26, interference wave type identification unit 27, and output unit 28 each realize a feature quantity calculation means, a determination means, an interference wave extraction means, an interference wave classification means, an interference wave representative model generation means, an interference wave type identification means, and an output means, respectively.

[0047] (Acquisition part 21) The acquisition unit 21 acquires data. The acquisition unit 21 stores the acquired data in the storage unit 40. As an example, the acquisition unit 21 acquires the observed wave OW supplied from the receiver 43. As another example, the acquisition unit 21 acquires the normal wave sample NWS from the receiver 43 or other device via the communication unit 42.

[0048] (Feature calculation unit 11) The feature calculation unit 11 calculates the feature quantities of radio waves.

[0049] As an example, the feature calculation unit 11 calculates the feature quantities of the radio waves (observed wave OW) received by the receiver 43. In this case, the feature calculation unit 11 calculates feature quantities in the same form as the feature quantities learned by the anomaly judgment model AJM (for example, any of the forms shown in Figure 4). The feature calculation unit 11 supplies the calculated feature quantities of the observed wave OW to the judgment unit 12.

[0050] As another example, the feature calculation unit 11 calculates the features of a normal wave sample NWS. The feature calculation unit 11 supplies the calculated features of the normal wave sample NWS to the normal wave classification unit 22 and the anomaly detection model learning unit 24.

[0051] (Normal wave classification section 22) The normal wave classification unit 22 classifies multiple normal wave sample NWS into one or more groups. For example, the normal wave classification unit 22 refers to the feature quantities of the normal wave sample NWS supplied from the feature quantity calculation unit 11 and classifies multiple normal wave sample NWS into one or more groups.

[0052] If multiple radio sources exist within the range that the receiver 43 can receive, the normal wave sample NWS will include radio waves from each of the multiple radio sources. Among the multiple radio waves, the characteristics of radio waves with high received intensity, such as radio waves from radio sources located close to the receiver or radio waves from high-power radio sources, will be clearly expressed in the normal wave sample NWS. Therefore, the normal wave classification unit 22 can classify the normal wave sample NWS into a finite number of groups.

[0053] Here, since the normal wave samples NWS are vectors, the normal wave classification unit 22 can group them into sets for each adjacent sample using numerical analysis methods. In this process, clustering methods in machine learning can be used, such as the group average method, Ward's method, shortest distance method, and longest distance method.

[0054] For example, the normal wave classification unit 22 sets an appropriate threshold and classifies normal wave sample NWS located at a distance below that threshold into the same group. For example, in Figure 5, the normal wave classification unit 22 classifies multiple normal wave sample NWS into three groups: normal wave group NWG1, normal wave group NGW2, and normal wave group NWG3.

[0055] (Normal wave representative model generation unit 23) The normal wave representative model generation unit 23 generates a normal wave representative model NWM. For example, the normal wave representative model generation unit 23 generates a normal wave representative model NWM from each of the normal wave samples NWS of one or more groups classified by the normal wave classification unit 22. For example, in Figure 5, the normal wave representative model generation unit 23 generates normal wave representative model NWM1 for normal wave group NWG1, normal wave representative model NWM2 for normal wave group NGW2, and normal wave representative model NWM3 for normal wave group NWG3.

[0056] As described above, the normal wave representative model generation unit 23 generates a normal wave representative model NWM using the arithmetic mean or median of each normal wave group. Furthermore, when using the arithmetic mean for the normal wave representative model NWM, the normal wave representative model generation unit 23 may exclude outliers beforehand.

[0057] (Anomaly detection model learning unit 24) The anomaly detection model learning unit 24 trains the anomaly detection model AJM. The method by which the anomaly detection model learning unit 24 trains the anomaly detection model AJM is as described above.

[0058] The anomaly detection model learning unit 24 may train one or more anomaly detection models AJM. For example, the anomaly detection model learning unit 24 may train anomaly detection model AJM1 using normal wave samples NWS included in normal wave group NWG1 as shown in Figure 5, train anomaly detection model AJM2 using normal wave samples NWS included in normal wave group NWG2, and train anomaly detection model AJM3 using normal wave samples NWS included in normal wave group NWG3.

[0059] (Judgment section 12) The determination unit 12 inputs the feature quantities of the observed wave OW calculated by the feature quantity calculation unit 11 into the anomaly determination model AJM to determine whether or not an anomaly wave is present in the observed wave OW. The determination unit 12 supplies the determination result, which determines whether or not an anomaly wave is present in the observed wave OW, to the interference wave extraction unit 13. For example, if the feature quantities of the observed wave OW include the feature quantity IW shown in Figure 5, the determination unit 12 supplies the interference wave extraction unit 13 with a determination result indicating that an anomaly wave is present in the observed wave OW.

[0060] (Normal wave identification section 25) The normal wave identification unit 25 identifies the normal wave representative model NWM that is closest to the observed wave OW. The normal wave identification unit 25 supplies the identified normal wave representative model NWM to the interference wave extraction unit 13.

[0061] For example, if the determination result determined by the determination unit 12 indicates that the observed wave OW contains an abnormal wave, the normal wave identification unit 25 identifies the normal wave representative model NWM that is closest to the observed wave OW.

[0062] One example of how the normal wave identification unit 25 identifies the normal wave representative model NWM that is closest to the observed wave OW is to calculate the arithmetic distance between the observed wave OW and each of the multiple normal wave representative model NWMs, and identify the normal wave representative model NWM with the smallest arithmetic distance as the normal wave representative model NWM that is closest to the observed wave OW.

[0063] (Interference wave extraction unit 13) The interference wave extraction unit 13 extracts interference waves from the observed wave OW. The interference wave extraction unit 13 stores the extracted interference waves as interference wave samples IWS in the storage unit 40.

[0064] For example, if the determination result determined by the determination unit 12 indicates that the observed wave OW contains an abnormal wave, the interference wave is extracted by subtracting the feature quantities of the normal wave representative model NWM supplied by the normal wave identification unit 25 from the feature quantities of the observed wave OW.

[0065] The method by which the interference wave extraction unit 13 extracts interference waves from the observed wave OW is not limited, and instead of the method described above, interference waves may be extracted from the observed wave OW by a known method.

[0066] (Interference wave classification unit 14) The interfering wave classification unit 14 classifies the interfering wave samples IWS into one or more groups according to their similarity. The method used by the interfering wave classification unit 14 to classify the interfering wave samples IWS into one or more groups according to their similarity may be the same method used by the normal wave classification unit 22 described above to classify multiple normal wave samples NWS into one or more groups.

[0067] Interference waves can be emitted not just once, but multiple times, and in some cases, for extended periods or repeatedly. Therefore, when radio wave measurements are performed for a long period, the same interference wave can be observed multiple times. In other words, the interference wave classification unit 14 can classify the interference wave samples (IWS) into one or more groups. This is particularly noticeable when the location of the interference wave source does not change. On the other hand, when the location of the interference wave source changes (for example, when the source is a moving car), the interference waves become sporadic and can be ignored as outliers during clustering.

[0068] (Interference wave representative model generation unit 26) The interference wave representative model generation unit 26 generates one or more interference wave representative models IWMs that represent one or more groups of interference waves classified according to their similarity. For example, the interference wave representative model generation unit 26 generates interference wave representative models IWMs for groups classified by the interference wave classification unit 14. The interference wave representative model generation unit 26 stores the generated interference wave representative models IWMs in the storage unit 40.

[0069] As described above, the interference wave representative model generation unit 26 generates the interference wave representative model IWM in the same manner as the normal wave representative model generation unit 23 generates the normal wave representative model NWM.

[0070] Furthermore, the representative interference wave model generation unit 26 determines whether or not to generate a representative interference wave model IWM. For example, the representative interference wave model IWM is generated by the representative interference wave model generation unit 26 if a specified amount or more of interference wave samples IWS are stored in the storage unit 40, or if the measurement period has elapsed for a specified period or longer.

[0071] (Interference wave type identification unit 27) The interference wave type identification unit 27 identifies the type of one or more groups into which the interference waves extracted by the interference wave extraction unit 13 have been classified, based on the interference wave classification unit 14. More specifically, the interference wave type identification unit 27 identifies the type of interference wave by identifying a similar interference wave representative model IWM from among the interference wave representative models IWM for the interference waves extracted by the interference wave extraction unit 13. The interference wave type identification unit 27 assigns the identified interference wave representative model IWM as a label to the interference wave and stores it in the storage unit 40 as a labeled interference wave LIW. Alternatively, the interference wave type identification unit 27 may store the labeled interference wave LIW in the storage unit 40 in association with the observed wave OW, which is the source of the interference wave extraction.

[0072] As a method for the interference wave type identification unit 27 to identify a representative interference wave model IWM similar to the extracted interference wave, a method similar to the method used by the normal wave identification unit 25 to identify a representative normal wave model NWM that is closest to the observed wave OW can be used.

[0073] Examples of types of interference waves include interference waves that appear during specific time periods (nighttime, lunch break, holidays, etc.), interference waves of a specific intensity, and interference waves of a specific bandwidth.

[0074] (Output section 28) The output unit 28 outputs data via the input / output unit 41 or the communication unit 42. For example, the output unit 28 outputs information indicating the type of interference wave identified by the interference wave type identification unit 27.

[0075] (Process flow for training the anomaly detection model AJM) The process for training the anomaly detection model AJM (training method S3) will be explained with reference to Figure 6. Figure 6 is a flowchart showing the flow of training method S3. Training method S3 is executed before the generation method S4 and information processing method S2, which will be described later. Training method S3 may also be executed periodically.

[0076] (Step S31) In step S31, the acquisition unit 21 acquires a normal wave sample NWS. The acquisition unit 21 stores the acquired normal wave sample NWS in the storage unit 40.

[0077] (Step S32) In step S32, the feature calculation unit 11 calculates the feature quantities of the normal wave sample NWS stored in the memory unit 40. The feature calculation unit 11 supplies the calculated feature quantities of the normal wave sample NWS to the normal wave classification unit 22 and the anomaly detection model learning unit 24.

[0078] (Step S33) In step S33, the normal wave classification unit 22 refers to the feature quantities of the normal wave sample NWS supplied from the feature quantity calculation unit 11 and classifies the multiple normal wave sample NWS into one or more groups.

[0079] (Step S34) In step S34, the normal wave representative model generation unit 23 generates a normal wave representative model NWM from each of the normal wave sample NWS of one or more groups classified by the normal wave classification unit 22. The normal wave representative model generation unit 23 stores the generated normal wave representative model NWM in the storage unit 40.

[0080] (Step S35) In step S35, the anomaly detection model learning unit 24 trains the anomaly detection model AJM using the features of the normal wave sample NWS supplied from the feature calculation unit 11.

[0081] The processing order of steps S33 and S34 and step S35 is not limited. For example, the information processing device 2 may execute step S35 first, then steps S33 and S34, or it may execute steps S33 and S34 and step S35 in parallel.

[0082] (Process flow for generating the representative interference wave model IWM) The process for generating the interfering wave representative model IWM (generation method S4) will be explained with reference to Figure 7. Figure 7 is a flowchart showing the flow of generation method S4. Generation method S4 is executed before the information processing method S2, which will be described later. Generation method S4 may also be executed periodically.

[0083] (Step S21) In step S21, the information processing device 2 performs radio wave measurement. More specifically, the information processing device 2 performs the following steps S211 and S212.

[0084] (Step S211) In step S211, the acquisition unit 21 acquires the observed wave OW. The acquisition unit 21 stores the acquired observed wave OW in the storage unit 40.

[0085] (Step S212) In step S212, the feature calculation unit 11 calculates the feature quantities of the observed wave OW stored in the memory unit 40. The feature calculation unit 11 supplies the calculated feature quantities to the determination unit 12, the normal wave identification unit 25, and the interference wave extraction unit 13.

[0086] (Step S22) In step S22, the determination unit 12 inputs the features of the observed wave OW calculated by the feature calculation unit 11 into the anomaly determination model AJM to determine whether or not an anomaly wave is included in the observed wave OW.

[0087] Step S22 may be executed when the acquisition unit 21 acquires a predetermined amount of the observed wave OW in step S211.

[0088] If it is determined in step S22 that the observed wave OW does not contain any abnormal waves (step S22: NO), the information processing device 2 executes the process in step S21 again.

[0089] (Step S23) If, in step S22, it is determined that the observed wave OW contains an abnormal wave (step S22: YES), then in step S23, the information processing device 2 extracts the interference wave. More specifically, the information processing device 2 executes the following steps S231 and S232.

[0090] (Step S231) In step S231, the normal wave identification unit 25 identifies the normal wave representative model NWM that is closest to the observed wave OW. The normal wave identification unit 25 supplies the identified normal wave representative model NWM to the interference wave extraction unit 13.

[0091] (Step S232) In step S232, the interference wave extraction unit 13 extracts the interference wave by subtracting the feature quantities of the normal wave representative model NWM supplied from the normal wave identification unit 25 from the feature quantities of the observed wave OW. The interference wave extraction unit 13 stores the extracted interference wave as an interference wave sample IWS in the storage unit 40.

[0092] (Step S41) In step S41, the interference wave representative model generation unit 26 determines whether or not to generate the interference wave representative model IWM.

[0093] If it is determined in step S41 that no representative interference wave model (IWM) is generated (step S41: NO), the information processing device 2 executes the process in step S21 again.

[0094] (Step S25) If it is determined in step S41 to generate an interfering wave representative model IWM (step S41: NO), then in step S25, the information processing device 2 generates an interfering wave representative model IWM. More specifically, the information processing device 2 executes the following steps S251 and S252.

[0095] (Step S251) In step S251, the interference wave classification unit 14 classifies the interference wave samples IWS into one or more groups according to their similarity.

[0096] (Step S252) In step S252, the interference wave representative model generation unit 26 generates an interference wave representative model IWM for the group in which the interference wave samples IWS were classified in step S251.

[0097] (Processing flow executed by information processing device 2) The flow of the processing (information processing method S2) executed by the information processing device 2 will be explained with reference to Figure 8. Figure 8 is a flowchart showing the flow of the information processing method S2.

[0098] (Steps S21 to S23) The processing in steps S21 to S23 is as described above, so we will omit the explanation.

[0099] (Step S24) In step S24, the interference wave representative model generation unit 26 determines whether or not to regenerate the interference wave representative model IWM.

[0100] (Step S25) If it is determined in step S24 to regenerate the representative interference wave model IWM (step 24: YES), then in step S25, the information processing device 2 generates the representative interference wave model IWM. The process in step S25 is as described above, so its explanation is omitted.

[0101] (Step S26) If it is determined in step S24 that the representative interference wave model IWM is not to be regenerated (step 24: YES), or if step S25 is executed, in step S26, the interference wave type identification unit 27 identifies the type of interference wave for the interference wave sample IWS by identifying a similar representative interference wave model IWM from among the representative interference wave model IWMs. The interference wave type identification unit 27 assigns the identified representative interference wave model IWM as a label to the interference wave and stores it in the storage unit 40 as a labeled interference wave LIW.

[0102] (Step S27) In step S27, the output unit 28 outputs information indicating the type identified by the interference wave type identification unit 27 in step S26.

[0103] (Effects of Information Processing Device 2) As described above, if the observed wave OW contains abnormal waves, the information processing device 2 extracts interference waves from the observed wave OW and classifies the extracted interference waves into one or more groups according to their similarity. Therefore, the information processing device 2 can identify which group of interference waves is contained in the observed wave OW, and thus can identify the interference wave source with high accuracy.

[0104] Furthermore, even when the observed wave OW contains multiple interfering waves, the information processing device 2 can classify each interfering wave into one or more groups. Therefore, the information processing device 2 can accurately identify the source of each of the multiple interfering waves.

[0105] Furthermore, the information processing device 2 identifies the type of interference wave. Therefore, since the information processing device 2 can identify the type of interference wave, it can pinpoint the interference wave source with high accuracy.

[0106] Furthermore, the information processing device 2 outputs information indicating the identified type. Therefore, the information processing device 2 can notify the user of the type of interference wave, enabling the user to efficiently deal with interference wave radio wave monitoring.

[0107] Furthermore, the information processing device 2 generates an interfering wave representative model IWM that represents a group of interfering waves, and identifies the type of interfering wave by identifying an interfering wave representative model similar to the interfering wave. Therefore, the information processing device 2 can suitably identify the type of interfering wave.

[0108] Furthermore, the information processing device 2 uses the features of the normal wave sample NWS to train the anomaly detection model AJM. Therefore, the information processing device 2 can suitably train (generate) the anomaly detection model AJM.

[0109] [Examples of implementation using software] Some or all of the functions of the information processing devices 1 and 2 (hereinafter also referred to as "the above devices") may be implemented by hardware such as integrated circuits (IC chips) or by software.

[0110] In the latter case, each of the above devices is implemented, for example, by a computer that executes instructions for a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as Computer C) is shown in Figure 9. Figure 9 is a block diagram showing the hardware configuration of Computer C, which functions as each of the above devices.

[0111] Computer C comprises at least one processor C1 and at least one memory C2. Memory C2 stores a program P that causes computer C to operate as each of the above-mentioned devices. In computer C, processor C1 reads program P from memory C2 and executes it, thereby realizing each of the above-mentioned devices.

[0112] For processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating Point Number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. For memory C2, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0113] Computer C may also be equipped with RAM (Random Access Memory) for loading program P at runtime and for temporarily storing various data. Furthermore, computer C may be equipped with communication interfaces for sending and receiving data with other devices. Additionally, computer C may be equipped with input / output interfaces for connecting input / output devices such as keyboards, mice, displays, and printers.

[0114] Furthermore, program P can be recorded on a non-temporary, tangible recording medium M that is readable by computer C. Such a recording medium M could be, for example, tape, disk, card, semiconductor memory, or programmable logic circuitry. Computer C can acquire program P via such a recording medium M. Program P can also be transmitted via a transmission medium. Such a transmission medium could be, for example, a communication network or broadcast waves. Computer C can also acquire program P via such a transmission medium.

[0115] Furthermore, each of the above functions of each of the above devices may be implemented by a single processor in a single computer, by multiple processors in a single computer working together, or by multiple processors in each of multiple computers working together. In addition, the programs for implementing each of the above functions in each of the above devices may be stored in a single memory in a single computer, distributed and stored in multiple memories in a single computer, or distributed and stored in multiple memories in each of multiple computers.

[0116] [Additional Note A] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0117] (Note A1) A feature calculation means for calculating the feature quantities of radio waves received by a receiver, A determination means determines whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated by the feature quantity calculation means into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination means determines that the abnormal wave is included, the interference wave extraction means extracts the interference wave which is an electromagnetic wave outside the predetermined range, Interference wave classification means classifies the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity, An information processing device equipped with the following features.

[0118] (Appendix A2) The system further comprises interference wave type identification means for identifying the type of group to which the interference waves extracted by the interference wave extraction means have been classified by the interference wave classification means. The information processing device described in Appendix A1.

[0119] (Note A3) The system further includes an output means for outputting information indicating the type identified by the interference wave type identification means. The information processing device described in Appendix A2.

[0120] (Note A4) The system further comprises interference wave representative model generation means for generating one or more interference wave representative models that represent each of the one or more groups, The interference wave type identification means identifies the type of interference wave by identifying similar representative interference wave models from among the representative interference wave models generated by the representative interference wave model generation means. The information processing device described in Appendix A2 or A3.

[0121] (Note A5) The system further comprises an anomaly detection model learning means for using the characteristic quantities of radio waves within the predetermined range to perform machine learning on the anomaly detection model. An information processing device as described in any one of the appendices A1 to A4.

[0122] [Additional Notes B] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0123] (Note B1) At least one processor performs a feature calculation process to calculate the feature quantities of the radio waves received by the receiver, The at least one processor takes the radio wave features as input and performs a determination process to determine whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the features calculated in the feature calculation process into an anomaly determination model that determines whether the radio wave features include features outside a predetermined range. If the at least one processor determines in the determination process that the abnormal wave is included, it performs an interference wave extraction process to extract the interference wave which is an electromagnetic wave outside the predetermined range, The at least one processor performs an interference wave classification process that classifies the interference waves extracted in the interference wave extraction process into one or more groups according to their similarity, Information processing methods including

[0124] (Note B2) The at least one processor further includes an interference wave type identification process that identifies the type of group into which the interference waves extracted in the interference wave extraction process were classified in the interference wave classification process. The information processing method described in Appendix B1.

[0125] (Note B3) The at least one processor further includes output processing that outputs information indicating the type identified in the interference wave type identification processing. The information processing method described in Appendix B2.

[0126] (Note B4) The at least one processor further includes interference wave representative model generation processing to generate one or more interference wave representative models that represent each of the one or more groups, In the interference wave type identification process, the at least one processor identifies the type of interference wave by identifying similar representative interference wave models from among the representative interference wave models generated by the representative interference wave model generation process. The information processing method described in Appendix B2 or B3.

[0127] (Note B5) The at least one processor further includes an anomaly detection model learning process that uses the radio wave features within the predetermined range to machine-learn the anomaly detection model. The information processing method described in any one of the appendices B1 to B4.

[0128] [Additional Note C] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0129] (Note C1) A program that makes a computer function as an information processing device. The aforementioned computer, A feature calculation means for calculating the feature quantities of radio waves received by a receiver, A determination means determines whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated by the feature quantity calculation means into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination means determines that the abnormal wave is included, the interference wave extraction means extracts the interference wave which is an electromagnetic wave outside the predetermined range, Interference wave classification means classifies the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity, An information processing program that functions as such.

[0130] (Note C2) The aforementioned computer, The interference wave classification means further functions as an interference wave type identification means for identifying the type of group to which the interference waves extracted by the interference wave extraction means have been classified. The information processing program described in Appendix C1.

[0131] (Note C3) The aforementioned computer, Further configured as an output means for outputting information indicating the type identified by the interference wave type identification means, The information processing program described in Appendix C2.

[0132] (Note C4) The aforementioned computer, Further configured as an interference wave representative model generation means for generating one or more interference wave representative models that represent each of the one or more groups, The interference wave type identification means identifies the type of interference wave by identifying similar representative interference wave models from among the representative interference wave models generated by the representative interference wave model generation means. The information processing program described in Appendix C2 or C3.

[0133] (Note C5) The aforementioned computer, The characteristic quantities of radio waves within the predetermined range are used to further enable the anomaly detection model to function as a machine learning means for the anomaly detection model. An information processing program described in any one of the appendices C1 to C4.

[0134] [Additional Note D] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0135] (Note D1) It comprises at least one processor, and the at least one processor is A feature calculation process that calculates the feature quantities of the radio waves received by the receiver, A determination process is performed to determine whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated in the feature quantity calculation process into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination process determines that the abnormal wave is included, an interference wave extraction process is performed to extract the interference wave which is an electromagnetic wave outside the predetermined range. Interference wave classification process, which classifies the interference waves extracted in the aforementioned interference wave extraction process into one or more groups according to their similarity, An information processing device that performs the following actions.

[0136] The information processing device may also include memory. Furthermore, the memory may store a program that causes at least one processor to execute each of the aforementioned processes.

[0137] (Note D2) The aforementioned at least one processor, Further, an interference wave type identification process is performed to identify the type of group that the interference waves extracted in the interference wave extraction process were classified into in the interference wave classification process. The information processing device described in Appendix D1.

[0138] (Note D3) The aforementioned at least one processor, Further output processing is performed to output information indicating the type identified in the interference wave type identification process. The information processing device described in Appendix D2.

[0139] (Note D4) The aforementioned at least one processor, Further interference wave representative model generation processing is performed to generate one or more representative interference wave models that represent each of the one or more groups, In the interference wave type identification process, the at least one processor identifies the type of interference wave by identifying similar interference wave representative models from among the interference wave representative models generated in the interference wave representative model generation process. The information processing device described in Appendix D2 or D3.

[0140] (Note D5) The aforementioned at least one processor, Further an anomaly detection model learning process is performed to train the anomaly detection model using the characteristic quantities of radio waves within the predetermined range. An information processing device as described in any one of the appendices D1 to D4.

[0141] [Additional Note E] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0142] (Note E1) A program that makes a computer function as an information processing device. To the aforementioned computer, A feature calculation process that calculates the feature quantities of the radio waves received by the receiver, A determination process is performed to determine whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated in the feature quantity calculation process into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination process determines that the abnormal wave is included, an interference wave extraction process is performed to extract the interference wave which is an electromagnetic wave outside the predetermined range. Interference wave classification process, which classifies the interference waves extracted in the aforementioned interference wave extraction process into one or more groups according to their similarity, A non-temporary recording medium that stores an information processing program that executes that program. [Explanation of Symbols]

[0143] 1, 2 Information Processing Devices 11 Feature Calculation Unit 12 Judgment section 13 Interference wave extraction section 14 Interference wave classification unit 21 Acquisition Department 22 Normal wave classification section 23. Normal wave representative model generation unit 24 Anomaly detection model learning unit 25 Normal wave identification section 26 Interference Wave Representative Model Generation Unit 27 Interference wave type identification unit 28 Output section AJM Anomaly Detection Model IWM Interference Wave Representative Model IWS Interferential Wave Sample LIW labeled interferometric waves NWM (Non-Wireless Normal Wave Representative Model) NWS normal wave sample OW observation waves

Claims

1. A feature calculation means for calculating the feature quantities of radio waves received by a receiver, A determination means determines whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated by the feature quantity calculation means into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination means determines that the abnormal wave is included, the interference wave extraction means extracts the interference wave which is an electromagnetic wave outside the predetermined range, An interference wave classification means classifies the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity, An information processing device equipped with the following features.

2. The system further comprises interference wave type identification means for identifying the type of group to which the interference waves extracted by the interference wave extraction means have been classified by the interference wave classification means. The information processing apparatus according to claim 1.

3. The system further includes an output means for outputting information indicating the type identified by the interference wave type identification means. The information processing apparatus according to claim 2.

4. The system further comprises interference wave representative model generation means for generating one or more representative interference wave models that represent each of the one or more groups, The interference wave type identification means identifies the type of interference wave by identifying similar representative interference wave models from among the representative interference wave models generated by the representative interference wave model generation means. The information processing apparatus according to claim 2 or 3.

5. The system further comprises an anomaly detection model learning means for using the characteristic quantities of radio waves within the predetermined range to perform machine learning on the anomaly detection model. The information processing apparatus according to claim 1 or 2.

6. At least one processor performs a feature calculation process to calculate the feature quantities of the radio waves received by the receiver, The at least one processor takes the radio wave feature quantities as input and performs a determination process to determine whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated in the feature quantity calculation process into an anomaly determination model that determines whether the radio wave feature quantities include radio wave feature quantities outside a predetermined range. If the at least one processor determines in the determination process that the abnormal wave is included, it performs an interference wave extraction process to extract the interference wave which is an electromagnetic wave outside the predetermined range, The at least one processor performs an interference wave classification process that classifies the interference waves extracted in the interference wave extraction process into one or more groups according to their similarity, Information processing methods including

7. A program that makes a computer function as an information processing device. The aforementioned computer, A feature calculation means for calculating the feature quantities of radio waves received by a receiver, A determination means determines whether the radio waves received by the receiver contain abnormal waves that are outside the predetermined range, by inputting the feature quantities calculated by the feature quantity calculation means into an anomaly determination model that takes the feature quantities of the radio waves as input and determines whether the feature quantities of the radio waves include feature quantities of radio waves outside a predetermined range, If the determination means determines that the abnormal wave is included, the interference wave extraction means extracts the interference wave which is an electromagnetic wave outside the predetermined range, An interference wave classification means classifies the interference waves extracted by the interference wave extraction means into one or more groups according to their similarity, An information processing program that functions as such.

Images