Radio evaluation system, radio evaluation method, radio device, and radio device control method

Abstract

To implement objective evaluation for a radio device such as on-vehicle one.SOLUTION: A radio evaluation system includes: an acquisition unit for acquiring a radio signal received by a radio device; an identification unit for identifying classification information on noise of the radio signal acquired by the acquisition unit using a learned model taking the radio signal as input and providing the classification information on the noise of the radio signal as output; and an output unit for outputting the classification information on the noise identified by the identification unit.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a radio evaluation system, a radio evaluation method, a radio device, and a method for controlling a radio device. [Background technology]

[0002] Conventionally, for in-vehicle radio devices, there is a need to evaluate the quality not only at a specific location but also while considering the movement of the vehicle. For example, since the signal reception environment changes as the vehicle equipped with the radio device moves, there is a need for a quality evaluation method that takes such environmental changes into account. Furthermore, the quality evaluation of radio devices sometimes involves sensory evaluation by users, and there is a need for a more objective evaluation method that does not rely on individuals.

[0003] For example, Patent Document 1 discloses an evaluation method for determining the deterioration of the reception condition of an in-vehicle radio device by simultaneously recording audio data under good noise-free reception conditions and audio data under conditions where the reception condition has deteriorated while driving, and determining the degree of deterioration of the reception condition from the difference. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2010-10848 Summary of the Invention [Problem to be solved by the invention]

[0005] In the method disclosed in Patent Document 1, audio data in good condition and audio data in bad condition must be received and used simultaneously in a coordinated manner at separate locations, and such a configuration also makes it difficult to achieve real-time processing.

[0006] Furthermore, noise in radio signals may be reduced after evaluating the quality of the radio signals. One type of noise is multipath noise, and a known technique is to detect the level of multipath noise and apply high-cut processing. On the other hand, one type of content transmitted via radio signals is music with a high amount of high-frequency components. Since high-frequency noise is less noticeable in such content, uniform application of high-cut processing when noise is detected may result in a decrease in user quality when the content is played. For example, applying high-cut processing to music with a high amount of high-frequency components is likely to result in a muffled output. Therefore, it is necessary to more appropriately evaluate noise based on the content of the radio signal and then reduce the noise.

[0007] In view of the above-mentioned problems, the present disclosure aims to realize an objective evaluation of radio devices such as those mounted in vehicles. [Means for solving the problem]

[0008] The present disclosure provides a radio evaluation system having an acquisition unit that acquires a radio signal received by a radio device, an identification unit that identifies noise classification information of the radio signal acquired by the acquisition unit using a trained model that takes the radio signal as input and outputs noise classification information of the radio signal, and an output unit that outputs the noise classification information identified by the identification unit.

[0009] The present disclosure also provides a radio device having a receiving unit that receives a radio signal, an identification unit that uses a trained model that receives a radio signal as an input and outputs noise classification information of the radio signal to identify noise classification information of the radio signal received by the receiving unit, an identification unit that uses the noise classification information identified by the identification unit to identify noise processing for the radio signal, and an output control unit that applies the noise processing identified by the identification unit to the radio signal and outputs sound.

[0010] The present disclosure also provides a method for evaluating a radio device, comprising: an acquisition step of acquiring a radio signal received by the radio device; an identification step of identifying noise classification information of the radio signal acquired in the acquisition step using a trained model that takes the radio signal as input and outputs noise classification information of the radio signal; and an output step of outputting the noise classification information identified in the identification step.

[0011] The present disclosure also provides a method for controlling a radio device, comprising: a receiving step of receiving a radio signal; an identification step of identifying noise classification information of the radio signal received in the receiving step using a trained model that takes the radio signal as input and outputs noise classification information of the radio signal; an identification step of identifying noise processing for the radio signal using the noise classification information identified in the identification step; and an output control step of applying the noise processing identified in the identification step to the radio signal and outputting audio.

[0012] Any combination of the above components, and conversion of the expression of the present disclosure into a method, device, system, storage medium, computer program, etc., are also valid aspects of the present disclosure. [Effects of the Invention]

[0013] According to the present disclosure, it is possible to realize objective evaluation of radio devices such as those mounted in vehicles. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a block diagram illustrating a learning model that can be used in a radio evaluation system according to a first embodiment. [Figure 2] FIG. 1 is a block diagram showing an example of the device configuration of a radio evaluation system according to a first embodiment. [Figure 3] Flowchart of radio evaluation process according to the first embodiment [Figure 4A] FIG. 10 is a diagram showing an example of the configuration of an evaluation screen for evaluation results according to the first embodiment; [Figure 4B]FIG. 10 is a diagram showing an example of the configuration of an evaluation screen for evaluation results according to the first embodiment; [Figure 4C] FIG. 10 is a diagram showing an example of the configuration of an evaluation screen for evaluation results according to the first embodiment; [Figure 5] Flowchart of adjustment processing according to the second embodiment [Figure 6] FIG. 10 is a graph illustrating the concept of signal adjustment according to the second embodiment. [Figure 7A] FIG. 10 is a graph illustrating the concept of applying high-cut processing according to the third embodiment. [Figure 7B] FIG. 10 is a graph illustrating the concept of applying high-cut processing according to the third embodiment. [Figure 8] FIG. 10 is a block diagram showing a configuration example of a radio device according to a third embodiment. [Figure 9] FIG. 10 is a graph illustrating the concept of applying high-cut processing according to the third embodiment. [Figure 10] FIG. 11 is a diagram showing an example of the configuration of a high-cut control table according to the third embodiment. [Figure 11] Flowchart of audio output processing according to the third embodiment [Figure 12] FIG. 10 is a block diagram showing a modified example of the system configuration according to the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, with appropriate reference to the accompanying drawings, embodiments specifically disclosing a radio evaluation system, a radio evaluation method, a radio device, and a radio device control method according to the present disclosure will be described in detail. However, more detailed description than necessary may be omitted. For example, detailed description of already well-known matters or redundant description of substantially identical configurations may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Note that the accompanying drawings and the following description are provided to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter recited in the claims.

[0016] <First Embodiment> [Learning model] First, a learning model for noise discrimination used in each embodiment of the present disclosure will be described. In the following description, "learning" or "machine learning" refers to generating a "trained model" by performing learning using training data and an arbitrary learning algorithm. The trained model is updated appropriately as learning progresses using multiple training data, and its output changes even for the same input. Therefore, the trained model is not limited to a state at any point in time. Here, a model used in learning will be referred to as a "learning model," and a learning model that has undergone a certain level of learning will be referred to as a "trained model."

[0017] Specific examples of "training data" will be described later, but the configuration thereof may be adjusted or changed depending on the learning algorithm used and its application. The training data is composed of pairs (hereinafter also referred to as "datasets") of data such as audio data, text data, image data, video data, numerical data, and point cloud data, and annotation data (hereinafter also referred to as "labels") that has been manually or mechanically assigned to the data. Alternatively, the training data is composed solely of data such as audio data, image data, text data, video data, numerical data, and point cloud data that has not been assigned labels. In this embodiment, data based on signals received by a radio device (hereinafter referred to as "radio signals") is particularly used as the training data.

[0018] Furthermore, the training data may include training data used for training itself, verification data used for verifying a trained model, and test data used for testing a trained model. In the following description, the term "training data" is used to refer collectively to data related to training, and the term "training data" is used to refer to data used when performing training itself. Note that it is not intended to clearly classify the training data into training data, verification data, and test data; for example, depending on the training, verification, and testing methods, all training data may also be training data.

[0019] Furthermore, the structure of the training data and the preprocessing of the training data may vary depending on the type of learning algorithm and the input and output contents of the learning model. For example, depending on the learning algorithm, labeled training data and unlabeled training data may be used in combination (semi-supervised machine learning). Therefore, the combinations of each data item described below are examples and are not limited to these.

[0020] 1 is a block diagram illustrating a learning model for noise discrimination according to this embodiment. Here, model 100, which represents a "learning phase" in which a trained model is generated using training data, and model 110, which represents a "use phase" in which noise discrimination is performed using the generated trained model, are conceptually shown.

[0021] Because the learning process typically involves a high processing load, it is preferable that the learning phase and the utilization phase be performed separately by different devices. In this case, the radio device that performs the utilization phase, i.e., the noise identification process, is different from the device that generates the trained model used in that process. The radio device can be configured to acquire and utilize the trained model as needed.

[0022] In the learning phase, learning data 101 is prepared in advance. The learning data 101 includes an audio signal 103 based on a radio signal received by a radio device and annotation data 102 corresponding to the audio signal 103. The annotation data 102 contains classification information of noise in the audio signal 103, which is added manually, for example.

[0023] In this embodiment, noise level is used as classification information for noise. The noise level is evaluated on a five-level scale from noise level 1 to noise level 5, with the smaller the value, the lower the noise level. When creating learning data, for example, if the noise in the content of the radio signal is noticeable, annotation data is added so that the noise level value becomes higher. On the other hand, if the noise in the content of the radio signal is not noticeable, annotation data is added so that the noise level value becomes lower. Note that the classification information for noise is not limited to noise level, and may be, for example, the type of noise as described below. The audio signal 103 may be audio data of a predetermined time width (for example, about 10 seconds).

[0024] In the learning process 104, a speech signal 103 of the learning data 101 is input to a learning model based on a predetermined learning algorithm, and one of the noise levels is output as a noise identification result 105. Furthermore, in parameter adjustment 106, various parameters of the learning model used in the learning process 104 are adjusted by comparing the noise identification result 105 with the annotation data 102 corresponding to the input speech signal 103. That is, in the parameter adjustment, the learning process is advanced by adjusting the parameters of the learning model so that the values of the annotation data 102 and the noise identification result 105 approach each other, thereby generating a trained model that provides a more appropriate output. In this way, by repeating the learning process using multiple data sets, a trained model that approaches the desired output is generated.

[0025] In the model 110, noise discrimination is performed using an already generated trained model. In the noise discrimination process 112, an audio signal 111 to be discriminated is input to the already generated trained model, and noise classification information of the audio signal 111 is output. Here, the noise level of each of the multiple audio signals 111 is discriminated.

[0026] In the discrimination result counting process 113, the noise levels identified for the multiple audio signals 111 in the noise discrimination process 112 are counted. As described above, in the case of a five-level evaluation, each of the five noise levels is counted.

[0027] [Device configuration] Fig. 2 is a block diagram showing an example of the configuration of a radio evaluation system according to this embodiment. In Fig. 2, arrows indicate an example of the flow of signals to each block. Note that the configuration shown in Fig. 2 is just one example. Therefore, the system may further include components other than those shown in Fig. 2.

[0028] The radio evaluation system 200 according to this embodiment receives radio signals corresponding to one or more services, evaluates the signals, and outputs a selected service from the one or more services as audio. Note that the function of the radio evaluation system 200 is not limited to radio signals acquired in real time, but may also evaluate radio signals that have been received and recorded in advance. Furthermore, the system may be configured to include multiple signal receiving units, each of which evaluates the radio signal received by the respective receiving units.

[0029] In the configuration example of FIG. 2, a radio device and a radio evaluation system are integrated into one unit; however, the radio evaluation system may be configured as a standalone system by omitting the audio output section. The radio device according to this embodiment is assumed to be, for example, an in-vehicle radio device mounted in a vehicle and operable by a user. The in-vehicle radio device may be configured as a single device that outputs radio signals, or may be configured as one function of a navigation device. The term "vehicle" used below is not limited to a passenger car, but may refer to any mobile object, such as a motorcycle, bus, or truck. The configuration shown in FIG. 2 is merely an example; a single component may be divided into multiple components, or multiple components may be integrated into one. Only the components related to the functions of this embodiment are shown here, and the device may be configured to provide other functions.

[0030] In the radio evaluation system 200, a signal receiving unit 201 receives a radio signal in a predetermined frequency band transmitted from a nearby base station by an antenna (not shown), and provides the signal to an audio demodulation unit 202. In this embodiment, the standard and type of the radio signal are not particularly limited, and any desired radio signal may be used.

[0031] The audio demodulation unit 202 receives (filters) radio signals of a frequency specified by a user's station selection, etc., from the radio signals received by the signal receiving unit 201, and outputs them as audio signals to the delay correction unit 203 and the noise identification unit 204.

[0032] The delay correction unit 203 controls the output timing etc. in response to a delay caused by processing by the noise identification unit 204 (described later) etc. The delay correction unit 203 outputs the corrected audio signal to the noise processing unit 205.

[0033] 1, the noise discrimination unit 204 applies noise discrimination processing 112 to the audio signal input from the audio demodulation unit 202 using a trained model. Then, the noise discrimination unit 204 outputs the discrimination result to the noise processing unit 205 and the noise counting unit 207.

[0034] The noise processing unit 205 performs noise reduction processing on the audio signal input from the delay correction unit 203 based on parameters corresponding to the noise level identified by the noise identification unit 204. The noise reduction method here is not particularly limited, and a noise reduction process defined in advance corresponding to the noise level may be applied. More specifically, the methods of the second and third embodiments described below may be used. The noise processing unit 205 then outputs the processed audio signal to the audio output unit 206.

[0035] The audio output unit 206 outputs audio based on the audio signal from the noise processing unit 205 using a speaker or the like (not shown).

[0036] 1, the noise counting unit 207 executes a process of counting the results of the noise levels identified by the noise identification process 112. The noise counting unit 207 outputs the counting results to the screen generation unit 208.

[0037] The screen generation unit 208 generates an evaluation screen based on the counting result from the noise counting unit 207. An example of the configuration of the UI screen generated here will be described later with reference to Figs. 4A to 4C.

[0038] A screen output unit 209 outputs the evaluation screen generated by the screen generation unit 208 using a display (not shown) or the like.

[0039] Each block shown in Fig. 2 may be realized by, for example, a control unit or a storage unit (not shown). The control unit (not shown) may be configured using, for example, a central processing unit (CPU), a micro processing unit (MPU), a digital signal processor (DSP), a graphical processing unit (GPU), or a field programmable gate array (FPGA). The storage unit (not shown) is a storage area for storing and holding various data, and may be configured, for example, by a read only memory (ROM) or a hard disk drive (HDD) that is a nonvolatile storage area, or a random access memory (RAM) that is a volatile storage area. For example, the control unit may realize some of the functions of the blocks shown in Fig. 2 by reading and executing various data and programs stored in the storage unit.

[0040] [Radio Evaluation Processing] Fig. 3 illustrates a flowchart of the radio evaluation process according to this embodiment. This process flow corresponds to the processes performed by the noise identification unit 204, noise counting unit 207, screen generation unit 208, and screen output unit 209, among the blocks of the radio evaluation system 200 shown in Fig. 2. For ease of explanation, the radio evaluation system 200 will be collectively described as the processing entity.

[0041] The radio evaluation system 200 acquires an audio signal to be evaluated (step S301).

[0042] The radio evaluation system 200 identifies the noise level of the audio signal acquired in step S301 using the trained model that has already been generated (step S302).

[0043] The radio evaluation system 200 performs a counting process of the noise levels identified in step S302 (step S303). As described above, if the noise level is evaluated on a five-point scale, the number of occurrences of each noise level may be counted. Note that the counting process here is not limited to counting the number of noise levels, and other counting processes or association with other information may be performed depending on the configuration of the evaluation screen described later.

[0044] The radio evaluation system 200 generates an evaluation screen based on the counting result in step S304 (step S304). At this time, the radio evaluation system 200 may also acquire various information required for the generated evaluation screen.

[0045] The radio evaluation system 200 displays the evaluation screen generated in step S304 on a display etc. Then, this processing flow ends.

[0046] [Example of evaluation screen configuration] 4A to 4C are diagrams showing examples of the configuration of an evaluation screen relating to the radio evaluation method according to the present embodiment.

[0047] 4A shows an example of the configuration of an evaluation screen 400 that shows the timing at which the noise level of an audio signal occurs. In FIG. 4A, the evaluation screen 400 includes a graph in which the horizontal axis indicates the noise occurrence time [s] and the vertical axis indicates the noise level. When generating the evaluation screen 400, time information related to the reception timing of the audio signal is also acquired and used along with the audio signal.

[0048] FIG. 4B shows an example of the configuration of an evaluation screen 410 that displays the noise level of an audio signal in association with the travel route of a vehicle equipped with an in-vehicle radio device. The noise level (NL) of the audio signal is plotted against the position where the audio signal was received along the travel route. When generating the evaluation screen 410, location information regarding the reception position of the audio signal is also acquired and used along with the audio signal. The location information may be position measurement information from a Global Navigation Satellite System (GNSS) such as a Global Positioning System (GPS), and may include latitude, longitude, altitude, etc. The location information such as GPS may be acquired, for example, from a GPS sensor (not shown) mounted on the vehicle.

[0049] FIG. 4C shows an example of the configuration of an evaluation screen 420 for comparing evaluation results determined based on radio signals received by multiple in-vehicle radio devices. This example shows a comparison of the evaluation results of two radio devices. Name 421 indicates the name of the radio device being evaluated, here "Radio A." Aggregate result 422 indicates the aggregation result of the noise level of the evaluation target indicated by name 421. Aggregate result 422 is shown in a graph format, with the horizontal axis indicating the noise level and the vertical axis indicating the number of noises, but this is not limited to this and may be shown only numerically. Name 423 indicates the name of the radio device being evaluated, here "Radio B." Aggregate result 424 indicates the aggregation result of the noise level of the evaluation target indicated by name 423. Aggregate result 424 is also not limited to a graph format and may be displayed in a different format, similar to aggregate result 422. Evaluation judgment 425 indicates the result of comparing each evaluation target based on the aggregation result and predetermined evaluation criteria. The evaluation criteria may be based on weighting of the number of noise levels or may be based on a threshold value for a predetermined noise level.

[0050] 4C , for example, the noise identification unit 204 acquires audio signals received by each of a plurality of radio devices and performs noise identification on each of them. For example, the information to be compared may be acquired by using a plurality of radio devices to be evaluated to acquire audio signals under the same conditions, and aggregating the results of performing noise identification on the audio signals.

[0051] The evaluation screens shown in Figures 4A to 4C are not limited to any one of them, and may be configured to be switchable by the user, for example. Furthermore, multiple tabulation results may be displayed side by side on one screen. Furthermore, the output format is not limited to tables and graphs as shown in Figures 4A to 4C, and may be output as text information or statistical information.

[0052] As described above, a radio evaluation system (e.g., 200) according to this embodiment includes an acquisition unit (e.g., 201, 202) that acquires a radio signal received by a radio device, an identification unit (e.g., 204) that identifies noise classification information of the acquired radio signal using a trained model that receives the radio signal as input and outputs noise classification information of the radio signal, and an output unit (e.g., 207, 208, 209) that outputs the identified noise classification information. This configuration makes it possible to objectively evaluate radio signals received by a radio device, such as an in-vehicle radio device.

[0053] The output unit generates an evaluation screen (e.g., 400, 410, 420) using the identified noise classification information and displays the evaluation screen. The evaluation screen (e.g., 400) includes a graph in which the time at which a radio signal was received is associated with the noise classification information of the radio signal. The acquisition unit further acquires location information of a mobile object equipped with a radio device, and the evaluation screen (e.g., 410) is configured to associate the position at which the radio signal was received on the mobile object's travel route with the noise classification information of the radio signal. The acquisition unit also acquires radio information received by each of the multiple radio devices, and the evaluation screen (e.g., 420) is configured to enable comparison of the noise classification information of the radio signals corresponding to each of the multiple radio devices. This configuration allows the user to easily understand the evaluation results for the radio device.

[0054] <Embodiment 2> A second embodiment of the present disclosure will be described. In the first embodiment, a configuration focusing on evaluation of an in-vehicle radio device has been described. In the second embodiment, a configuration in which adjustment of parameters related to noise reduction processing of a radio signal is performed based on a noise identification result will be described.

[0055] In this embodiment, two radio devices will be used for explanation: one that serves as a quality standard, and one that performs parameter adjustment taking noise into consideration. For convenience, the radio device having the quality that serves as the standard will be referred to as the "reference device." Furthermore, the radio device that is the target of adjusting the control parameters for noise reduction processing based on the quality of the reference device will be referred to as the "device to be adjusted." The reference device is assumed to be configured in advance to be able to output audio signals at a certain quality.

[0056] [Adjustment Processing] FIG. 5 illustrates a flowchart of the adjustment process according to this embodiment. The trained model for noise discrimination used in the adjustment process according to this embodiment has the same configuration as that of Embodiment 1. This processing flow may be executed, for example, by an information processing device that can acquire both the radio signal received by the reference device and the radio signal received by the device to be adjusted. The information processing device may be configured as having the same device configuration as the radio evaluation system 200 shown in FIG. 2, or may be configured as a device with a different device configuration. Here, the processing will be described assuming that the radio evaluation system 200 is the subject of processing.

[0057] The radio evaluation system 200 acquires an audio signal A received by a reference device (step S501).

[0058] The radio evaluation system 200 applies the trained model to the audio signal A acquired in step S501, thereby identifying the noise level in the audio signal A (step S502).

[0059] The radio evaluation system 200 compiles the noise levels identified in step S502 (step S503).

[0060] The radio evaluation system 200 acquires an audio signal B received by the device to be adjusted (step S504). It is more preferable that the audio signal B acquired here is acquired under the same conditions as those in the acquisition step in step S501.

[0061] The radio evaluation system 200 applies the trained model to the audio signal B acquired in step S504, thereby identifying the noise level in the audio signal B (step S505). The trained model used in this step is the same as the trained model used in step S502.

[0062] The radio evaluation system 200 compiles the noise levels identified in step S505 (step S506).

[0063] The radio evaluation system 200 compares the counting result in step S503 with the counting result in step S506 (step S507). For example, the highest noise level identified in a predetermined time span may be compared. Alternatively, the number of noise levels identified in a predetermined time span may be weighted and the results may be compared. The radio evaluation system 200 determines whether the noise level of audio signal A is higher than the noise level of audio signal B. If the noise level of audio signal A is high (step S507: YES), this processing flow ends. In other words, it is determined that the signal quality of the device to be adjusted is higher than that of the reference device and no adjustment is necessary. On the other hand, if the noise level of audio signal B is high (step S507: NO), the radio evaluation system 200 proceeds to step S508.

[0064] The radio evaluation system 200 adjusts the amount of high cut for the audio signal so that the noise level of audio signal B is equal to or lower than the noise level of audio signal A (step S508). In this embodiment, parameters are adjusted to improve the quality of the device to be adjusted so that the quality is equal to or better than that of the reference device. Methods for adjusting the amount of high cut here include, for example, adjusting parameters of a predefined high cut algorithm, switching the application process, and changing the threshold value that defines the application range of the high cut.

[0065] The radio evaluation system 200 compares audio signal A and audio signal B and derives the fluctuation of the high cut (step S509). For example, Fast Fourier Transfer (FFT) processing is applied to each of audio signal A of the reference device and audio signal B of the device to be adjusted, and the frequency characteristics of a predetermined frequency band are further derived. Then, the difference between frequency characteristic A of audio signal A as a reference and frequency characteristic B of audio signal B is taken as the amount of change. The amount of change is acquired over time and is derived as the fluctuation of the high cut.

[0066] Based on the fluctuation (difference) of the high cuts derived in step S509, the radio evaluation system 200 determines whether the amount of high cut for audio signal A is equal to or greater than the amount of high cut for audio signal B. If the amount of high cut for audio signal A is equal to or greater than the amount of high cut for audio signal B (step S510: YES), the processing of the radio evaluation system 200 proceeds to step S511. On the other hand, if the amount of high cut for audio signal A is smaller than the amount of high cut for audio signal B (step S510: NO), the processing of the radio evaluation system 200 returns to step S505 and repeats the processing. In this case, the amount of high cut for audio signal B is adjusted again so that it does not become greater than the amount of high cut for audio signal A. In this case, the amount of high cut for audio signal B may be adjusted to fall within a certain range, using the amount of high cut for audio signal A as a reference.

[0067] The radio evaluation system 200 updates the control parameters of the device to be adjusted based on the amount of high cut at this point (step S511), and then ends this processing flow.

[0068] FIG. 6 is a graph showing an example of noise level adjustment. The horizontal axis of each graph indicates time [s], and the vertical axis indicates noise level. Time corresponds to each other in each graph. The top graph shows the change in noise level of audio signal A. The middle graph shows the noise level of audio signal B before adjustment. The bottom graph shows the noise level of audio signal B after adjustment.

[0069] The noise level of audio signal B is adjusted to be suppressed using the noise level of audio signal A in the upper row as a reference. At this time, the control parameters are adjusted so that high cut is not applied above a predetermined range using audio signal A as a reference. In other words, the control parameters of the device to be adjusted are adjusted so that the reference device and the device to be adjusted have the same quality.

[0070] The above-described control parameter adjustment may be performed at any timing. The above-described control parameter adjustment technique may be applied, for example, when adjusting the parameters of a radio device before product shipment. Alternatively, the control parameter adjustment technique may be applied to a radio device that has already been shipped, by providing a new data set corresponding to a reference device and executing the above-described control parameter adjustment technique on the radio device side to subsequently adjust the quality of the new data set (for reference) to that of the new data set.

[0071] As described above, according to this embodiment, it is possible to adjust the control parameters for processing noise based on the evaluation results of a plurality of radio devices, using the quality of one of the radio devices as a standard.

[0072] <Third Embodiment> A third embodiment of the present disclosure will be described. In the first embodiment, a configuration focusing on evaluation of a radio device was described. In the third embodiment, a configuration in which noise processing related to the output of a radio signal, particularly switching of application of high cut, is performed based on the noise identification result will be described.

[0073] 7A and 7B are graphs illustrating the concept and effect of applying a high cut to an audio signal. In each of the graphs shown in Fig. 7A and 7B, the vertical axis represents dB and the horizontal axis represents frequency [Hz].

[0074] Graph 700 shows an example of an audio signal for content in which high-frequency noise is relatively noticeable. Region 701 in graph 700 indicates the range in which a user is likely to recognize noise when the audio signal is played back. Therefore, it is preferable to apply high-cut processing to the audio signal included in region 701 and adjust the shape to approximate dashed line 702, thereby suppressing noise. This enables playback of higher quality content.

[0075] Graph 710 shows an example of an audio signal for content in which high-frequency noise is relatively inconspicuous. With an audio signal like that shown in graph 710, the original audio contains a large amount of high-frequency components, so even if high-frequency noise is included, it is unlikely for the user to recognize it as noise. Suppose that high-cut processing in area 711 is applied to an audio signal like that shown in graph 710, as in FIG. 7A , and the signal is adjusted as shown by dashed line 712. In this case, the high-frequency components of the original audio signal, as shown in area 713, are also removed. As a result, the original audio signal is lost, resulting in a muffled sound, for example, which leads to a decrease in audio quality.

[0076] In consideration of the above-mentioned problems, in this embodiment, the application of high cut is switched when a radio signal is output.

[0077] [Device configuration] Fig. 8 is a block diagram showing an example of the configuration of a radio device according to this embodiment. In Fig. 8, arrows indicate an example of the flow of signals to each block. Note that the configuration shown in Fig. 8 is just one example. Therefore, the radio device may further include components other than those shown in Fig. 8.

[0078] Radio device 800 according to the present embodiment receives radio signals corresponding to one or more services and outputs a selected service from the one or more services as audio. Radio device 800 may be fixedly installed in a vehicle or may be detachably configured.

[0079] The radio device according to the present embodiment is configured to be mounted in a vehicle, for example, and operable by a user. The radio device may be configured as a single device that outputs radio signals, or may be configured as one function of a so-called navigation device.

[0080] In radio device 800, signal receiving unit 801 receives radio signals in a predetermined frequency band transmitted from a nearby base station using an antenna (not shown), and provides the signals to audio demodulation unit 802. In this embodiment, the standard and type of the radio signal are not particularly limited, and any desired radio signal may be used.

[0081] The audio demodulation unit 802 receives (filters) radio signals of a frequency specified by a user's station selection, etc., from the radio signals received by the signal receiving unit 801, and outputs them as audio signals to the delay correction unit 803 and the noise identification unit 804.

[0082] The delay correction unit 803 controls the output timing etc. in response to a delay caused by processing by the noise identification unit 204 (described later) etc. The delay correction unit 803 outputs the corrected audio signal to the noise processing unit 806.

[0083] 1 described in the first embodiment, the noise discrimination unit 804 applies the noise discrimination process 112 to the audio signal input from the audio demodulation unit 802 using the trained model. Then, the noise discrimination unit 804 outputs the discrimination result to the high cut control table reference unit 805.

[0084] The high cut control table reference unit 805 references a high cut control table that defines parameters for performing high cut, based on the identification result of the noise identification unit 804. An example of the high cut control table will be described later with reference to Fig. 9. The high cut control table reference unit 805 outputs the reference result to the noise processing unit 806.

[0085] The noise processing unit 806 performs noise removal processing on the audio signal input from the delay correction unit 203 based on parameters corresponding to the reference result from the high cut control table reference unit 805. Then, the noise processing unit 806 outputs the processed audio signal to the audio output unit 807.

[0086] The audio output unit 807 outputs audio based on the audio signal from the noise processing unit 806 using a speaker (not shown) or the like.

[0087] Each block shown in Fig. 8 may be realized by, for example, a control unit or a storage unit (not shown). The control unit (not shown) may be configured using, for example, a CPU, an MPU, a DSP, a GPU, or an FPGA. The storage unit (not shown) is a storage area for storing and holding various data, and may be configured, for example, by a non-volatile storage area such as a ROM or HDD, or a volatile storage area such as a RAM. For example, the control unit may read and execute various data and programs stored in the storage unit, thereby realizing some of the functions of the blocks shown in Fig. 8.

[0088] In this embodiment, high-cut processing is performed on audio signals in which noise is easily noticeable, as shown in Fig. 7A. On the other hand, high-cut processing is suppressed for audio signals in which noise is less noticeable. As shown in Fig. 9, for an audio signal in which noise is less noticeable, such as graph 710, the range in which high-cut is applied is changed, as shown by dashed line 901, for example, and removal of the original signal waveform is suppressed.

[0089] In this embodiment, for an audio signal such as that shown in FIG. 7A, training data is defined as having a high noise level, and training processing is performed. As a result, when a trained model is applied to a signal such as that shown in FIG. 7A, the signal is identified as having a high noise level, and the effect of applying high-cut filtering is enhanced. On the other hand, for an audio signal such as that shown in FIG. 7B, training data is defined as having a low noise level, and training processing is performed. As a result, when a trained model is applied to a signal such as that shown in FIG. 7B, the signal is identified as having a low noise level, and the effect of applying high-cut filtering is suppressed.

[0090] FIG. 10 shows an example of the configuration of a high cut control table stored and managed by the high cut control table reference unit 805 according to this embodiment. Two examples are shown here. The high cut control table 1001 shows an example of a table in which the high cut control effect amount [dB] is defined according to the noise level value according to the trained model. When the high cut control table 1001 is used, the high cut amount is adjusted according to the noise level. The high cut control effect amount corresponds to the high cut amount.

[0091] The high-cut control table 1002 shows an example of a table used when the trained model is configured to identify noise types instead of noise levels. In this case, the trained model undergoes a learning process for an input audio signal so as to identify the type of noise contained in the audio signal. Here, multipath noise, electrical component noise, and weak electric field noise are given as examples of noise types. Noise processing, for which control parameters are pre-set, is shown in association with each type of noise. For example, when multipath noise is identified by the trained model, high-cut A processing is applied. When electrical component noise is identified, noise canceling processing is applied. When weak electric field noise is identified, high-cut B processing and soft mute processing are applied. These noise processing methods are not particularly limited, and any control parameters and application modes are pre-defined and configured to be switchable by the noise processing unit 806.

[0092] 10 are merely examples and are not intended to be limiting. The correspondence relationship in the high cut control table may be determined manually or may be defined based on the relationship between the "reference device" and the "device to be adjusted" as described in the second embodiment.

[0093] [Audio output processing] FIG. 11 illustrates a flowchart of the audio output process according to this embodiment. The trained model for noise identification used in the audio output process has the same configuration as in the first embodiment. This processing flow is realized by the cooperation and execution of the various components of the radio device 800 shown in FIG. 8. For ease of explanation, the processing will be collectively described as being performed by the radio device 800.

[0094] The radio device 800 acquires a radio signal (step S1101).

[0095] The radio device 800 performs demodulation processing on the radio signal acquired in step S1101 to acquire an audio signal (step S1102).

[0096] The radio device 800 performs noise discrimination by applying the trained model to the audio signal demodulated in step S1102 (step S1103). Here, the trained model that outputs the noise level is used, assuming the high cut control table 1001 in FIG.

[0097] The radio device 800 derives the amount of high cut for the audio signal based on the noise level identified in step S1103 (step S1104). In this example, the radio device 800 refers to a high cut control table 1001 shown in FIG. 10 to identify the amount of high cut corresponding to the noise level.

[0098] The radio device 800 performs noise reduction processing on the audio signal based on the amount of high cut calculated in step S1104 (step S1105).

[0099] The radio device 800 outputs audio using the radio signal to which the noise reduction process has been applied in step S1105 (step S1106). Then, this processing flow ends. Note that the above processing flow may be repeated while receiving an audio signal.

[0100] (Variation) The high-cut control table used in radio device 800 shown in Fig. 8 may be updated as needed. For example, the high-cut generation table may be generated and updated by table generation system 1200 provided on a network (not shown), and radio device 800 may be configured to acquire the table as needed.

[0101] Fig. 12 shows an example of a system configuration for updating a high-cut control table. The configuration of radio device 800 is assumed to be the same as that of Fig. 8, and detailed description thereof will be omitted here. Table generation system 1200 is configured to be able to communicate with radio device 800 via a network (not shown). Table generation system 1200 may be configured as, for example, a cloud-based computing device.

[0102] The reference signal receiving unit 1201 receives a reference signal, which is a reference signal. The audio demodulation unit 1202 demodulates the reference signal received by the reference signal receiving unit 1201. The noise identifying unit 1203 identifies the noise level of the reference signal using a trained model for identifying noise levels. This trained model is the same as the trained model used by the noise identifying unit 804 of the radio device 800. The high cut amount derivation unit 1204 derives the high cut amount based on the noise level identified by the noise identifying unit 1203. Then, the high cut amount derivation unit 1204 associates the noise level with the high cut amount and passes it to the high cut control table generation unit 1208.

[0103] The evaluation signal receiving unit 1205 receives the evaluation signal with a configuration equivalent to that of the radio device to be evaluated. In this example, the evaluation signal receiving unit 1205 is assumed to have a configuration equivalent to that of the signal receiving unit 801 of the radio device 800. Furthermore, the content of the received audio signal is the same for the reference signal and the evaluation signal. The audio demodulation unit 1206 demodulates the evaluation signal received by the evaluation signal receiving unit 1205. The noise identification unit 1207 identifies the noise level of the evaluation signal using a trained model for identifying noise levels. This trained model is the same as the trained model used in the noise identification unit 804 and noise identification unit 1203 of the radio device 800. The noise identification unit 1207 passes the noise level identification result to the high-cut control table generation unit 1208.

[0104] The high-cut control table generation unit 1208 generates a high-cut control table based on the correspondence between the noise level of the reference signal, the amount of high-cut, and the noise level of the evaluation signal obtained for the same content. As a result, for example, the high-cut control table 1001 shown in Fig. 10 is generated for the radio device 800. The high-cut control table here may be generated in the same manner as the flowchart of Fig. 5 described in the second embodiment. The generated high-cut control table is then passed to the high-cut control table provision unit 1209.

[0105] The high cut control table provider 1209 manages, updates, and stores the generated high cut control table. The high cut control table provider 1209 then provides the high cut control table to the radio device 800 at a predetermined timing. The predetermined timing may be configured to periodically transmit the high cut control table to the radio device 800, or may be configured to transmit the table in response to a request from the radio device 800.

[0106] As described above, a radio device (e.g., 800) according to this embodiment includes a receiving unit (e.g., 801, 802) that receives a radio signal, an identifying unit (e.g., 804) that identifies noise classification information of the radio signal received by the receiving unit using a trained model that receives the radio signal as input and outputs noise classification information of the radio signal, an identifying unit (e.g., 805) that identifies noise processing for the radio signal using the identified noise classification information, and an output control unit (e.g., 806, 807) that applies the identified noise processing to the radio signal and outputs audio. This enables the radio device to objectively evaluate the audio signal received and apply appropriate noise processing in accordance with the evaluation results. In particular, by classifying noise using a trained model that has been trained taking into account the conspicuousness of noise, the application of excessive noise processing can be suppressed, thereby improving the output quality of the radio device.

[0107] The identification unit also identifies noise processing for the radio signal using a table (e.g., 1001, 1002) in which noise processing corresponding to the classification information is predefined. This makes it possible to predefine appropriate noise processing and parameters according to the noise classification information.

[0108] <Other embodiments> In the above embodiment, a configuration has been described in which high-cut processing for noise in high-frequency components is assumed. However, the present disclosure is not limited to this, and the method of the present disclosure may be applied to noise in any band. In this case, a trained model corresponding to the noise of interest may be generated and then used.

[0109] Furthermore, the frequency to be subjected to the high-cut processing is assumed to be, for example, about 7 to 15 kHz, but is not limited to this.

[0110] In addition, the functions of one or more of the above-described embodiments can be realized by supplying a program and application to a system or device via a network or storage medium, and having one or more processors in the computer of the system or device read and execute the program.

[0111] Alternatively, it may be realized by a circuit that realizes one or more functions (for example, an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array)).

[0112] Although various embodiments have been described above with reference to the drawings, it goes without saying that the present disclosure is not limited to these examples. It is clear to those skilled in the art that various modifications, alterations, substitutions, additions, deletions, and equivalents may be made within the scope of the claims, and it is understood that these also fall within the technical scope of the present disclosure. Furthermore, the components of the various embodiments described above may be combined in any manner without departing from the spirit of the invention.

[0113] (Addendum) The above description of the embodiments discloses the following techniques. (Technology 1) an acquisition unit that acquires a radio signal received by the radio device; an identification unit that uses a trained model that receives a radio signal as an input and outputs noise classification information of the radio signal to identify noise classification information of the radio signal acquired by the acquisition unit; an output unit that outputs classification information of the noise identified by the identification unit; A radio rating system having: This configuration makes it possible to objectively evaluate radio signals received by a radio device such as an in-vehicle radio device.

[0114] (Technology 2) The radio evaluation system according to technology 1, wherein the output unit generates an evaluation screen using classification information of the noise identified by the identification unit and displays the evaluation screen. This configuration allows the user to easily understand the evaluation results for the radio device.

[0115] (Technology 3) The radio evaluation system according to technology 2, wherein the evaluation screen includes a graph in which the time at which the radio signal was received is associated with classification information of noise in the radio signal. According to this configuration, the user can easily grasp the evaluation result for the radio device, in particular the relationship between the reception timing of the radio signal and the classification information.

[0116] (Technology 4) The acquisition unit further acquires location information of a mobile body on which the radio device is mounted, The radio evaluation system according to Technology 2 or Technology 3, wherein the evaluation screen is configured by associating the position where the radio signal was received on the moving path of the moving body with classification information of noise of the radio signal. According to this configuration, the user can easily grasp the evaluation result for the radio device, in particular the relationship between the reception position of the radio signal and the classification information.

[0117] (Technology 5) the acquisition unit acquires radio information received by each of a plurality of radio devices; The radio evaluation system according to any one of Technology 2 to Technology 4, wherein the evaluation screen is configured to enable comparison of noise classification information of radio signals corresponding to each of the plurality of radio devices. This configuration allows the user to easily understand the evaluation results for the radio device, particularly the results of a comparison of the quality of a plurality of radio devices.

[0118] (Technology 6) a receiving unit for receiving a radio signal; an identification unit that uses a trained model that receives a radio signal as an input and outputs noise classification information of the radio signal to identify noise classification information of the radio signal received by the receiving unit; an identification unit that identifies a noise processing method for the radio signal using classification information of the noise identified by the identification unit; an output control unit that applies the noise processing specified by the specifying unit to the radio signal and outputs the result as sound; A radio device having: This configuration enables the radio device to objectively evaluate the audio signal received and apply appropriate noise processing in response to the evaluation results. In particular, by classifying noise using a trained model that takes into account the conspicuousness of noise, the application of excessive noise processing can be suppressed, thereby improving the output quality of the radio device.

[0119] (Technology 7) The radio device according to technology 6, wherein the specification unit specifies noise processing for the radio signal using a table in which noise processing corresponding to the classification information is predefined. This makes it possible to predefine appropriate noise processing and parameters according to noise classification information.

[0120] (Technology 8) an acquisition step of acquiring a radio signal received by a radio device; an identification step of identifying noise classification information of the radio signal acquired in the acquisition step using a trained model that receives a radio signal as input and outputs noise classification information of the radio signal; an output step of outputting classification information of the noise identified in the identification step; A method for evaluating a radio device having the above configuration. This configuration makes it possible to objectively evaluate radio signals received by a radio device such as an in-vehicle radio device.

[0121] (Technology 9) a receiving step of receiving a radio signal; an identification process for identifying noise classification information of the radio signal received in the receiving process using a trained model that receives a radio signal as input and outputs noise classification information of the radio signal; a step of specifying a noise processing method for the radio signal using the classification information of the noise identified in the identification step; an output control step of applying the noise processing specified in the specifying step to the radio signal and outputting the result as audio; A method for controlling a radio device having the above configuration. This configuration enables the radio device to objectively evaluate the audio signal received and apply appropriate noise processing in response to the evaluation results. In particular, by classifying noise using a trained model that takes into account the conspicuousness of noise, the application of excessive noise processing can be suppressed, thereby improving the output quality of the radio device. [Industrial Applicability]

[0122] The present disclosure is useful as a radio evaluation system, a radio evaluation method, a radio device, and a method for controlling a radio device. [Explanation of symbols]

[0123] 200...Radio Evaluation System 201...Signal receiving unit 202...Audio demodulation section 203...Delay correction unit 204...Noise discrimination unit 205...Noise processing section 206...Audio output unit 207...Noise counting section 208...Screen generation section 209...Screen output section 800...Radio equipment 801...Signal receiving unit 802...Audio demodulation section 803...Delay correction unit 804...Noise discrimination unit 805...High cut control table reference section 806...Noise processing section 807...Audio output section 1200...Table generation system 1201...Reference signal receiving unit 1202...Audio demodulation section 1203...Noise discrimination unit 1204...High cut amount derivation part 1205...Evaluation signal receiving unit 1206...Audio demodulation section 1207...Noise discrimination unit 1208...High cut control table generation unit 1209…High-cut control table provider

Claims

1. an acquisition unit that acquires a radio signal received by the radio device; an identification unit that uses a trained model that receives a radio signal as an input and outputs noise classification information of the radio signal to identify noise classification information of the radio signal acquired by the acquisition unit; an output unit that outputs classification information of the noise identified by the identification unit; A radio rating system having:

2. The radio evaluation system according to claim 1 , wherein the output unit generates an evaluation screen using classification information of the noise identified by the identification unit, and displays the evaluation screen.

3. The radio evaluation system according to claim 2 , wherein the evaluation screen includes a graph in which the time at which the radio signal was received is associated with classification information of noise in the radio signal.

4. The acquisition unit further acquires location information of a mobile body on which the radio device is mounted, 3. The radio evaluation system according to claim 2, wherein the evaluation screen is configured by associating a position on a moving route of the mobile object where the radio signal was received with classification information of noise of the radio signal.

5. the acquisition unit acquires radio information received by each of a plurality of radio devices; The radio evaluation system according to claim 2 , wherein the evaluation screen is configured to enable comparison of classification information of noise of radio signals corresponding to each of the plurality of radio devices.

6. a receiving unit for receiving a radio signal; an identification unit that uses a trained model that receives a radio signal as an input and outputs noise classification information of the radio signal to identify noise classification information of the radio signal received by the receiving unit; an identification unit that identifies a noise processing method for the radio signal using classification information of the noise identified by the identification unit; an output control unit that applies the noise processing specified by the specifying unit to the radio signal and outputs the result as sound; A radio device having:

7. The radio device according to claim 6 , wherein the specifying unit specifies the noise processing for the radio signal using a table in which noise processing corresponding to the classification information is predefined.

8. an acquisition step of acquiring a radio signal received by a radio device; an identification step of identifying noise classification information of the radio signal acquired in the acquisition step using a trained model that receives a radio signal as input and outputs noise classification information of the radio signal; an output step of outputting classification information of the noise identified in the identification step; A method for evaluating a radio device having the above configuration.

9. a receiving step of receiving a radio signal; an identification process for identifying noise classification information of the radio signal received in the receiving process using a trained model that receives a radio signal as input and outputs noise classification information of the radio signal; a step of specifying a noise processing method for the radio signal using the classification information of the noise identified in the identification step; an output control step of applying the noise processing specified in the specifying step to the radio signal and outputting the result as audio; A method for controlling a radio device having the above configuration.

Images