Device wake-up method, storage medium and electronic device

By separating the received data of smart devices, calculating the noise immunity and determining the distance between the device and the sound source, the problem of low accuracy during multiple devices is solved, and higher accuracy of device wake-up is achieved.

CN116206618BActive Publication Date: 2024-11-19HAIER YOUJIA INTELLIGENT TECH (BEIJING) CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211711795.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-29
Publication Date
2024-11-19
Estimated Expiration
2042-12-29

AI Technical Summary

Technical Problem

In the prior art, when multiple devices wake up at the same time, the energy information received by the microphone is easily disturbed by reverb and external noise, resulting in low accuracy of the device wake-up.

Method used

After obtaining the initial audio, the received data of each smart device is separated and the sound source is processed to obtain the wake-up audio data and the noise audio data, and the noise immunity is calculated. Based on this, the distance between the device and the sound source is determined and the distance is selected is selected. the device wakes up.

Benefits of technology

It reduces the interference of noisy audio data on device wake-up, improves the accuracy of device wake-up, and solves the problem of low device wake-up accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116206618B_ABST
    Figure CN116206618B_ABST
Patent Text Reader

Abstract

The present application discloses a device wake-up method, storage medium and electronic device, and relates to the field of smart home / intelligent family technology. The device wake-up method includes: determining at least two smart devices to be awakened that have received initial audio, and obtaining the reception data of the initial audio for each of the at least two smart devices to be awakened; performing sound source separation processing on the reception data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device; obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating the proportional coefficient of each wake-up audio data in the corresponding audio data sum; calculating the distance value based on the proportional coefficient, and determining the smart device with the smallest distance value from the at least two smart devices to be awakened for awakening. The present application solves the technical problem of low accuracy of device wake-up in the related art.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of smart home / intelligent family technology, and more specifically, to a device wake-up method, a storage medium, and an electronic device. Background Art

[0002] With the continuous development of voice interaction applications, the interaction of a single device no longer meets people's needs. The coordinated awakening of multiple devices in the same scene can best enable people to experience the futuristic sense of technology brought by artificial intelligence.

[0003] However, when multiple devices are awakened at the same time, the common method in the prior art is to use the energy information received by the microphone in the device to reflect the distance between the device to be awakened and the sound source, and then decide which device to respond first. This not only requires additional energy calibration work, but is also easily affected by reverberation and external noise factors, making it impossible for the energy information received by the microphone to truly and accurately reflect the distance between the device to be awakened and the sound source, thereby resulting in low accuracy in device awakening.

[0004] Therefore, there is a problem in the related art that the accuracy of device wake-up is low. Summary of the invention

[0005] The embodiments of the present application provide a device wake-up method, a storage medium, and an electronic device to at least solve the technical problem of low accuracy of device wake-up in the related art.

[0006] According to one aspect of an embodiment of the present application, a device wake-up method is provided, comprising: when initial audio is obtained, determining at least two smart devices to be awakened that have received the initial audio, and obtaining reception data of the initial audio by each of the at least two smart devices to be awakened; when the reception data is audio data for awakening the smart device, performing sound source separation processing on the reception data of each of the smart devices to obtain each pair of wake-up audio data and noise audio data corresponding to each of the smart devices; obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating the proportional coefficient of each of the wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the corresponding noise resistance capability of each of the smart devices; calculating the distance value between each of the smart devices and the audio source that emits the initial audio based on the proportional coefficient, and determining the smart device with the smallest distance value from the at least two smart devices to be awakened for awakening.

[0007] According to another aspect of an embodiment of the present application, a device wake-up device is also provided, including: an acquisition unit, which is used to determine at least two smart devices to be awakened that have received the initial audio when the initial audio is acquired, and obtain the reception data of the initial audio for each of the at least two smart devices to be awakened; a sound source separation unit, which is used to perform sound source separation processing on the reception data of each of the above-mentioned smart devices when the above-mentioned reception data is audio data used to wake up the above-mentioned smart device, so as to obtain each pair of wake-up audio data and noise audio data corresponding to each of the above-mentioned smart devices; a calculation unit, which is used to obtain the audio data sum of the above-mentioned each pair of wake-up audio data and noise audio data, and calculate the proportional coefficient of each of the above-mentioned wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the corresponding noise resistance capability of each of the above-mentioned smart devices; a wake-up unit, which is used to calculate the distance value between each of the above-mentioned smart devices and the audio sound source that emits the above-mentioned initial audio based on the above-mentioned proportional coefficient, and determine the smart device with the smallest distance value from the at least two smart devices to be awakened for wake-up.

[0008] As an optional solution, the sound source separation unit includes: a first acquisition module, used to acquire the first audio data and the second audio data obtained by performing the sound source separation processing on the received data of each of the smart devices; a calculation module, used to calculate the first confidence corresponding to the first audio data and the second confidence corresponding to the second audio data; a first determination module, used to compare the first confidence and the second confidence, and determine the audio data with high confidence as the wake-up audio data, and determine the audio data with low confidence as the noise audio data.

[0009] As an optional solution, the wake-up unit includes: a second determination module, used to determine the above-mentioned proportional coefficient as the anti-noise parameter corresponding to each of the above-mentioned smart devices; a third determination module, used to determine the above-mentioned distance value between each of the above-mentioned smart devices and the audio source that emits the above-mentioned initial audio based on the above-mentioned anti-noise parameter.

[0010] As an optional scheme, the above-mentioned third determination module includes: an acquisition submodule, used to obtain the diffusion ratio corresponding to the above-mentioned initial data received by the above-mentioned each smart device; a determination submodule, used to determine the direct mixing ratio corresponding to each of the above-mentioned smart devices based on the above-mentioned anti-noise parameters and the above-mentioned diffusion ratio, wherein the above-mentioned direct mixing ratio is used to indicate the above-mentioned distance value between each of the above-mentioned smart devices and the audio sound source that emits the above-mentioned initial audio.

[0011] As an optional solution, the acquisition submodule includes: an acquisition subunit, used to acquire the coherence function and the scattering noise field corresponding to the initial data based on the initial data received by each of the smart devices; and a determination subunit, used to integrate the coherence function and the scattering noise field to obtain the diffusion ratio corresponding to the initial data.

[0012] As an optional solution, the sound source separation unit includes: a fourth determination module, used to determine the preset environment matching the at least two smart devices to be awakened; a second acquisition module, used to acquire the audio recognition model corresponding to the preset environment, wherein the audio recognition model is a neural network model for recognizing audio obtained by training using audio samples of the preset environment; a fifth determination module, used to input the received data of each of the smart devices into the audio recognition model in turn to obtain each pair of wake-up audio data and noise audio data corresponding to each of the smart devices.

[0013] As an optional solution, the above-mentioned device also includes: a third acquisition module, used to obtain multiple audio samples of the above-mentioned preset environments before determining the preset environments that match the above-mentioned at least two smart devices to be awakened, wherein the audio samples of the above-mentioned preset environments carry marked noise audio data in the above-mentioned preset environments; a training module, used to train the initial audio recognition model using the above-mentioned multiple audio samples of the above-mentioned preset environments before determining the preset environments that match the above-mentioned at least two smart devices to be awakened, until the trained audio recognition model is obtained.

[0014] According to another aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above device wake-up method.

[0015] According to another aspect of an embodiment of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the device wake-up method through the computer program.

[0016] In an embodiment of the present application, when the initial audio is obtained, at least two smart devices to be awakened that have received the initial audio are determined, and the reception data of the initial audio by each of the at least two smart devices to be awakened is obtained; when the received data is audio data for awakening the smart device, the received data of each of the smart devices is subjected to sound source separation processing to obtain each pair of wake-up audio data and noise audio data corresponding to each of the smart devices; the audio data sum of each pair of wake-up audio data and noise audio data is obtained, and the proportional coefficient of each of the wake-up audio data in the corresponding audio data sum is calculated, wherein the proportional coefficient is used to represent the anti-noise capability corresponding to each of the smart devices; the distance value between each of the smart devices and the audio source that emits the initial audio is calculated based on the proportional coefficient, and the smart device with the smallest distance value is determined from the at least two smart devices to be awakened for awakening;

[0017] By using the above-mentioned device wake-up method, the means in the prior art of directly calculating energy information based on the received data obtained by the smart device to reflect the distance between the smart device and the sound source is abandoned. Instead, after the smart device obtains the received data, sound source separation processing is first performed, and then the distance between each smart device and the sound source is determined based on the wake-up audio data and noise audio data obtained by the sound source separation processing, thereby reducing the interference of the noise audio data contained in the initial received data on the distance between the smart device and the sound source, and thus solving the defect that external noise affects the audio data received by the microphone in the smart device, thereby achieving the technical effect of improving the accuracy of device wake-up, and solving the technical problem of low device wake-up accuracy in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 is a schematic diagram of a hardware environment of a device wake-up method according to an embodiment of the present application;

[0021] Figure 2 is a schematic diagram of a process of an optional device wake-up method according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of an optional device wake-up method according to an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of another optional device wake-up method according to an embodiment of the present application;

[0024] Figure 5 is a schematic diagram of an optional information processing device according to an embodiment of the present invention;

[0025] Figure 6 is a schematic structural diagram of an optional electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] According to one aspect of an embodiment of the present application, a method for interacting with a smart home device is provided. The method for interacting with a smart home device is widely used in smart home (Smart Home), smart home, smart home device ecosystem, smart home (IntelligenceHouse) ecosystem and other whole-house intelligent digital control application scenarios. Optionally, in this embodiment, the method for interacting with the smart home device can be applied to Figure 1 In the hardware environment composed of the terminal device 102 and the server 104 shown in FIG. Figure 1As shown, the server 104 is connected to the terminal device 102 via a network, and can be used to provide services (such as application services, etc.) for the terminal or a client installed on the terminal. A database can be set on the server or independently of the server to provide data storage services for the server 104. Cloud computing and / or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.

[0029] The network may include but is not limited to at least one of the following: wired network, wireless network. The wired network may include but is not limited to at least one of the following: wide area network, metropolitan area network, local area network, and the wireless network may include but is not limited to at least one of the following: WIFI (Wireless Fidelity), Bluetooth. The terminal device 102 may be but is not limited to a PC, a mobile phone, a tablet computer, a smart air conditioner, a smart range hood, a smart refrigerator, a smart oven, a smart stove, a smart washing machine, a smart water heater, a smart washing device, a smart dishwasher, a smart projection device, a smart TV, a smart clothes drying rack, a smart curtain, a smart audio and video, a smart socket, a smart speaker, a smart fresh air device, a smart kitchen and bathroom device, a smart bathroom device, a smart sweeping robot, a smart window cleaning robot, a smart mopping robot, a smart air purification device, a smart steamer, a smart microwave oven, a smart kitchen treasure, a smart purifier, a smart water dispenser, a smart door lock, etc.

[0030] Optionally, as an optional implementation, as Figure 2 As shown, the device wake-up method includes:

[0031] S202, when the initial audio is obtained, determining at least two smart devices to be awakened that have received the initial audio, and obtaining reception data of the initial audio by each of the at least two smart devices to be awakened;

[0032] S204, when the received data is audio data for waking up the smart device, performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device;

[0033] S206, obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating a proportional coefficient of each wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the corresponding anti-noise capability of each smart device;

[0034] S208, calculating the distance value between each smart device and the audio source that emits the initial audio based on the proportional coefficient, and determining the smart device with the smallest distance value from at least two smart devices to be awakened to awaken.

[0035] Optionally, in this embodiment, the executor of the above-mentioned device wake-up method may include but is not limited to a system that can uniformly process the data received by each smart device, such as a local system or a cloud system. Taking the cloud system as an example, after each smart device receives its own audio data, it will upload it to the cloud system, and the cloud system will then summarize and process the audio data, and when the audio data carries a wake-up instruction, further determine which smart device is the target device to be awakened by the wake-up instruction. For example, if the cloud system determines that the above-mentioned target device is smart device A, it will return the wake-up instruction to smart device A to wake up smart device A.

[0036] Optionally, in this embodiment, the device wake-up method can be applied in, but is not limited to, smart home application scenarios. In smart home application scenarios, there are multiple smart devices that can respond to voice command information issued by users, be awakened, and perform operations corresponding to user commands. With the continuous development of voice interaction applications, multiple smart devices in smart home scenarios can respond to some voice commands of users and be awakened. In this way, when multiple smart devices are awakened at the same time, there is a problem of how to determine which smart device should be awakened first.

[0037] To address this issue, the prior art often uses the energy information corresponding to the voice command information received by the microphone in the smart device to reflect the distance between the device to be awakened and the sound source, and then decides which device to respond first. This not only requires additional energy calibration work, but is also easily affected by reverberation and external noise factors, making it impossible for the energy information received by the microphone to truly and accurately reflect the distance between the device to be awakened and the sound source, thereby resulting in low accuracy in device awakening.

[0038] To further illustrate, Figure 3 As shown, in a smart home environment, a user 302 sends a first audio 308 to wake up the smart device closest to him / her. However, at this time, an interfering second audio 310 is transmitted from the vicinity of a second smart device 306. Thus, under the influence of the second audio 310, the energy information corresponding to the second smart device 306 is greater than the energy information corresponding to the first smart device 304, and the second smart device 306 responds to the wake-up priority.

[0039] It should be noted that the user 302 actually wants to wake up the first smart device 304 that is closest to him / her, so that he / she can realize subsequent interaction with the smart device within a shorter distance. However, under the interference of external noise, the device priority wake-up method based on energy information has the defect of low wake-up accuracy due to the interference of external noise.

[0040] The above-mentioned device wake-up method of the present embodiment is used to abandon the means in the prior art of directly calculating energy information based on the received data obtained by the smart device to reflect the distance between the smart device and the sound source. Instead, after the smart device obtains the received data, sound source separation processing is first performed, and then the distance between each smart device and the sound source is determined based on the wake-up audio data and noise audio data obtained by the sound source separation processing, thereby reducing the interference of the noise audio data contained in the initial received data on the distance between the smart device and the sound source, and further solving the defect that external noise affects the audio data received by the microphone in the smart device, thereby achieving the technical effect of improving the accuracy of device wake-up and solving the defect of low wake-up accuracy caused by the interference of external noise on the device wake-up.

[0041] Optionally, in this embodiment, the smart device may be, but is not limited to, a device that can receive and respond to voice command information from a user, and may include, but is not limited to, a smart speaker, a smart water heater, a smart TV, etc.

[0042] Optionally, in this embodiment, the initial audio may be, but is not limited to, audio emitted by a user and containing voice command information of the user, wherein the audio source of the initial audio corresponds to the position of the user when emitting the initial audio.

[0043] Optionally, in this embodiment, the data received by each smart device for the initial audio may be but is not limited to audio data obtained through a microphone array in the smart device, wherein the data received for the initial audio may be but is not limited to including audio information corresponding to the initial audio for waking up the smart device, and noise information transmitted from the outside that interferes with waking up the smart device.

[0044] Optionally, in this embodiment, the sound source separation processing may be but is not limited to using aux-iva to process the received data of the smart device to obtain the wake-up audio data and noise audio data corresponding to the smart device.

[0045] It should be noted that after performing sound source separation processing on the data received by the smart device, two channels of audio data can be obtained, but not limited to. Due to the replacement problem, it is impossible to determine which of the two channels of audio data corresponds to the wake-up signal of the target, and which channel of data corresponds to the noise signal that interferes with the wake-up of the device. In this regard, it is possible, but not limited to, to obtain and calculate the confidence corresponding to each of the two channels of audio data, and select the audio data with a high confidence as the wake-up signal of the corresponding target, wherein the confidence can be, but is not limited to, the probability information used to indicate that the audio data is the wake-up signal of the corresponding target.

[0046] Optionally, in this embodiment, after determining the wake-up audio data and noise audio data corresponding to each smart device, the distance value between each smart device and the audio source of the initial audio is determined based on the wake-up audio data and the noise audio data, and the smart device with the smallest distance value is woken up.

[0047] Through the embodiments provided by the present application, when the initial audio is obtained, at least two smart devices to be awakened that have received the initial audio are determined, and the reception data of the initial audio by each smart device among the at least two smart devices to be awakened is obtained; when the received data is audio data for awakening the smart device, the received data of each smart device is subjected to sound source separation processing to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device; the audio data sum of each pair of wake-up audio data and noise audio data is obtained, and the proportional coefficient of each wake-up audio data in the corresponding audio data sum is calculated, wherein the proportional coefficient is used to represent the corresponding anti-noise capability of each smart device; the distance value between each smart device and the audio sound source that emits the initial audio is calculated based on the proportional coefficient, and the smart device with the smallest distance value is determined from the at least two smart devices to be awakened for awakening;

[0048] By utilizing the above-mentioned device wake-up method, the means in the prior art of directly calculating energy information based on the received data obtained by the smart device to reflect the distance between the smart device and the sound source is abandoned. Instead, after the smart device obtains the received data, sound source separation processing is first performed, and then the distance between each smart device and the sound source is determined based on the wake-up audio data and noise audio data obtained by the sound source separation processing, thereby reducing the interference of the noise audio data contained in the initial received data on the distance between the smart device and the sound source, thereby solving the defect that external noise affects the audio data received by the microphone in the smart device, thereby achieving the technical effect of improving the accuracy of device wake-up.

[0049] As an optional solution, the received data of each smart device is subjected to sound source separation processing to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device, including:

[0050] S1, obtaining first audio data and second audio data obtained by performing sound source separation processing on received data of each smart device;

[0051] S2, calculating a first confidence level corresponding to the first audio data and a second confidence level corresponding to the second audio data;

[0052] S3, comparing the first confidence and the second confidence, and determining the audio data with high confidence as wake-up audio data, and determining the audio data with low confidence as noise audio data.

[0053] Optionally, in this embodiment, the first audio data and the second audio data obtained after sound source separation processing of the received data of each smart device are obtained, the first confidence corresponding to the first audio data and the second confidence corresponding to the second audio data are calculated and compared, the audio data with high confidence is determined as wake-up audio data, and the audio data with low confidence is determined as noise audio data.

[0054] Optionally, in this embodiment, the confidence level may be, but is not limited to, probability information indicating that the audio data is a wake-up signal of a corresponding target.

[0055] Through the embodiments provided by the present application, the first audio data and the second audio data obtained by performing sound source separation processing on the received data of each smart device are obtained; the first confidence corresponding to the first audio data and the second confidence corresponding to the second audio data are calculated; the first confidence and the second confidence are compared, and the audio data with high confidence is determined as the wake-up audio data, and the audio data with low confidence is determined as the noise audio data. After the sound source separation processing obtains two channels of audio data, the confidence of the wake-up signal used to indicate the target corresponding to the audio data is calculated, and then the wake-up audio data and the noise audio data are determined, thereby achieving the purpose of improving the accuracy of obtaining the wake-up audio data and the noise audio data, and improving the accuracy of obtaining the distance value between the smart device and the audio sound source of the initial audio, thereby achieving the technical effect of improving the accuracy of device wake-up.

[0056] As an optional solution, a distance value between each smart device and an audio source emitting the initial audio is calculated based on a proportionality coefficient, including:

[0057] S1, determining the proportionality coefficient as the anti-noise parameter corresponding to each smart device;

[0058] S2, based on the anti-noise parameter, determining a distance value between each smart device and an audio source that emits the initial audio.

[0059] Optionally, in this embodiment, the anti-noise parameter may be used, but is not limited to, to determine the distance between the smart device and an audio source that emits the initial audio.

[0060] For example, when the received data of the smart device is obtained and sound source separation processing is performed to obtain the first audio data Z1(l, f) and the second audio data Z2(l, f), and it is determined that the first confidence corresponding to the first audio data Z1(l, f) is greater than the second confidence corresponding to the second audio data Z2(l, f), the first audio data Z1(l, f) is determined as the wake-up audio data Z1(l, f), and the second audio data Z2(l, f) is determined as the noise audio data Z2(l, f).

[0061] It should be noted that l is used to indicate the time information of the received audio data, and f is used to indicate the frequency information of the received audio data.

[0062] To further illustrate, the first eigenvector |Z1(l,f)| corresponding to the wake-up audio data Z1(l,f) is obtained, and based on the wake-up audio data Z1(l,f) and the noise audio data Z2(l,f), the second eigenvector |X(l,f)| corresponding to the initial audio data is obtained, where X(l,f) is used to indicate the initial audio data indicated by the wake-up audio data Z1(l,f) and the noise audio data Z2(l,f), and the following calculation relationship exists:

[0063] X(l,f)=Z1(l,f)+Z2(l,f) (1)

[0064]

[0065] It should be noted that Vad(l,f) is the anti-noise parameter corresponding to the smart device, and the value of Vad(l,f) is between 0 and 1; |Z T (l,f)| used to indicate Z T The eigenvector corresponding to (l,f), Z T (l,f) is used to indicate wake-up audio data, T=1 or 2.

[0066] It should be noted that, in this embodiment, in the case of the wake-up audio data Z1 (l, f) and the noise audio data Z2 (l, f), T=1, Z T (l,f) is Z1(l,f).

[0067] Through the embodiments provided by the present application, the proportional coefficient is determined as the anti-noise parameter corresponding to each smart device; based on the anti-noise parameter, the distance value between each smart device and the audio source that emits the initial audio is determined. The anti-noise parameter determined by the first eigenvector and the second eigenvector is used to adjust and determine the distance value between each smart device and the audio source that emits the initial audio, thereby achieving the technical effect of improving the accuracy of obtaining the distance value between the smart device and the audio source that emits the initial audio, thereby achieving the technical effect of improving the accuracy of device wake-up.

[0068] As an optional solution, based on the anti-noise parameter, a distance value between each smart device and an audio sound source emitting the initial audio is determined, including:

[0069] S1, obtaining the diffusion ratio corresponding to the initial data received by each smart device;

[0070] S2, determining a direct-mix ratio corresponding to each smart device based on the anti-noise parameter and the diffusion ratio, wherein the direct-mix ratio is used to indicate a distance value between each smart device and an audio sound source that emits an initial audio.

[0071] Optionally, in this embodiment, the direct-mix ratio corresponding to the smart device is used to indicate the distance value between the smart device and the audio source that emits the initial audio, and the direct-mix ratio corresponding to the smart device is inversely proportional to the above distance value. The larger the direct-mix ratio value corresponding to the smart device, the smaller the above distance value; the smaller the direct-mix ratio value corresponding to the smart device, the larger the above distance value.

[0072] Optionally, in this embodiment, the diffusion ratio corresponding to the initial data received by the smart device is obtained. Wherein, l is used to indicate the time information of the received audio data, and f is used to indicate the frequency information of the received audio data.

[0073] Optionally, in this embodiment, based on the anti-noise parameter Vad(l, f) and the diffusion ratio To determine the direct-mixed ratio DRR corresponding to the smart device, there is the following calculation formula:

[0074]

[0075]

[0076] Wherein, lb is used to indicate the first frame corresponding to the current audio data, and lt is used to indicate the last frame corresponding to the current audio data. Used to indicate the direct-mixed ratio corresponding to the current frame.

[0077] Through the embodiment provided by the present application, the diffusion ratio corresponding to the initial data received by each smart device is obtained; based on the anti-noise parameter and the diffusion ratio, the direct mixing ratio corresponding to each smart device is determined, wherein the direct mixing ratio is used to indicate the distance value between each smart device and the audio sound source that emits the initial audio. The direct mixing ratio corresponding to each device determined by the anti-noise parameter and the diffusion ratio reflects the above distance value, thereby achieving the technical effect of improving the accuracy of obtaining the above distance value, thereby achieving the accuracy of device wake-up.

[0078] As an optional solution, the diffusion ratio corresponding to the initial data received by each smart device is obtained, including:

[0079] S1, obtain the coherence function and scattered noise field corresponding to the initial data;

[0080] S2, integrating the coherence function and the scattering noise field to obtain a diffusion ratio corresponding to the initial data.

[0081] Optionally, in this embodiment, before obtaining the coherence function and scattered noise field corresponding to the initial data, the frequency domain data corresponding to each microphone in the smart device is obtained. Taking the microphone array in the smart device uniformly using two microphones as an example, the received signals are represented by X(t) and Y(t). The cross-spectrum and self-spectrum operations are performed on the microphone frequency domain data X(l,f) and Y(l,f) of each smart device, and the calculation formula is as follows:

[0082] P xy (l,f)=αP xy (l,f)+(1-α)X(l,f)Y * (l,f)(5)

[0083] P x (l,f)=αP x (l,f)+(1-α)X(l,f)X * (l,f) (6)

[0084] P y (l,f)=αP y (l,f)+(1-α)Y(l,f)Y * (l,f) (7)

[0085] Wherein, X(l,f) is used to indicate the audio data collected by the first microphone, X * (l,f) is used to indicate the conjugate data of the audio data collected by the first microphone; Y(l,f) is used to indicate the audio data collected by the second microphone, * (l,f) is used to indicate the conjugate data of the audio data collected by the second microphone; the above-mentioned two microphones include a first microphone and a second microphone; l is used to indicate the time information of the received audio data, and f is used to indicate the frequency information of the received audio data; a is a smoothing factor, and in order to reflect the instantaneous nature of the direct-to-mix ratio change, 0.5 is used in the embodiment of the present application.

[0086] Optionally, in this embodiment, the coherence function and the scattered noise field corresponding to the initial data are obtained, and there is the following calculation formula:

[0087]

[0088] R n (f) = sinc(2fd / c)(9)

[0089] Among them, R x (l,f) is a coherence function corresponding to the initial data, which can be used, but is not limited to, to measure the linear correlation between the audio data collected by the first microphone and the audio data collected by the second microphone; R n(f) is the scattered noise field corresponding to the initial data, sinc is the sinc function, d is the distance between the first microphone and the second microphone, and c is the speed of sound; l is used to indicate the time information of the received audio data, and f is used to indicate the frequency information of the received audio data.

[0090] It should be noted that R x (l,f) is a relative value. If the two microphone signals are amplified by the same multiple at the same time, it will not affect R x value.

[0091] Optionally, in this embodiment, based on the coherence function and the scattered noise field, the diffusion ratio corresponding to the initial data is determined, including the following calculation formula:

[0092]

[0093] in, is the diffusion ratio corresponding to the initial data, Re and {} mean taking the real part for the next calculation.

[0094] Through the embodiments provided by the present application, based on the initial data received by each smart device, the coherence function and the scattering noise field corresponding to the initial data are obtained; the coherence function and the scattering noise field are integrated and processed to obtain the diffusion ratio corresponding to the initial data.

[0095] As an optional solution, the received data of each smart device is subjected to sound source separation processing to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device, including:

[0096] S1, determining preset environments matching at least two smart devices to be awakened;

[0097] S2, obtaining an audio recognition model corresponding to a preset environment, wherein the audio recognition model is a neural network model for recognizing audio obtained by training using audio samples of the preset environment;

[0098] S3, inputting the received data of each smart device into the audio recognition model in sequence to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device.

[0099] Optionally, in this embodiment, the preset environment matched by at least two smart devices to be awakened may be, but is not limited to, used to indicate noise data under the preset environment.

[0100] Optionally, in this embodiment, the audio recognition model can be obtained by training based on, but not limited to, open source data and noise data under a preset environment, wherein the open source data can include, but is not limited to, DNS2021 and MUSAN.

[0101] Optionally, in this embodiment, the audio recognition model may be, but is not limited to, a neural network model for recognizing audio obtained by training using audio samples of a preset environment, wherein the audio samples of the preset environment may be, but is not limited to, noise data under the preset environment.

[0102] It should be noted that the received data of each smart device is input into the audio recognition model in sequence to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device.

[0103] Through the embodiments provided by the present application, the preset environments that match at least two smart devices to be awakened are determined; the audio recognition model corresponding to the preset environment is obtained, wherein the audio recognition model is a neural network model for recognizing audio obtained by training with audio samples of the preset environment; the received data of each smart device is input into the audio recognition model in sequence to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device. The audio recognition model obtained by training with the pre-matched preset environment identifies and outputs the wake-up audio data and noise audio data corresponding to the smart device, thereby achieving the purpose of improving the acquisition of the wake-up audio data and noise audio data, thereby improving the accuracy of the sound source separation processing, reducing the influence of external noise on the device wake-up, and achieving the technical effect of improving the accuracy of the device wake-up.

[0104] As an optional solution, before determining the preset environments matching at least two smart devices to be awakened, the method further includes:

[0105] S1, obtaining audio samples of multiple preset environments, wherein the audio samples of the preset environments carry marked noise audio data in the preset environments;

[0106] S2, using audio samples of multiple preset environments to train the initial audio recognition model until a trained audio recognition model is obtained.

[0107] It should be noted that audio samples of multiple preset environments are obtained, wherein the audio samples of the preset environments carry labeled noise audio data in the preset environments; and the initial audio recognition model is trained using the audio samples of the multiple preset environments until a trained audio recognition model is obtained.

[0108] Through the embodiments provided by the present application, audio samples of multiple preset environments are obtained, wherein the audio samples of the preset environments carry marked noise audio data in the preset environments; the initial audio recognition model is trained using the audio samples of multiple preset environments until a trained audio recognition model is obtained. By training the audio recognition model, not only a wide range of noises are covered, but also the impact noise can be estimated based on the output wake-up audio data and the noise audio data, thereby reducing the impact of external noise on device wake-up, thereby achieving the technical effect of improving the accuracy of device wake-up.

[0109] As an optional solution, based on the above device wake-up method, a method for determining the anti-noise parameter Vad(l, f) is described as follows:

[0110] It should be noted that for impact noises with a long duration and large spectrum fluctuations, such as knocking on the table, closing the door, clapping, etc., the CDR estimation will be too large, and the result of obtaining the distance value between the smart device and the audio sound source that emits the initial audio will have the defect of being too large. Traditional signal processing methods cannot accurately estimate the power spectrum of this type of noise. The present invention provides a Vad(l,f) determination model based on deep learning to improve noise resistance.

[0111] Model features: The noisy signal is passed through stft to obtain the corresponding frequency domain signal. The frequency domain signal is a complex number. The bark feature is selected to amplify the low frequency and compress the high frequency. The bark domain feature can be understood as amplitude information. This embodiment takes 32 dimensions as an example, and the frequency range of the mapping is selected to be 200-3200Hz.

[0112] Datasets: Open source DNS2021 and MUSAN, as well as noise collected from various home scenes.

[0113] Model structure: The first layer is a one-dimensional convolutional layer conv1d: in_channels = 32, out_channels = 32, kernel_size = 4, stride = 1, padding = 1, kernel_size = 4, which is convenient for optimization with arm_neon on the mobile terminal; further is the LSTM layer: Input_size = 32, hidden_size = 32, num_layers = 1; further is the linear layer Input_size = 32, out_size = 32 and the activation layer sigmoid; Adam is selected for model training; its parameter quantity is 13632.

[0114] Model output: b represents the bark frequency band, and Y is the clean signal. After the bark is converted to the mask value of 200-3200Hz in the frequency domain, Vad(l, f) can be obtained. Since the last layer is sigmoid, the value range of Vad(l, f) is also [0, 1].

[0115] It should be noted that the above-mentioned deep learning-based Vad(l, f) determination model can be used to estimate impact noise and also cover more types of noise.

[0116] It should be noted that the above-mentioned Vad(l, f) determination model based on deep learning uses an extremely small model structure, reduces memory and computing power, and is sufficient to run on smart devices.

[0117] As an optional solution, the above device wake-up method is applied to the application scenario of sound source separation to reduce the wake-up interference of external noise on smart devices in a home environment. The specific steps are as follows:

[0118] Step S402, based on the received data of the initial data corresponding to each group of microphones, a coherence function and a scattered noise field are calculated;

[0119] Step S404, determining a diffusion ratio corresponding to the smart device based on the coherence function and the scattered noise field;

[0120] Step S406, using aux-iva to perform sound source separation processing on the received data of the initial data, and obtaining wake-up audio data and noise audio data, wherein the confidence corresponding to the wake-up audio data is greater than the confidence corresponding to the noise audio data;

[0121] Step S408, determining an anti-noise parameter corresponding to the smart device based on the wake-up audio data and the noise audio data;

[0122] Step S410, determining the direct-mixed ratio of the smart device in the current frame based on the diffusion ratio and anti-noise parameter corresponding to the smart device;

[0123] Step S412, based on the direct-to-mix ratio of the smart device in the current frame, determining the direct-to-mix ratio corresponding to the smart device in the wake-up word stage;

[0124] It should be noted that the calculation method of the above-mentioned coherence function, scattered noise field, diffusion ratio, wake-up audio data and noise audio data, anti-noise parameters, direct mixing ratio of the current frame, and direct mixing ratio corresponding to the smart device can refer to the example shown in the above-mentioned device wake-up method, and will not be repeated here in this example.

[0125] It is understandable that in the specific implementation of this application, related data such as user information is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0126] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0127] According to another aspect of the embodiments of the present application, a device awakening apparatus for implementing the above device awakening method is also provided. Figure 5 As shown, the device comprises:

[0128] The acquiring unit 502 is configured to, when the initial audio is acquired, determine at least two smart devices to be awakened that have received the initial audio, and acquire reception data of the initial audio by each of the at least two smart devices to be awakened;

[0129] The sound source separation unit 504 is used to perform sound source separation processing on the received data of each smart device when the received data is audio data for waking up the smart device, so as to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device;

[0130] A calculating unit 506 is used to obtain the audio data sum of each pair of wake-up audio data and noise audio data, and calculate a proportional coefficient of each wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the anti-noise capability corresponding to each smart device;

[0131] The awakening unit 508 is used to calculate the distance value between each smart device and the audio sound source that emits the initial audio based on the proportional coefficient, and determine the smart device with the smallest distance value from at least two smart devices to be awakened to wake up.

[0132] As an optional solution, the above-mentioned sound source separation unit 504 includes: a first acquisition module, used to acquire the first audio data and the second audio data obtained by performing sound source separation processing on the received data of each smart device; a calculation module, used to calculate the first confidence corresponding to the first audio data and the second confidence corresponding to the second audio data; a first determination module, used to compare the first confidence and the second confidence, and determine the audio data with high confidence as wake-up audio data, and determine the audio data with low confidence as noise audio data.

[0133] As an optional solution, the wake-up unit 508 includes: a second determination module, used to determine the proportional coefficient as the anti-noise parameter corresponding to each smart device; and a third determination module, used to determine the distance value between each smart device and the audio source that emits the initial audio based on the anti-noise parameter.

[0134] As an optional solution, the above-mentioned third determination module includes: an acquisition submodule, used to obtain the diffusion ratio corresponding to the initial data received by each smart device; a determination submodule, used to determine the direct mixing ratio corresponding to each smart device based on the anti-noise parameter and the diffusion ratio, wherein the direct mixing ratio is used to indicate the distance value between each smart device and the audio sound source that emits the initial audio.

[0135] As an optional solution, the acquisition submodule includes: an acquisition subunit, used to acquire the coherence function and the scattering noise field corresponding to the initial data; and a determination subunit, used to integrate the coherence function and the scattering noise field to obtain the diffusion ratio corresponding to the initial data.

[0136] As an optional solution, the sound source separation unit 504 includes: a fourth determination module, used to determine a preset environment matching at least two smart devices to be awakened; a second acquisition module, used to acquire an audio recognition model corresponding to the preset environment, wherein the audio recognition model is a neural network model for recognizing audio obtained by training using audio samples of the preset environment; a fifth determination module, used to input the received data of each smart device into the audio recognition model in turn to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device.

[0137] As an optional solution, the above-mentioned device also includes: a third acquisition module, used to obtain audio samples of multiple preset environments before determining the preset environments that match at least two smart devices to be awakened, wherein the audio samples of the preset environments carry marked noise audio data in the preset environments; a training module, used to train the initial audio recognition model using audio samples of multiple preset environments before determining the preset environments that match at least two smart devices to be awakened, until a trained audio recognition model is obtained.

[0138] For specific embodiments, reference may be made to the examples shown in the above device wake-up method, which will not be described in detail in this example.

[0139] According to another aspect of the embodiment of the present application, an electronic device for implementing the above-mentioned device wake-up method is also provided. Figure 6 As shown, the electronic device includes a memory 602 and a processor 604. The memory 602 stores a computer program. The processor 804 is configured to execute the steps in any of the above method embodiments through the computer program.

[0140] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0141] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0142] S1, when the initial audio is obtained, determining at least two smart devices to be awakened that have received the initial audio, and obtaining reception data of the initial audio by each of the at least two smart devices to be awakened;

[0143] S2, when the received data is audio data for waking up the smart device, performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device;

[0144] S3, obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating the proportional coefficient of each wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the corresponding anti-noise capability of each smart device;

[0145] S4, calculating the distance value between each smart device and the audio source that emits the initial audio based on the proportional coefficient, and determining the smart device with the smallest distance value from at least two smart devices to be awakened to awaken.

[0146] Alternatively, a person skilled in the art may understand that: Figure 6 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, or other terminal devices. Figure 6 The structure of the electronic device is not limited. Figure 6 more or fewer components such as network interfaces, etc.) as shown in the Figure 6 Different configurations shown.

[0147] Among them, the memory 602 can be used to store software programs and modules, such as the program instructions / modules corresponding to the device wake-up method and apparatus in the embodiments of the present application. The processor 604 executes various functional applications and data processing by running the software programs and modules stored in the memory 602, that is, to implement the above-mentioned device wake-up method. The memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 602 may further include a memory remotely located relative to the processor 604, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Among them, the memory 602 may specifically be used, but is not limited to, to store information such as initial audio reception data, wake-up audio data, noise audio data, etc. As an example, such as Figure 6 As shown, the memory 602 may include but is not limited to the acquisition unit 502, the sound source separation unit 504, the calculation unit 506 and the awakening unit 508 in the device awakening device. In addition, it may also include but is not limited to other module units in the device awakening device, which will not be repeated in this example.

[0148] Optionally, the transmission device 606 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wired network and a wireless network. In one example, the transmission device 606 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices and routers via a network cable so as to communicate with the Internet or a local area network. In one example, the transmission device 806 is a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0149] In addition, the electronic device further includes: a display 608 for displaying information such as the initial audio reception data, wake-up audio data, noise audio data, etc.; and a connection bus 610 for connecting various module components in the electronic device.

[0150] In other embodiments, the terminal device or server may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting the multiple nodes through network communication. Among them, the nodes may form a peer-to-peer (P2P) network, and any form of computing device, such as a server, terminal or other electronic device, may become a node in the blockchain system by joining the peer-to-peer network.

[0151] According to one aspect of the present application, a computer program product is provided, the computer program product comprising a computer program / instruction, the computer program / instruction comprising a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit, various functions provided by the embodiments of the present application are executed.

[0152] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0153] It should be noted that the computer system of the electronic device is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0154] The computer system includes a central processing unit (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) or the program loaded from the storage part to the random access memory (RAM). In the random access memory, various programs and data required for system operation are also stored. The central processing unit, the read-only memory and the random access memory are connected to each other through a bus. The input / output interface (I / O interface) is also connected to the bus.

[0155] The following components are connected to the input / output interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a local area network card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the input / output interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed.

[0156] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit, various functions defined in the system of the present application are executed.

[0157] According to one aspect of the present application, a computer-readable storage medium is provided, and a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the above-mentioned various optional implementations.

[0158] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0159] S1, when the initial audio is obtained, determining at least two smart devices to be awakened that have received the initial audio, and obtaining reception data of the initial audio by each of the at least two smart devices to be awakened;

[0160] S2, when the received data is audio data for waking up the smart device, performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device;

[0161] S3, obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating the proportional coefficient of each wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the corresponding anti-noise capability of each smart device;

[0162] S4, calculating the distance value between each smart device and the audio source that emits the initial audio based on the proportional coefficient, and determining the smart device with the smallest distance value from at least two smart devices to be awakened to awaken.

[0163] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0164] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0165] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0166] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0167] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0170] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A device wake-up method, characterized in that: include: In the case of acquiring the initial audio, determining at least two smart devices to be awakened that have received the initial audio, and acquiring reception data of the initial audio by each of the at least two smart devices to be awakened; In the case where the received data is audio data for waking up the smart device, performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device; Obtaining the audio data sum of each pair of wake-up audio data and noise audio data, and calculating a proportional coefficient of each of the wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the anti-noise capability corresponding to each of the smart devices; Determine the proportionality coefficient as the anti-noise parameter corresponding to each smart device; Obtaining a diffusion ratio corresponding to the initial audio received by each smart device; Determine a direct-mix ratio corresponding to each smart device based on the anti-noise parameter and the diffusion ratio, wherein the direct-mix ratio is used to indicate a distance value between each smart device and an audio sound source that emits the initial audio; The smart device with the smallest distance value is determined from the at least two smart devices to be awakened to be awakened.

2. The method according to claim 1, characterized in that The performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device includes: Acquire the first audio data and the second audio data obtained by performing the sound source separation process on the received data of each smart device; Calculating a first confidence level corresponding to the first audio data and a second confidence level corresponding to the second audio data; The first confidence level and the second confidence level are compared, and the audio data with a higher confidence level is determined as the wake-up audio data, and the audio data with a lower confidence level is determined as the noise audio data.

3. The method according to claim 1, characterized in that The obtaining of the diffusion ratio corresponding to the initial audio received by each smart device includes: Acquiring a coherence function and a scattered noise field corresponding to the initial audio; The coherence function and the scattered noise field are integrated to obtain the diffusion ratio corresponding to the initial audio.

4. The method according to claim 1, characterized in that: The performing sound source separation processing on the received data of each smart device to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device includes: Determining a preset environment matching the at least two smart devices to be awakened; Acquire an audio recognition model corresponding to the preset environment, wherein the audio recognition model is a neural network model for recognizing audio obtained by training using audio samples of the preset environment; The received data of each smart device is input into the audio recognition model in sequence to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device.

5. The method according to claim 4, characterized in that Before determining the preset environments matching the at least two smart devices to be awakened, the method further includes: Acquire a plurality of audio samples of the preset environment, wherein the audio samples of the preset environment carry marked noise audio data in the preset environment; The initial audio recognition model is trained using the audio samples of the plurality of preset environments until a trained audio recognition model is obtained.

6. A device waking up a device, characterized in that: include: an acquiring unit, configured to, when the initial audio is acquired, determine at least two smart devices to be awakened that have received the initial audio, and acquire reception data of the initial audio by each of the at least two smart devices to be awakened; a sound source separation unit, configured to perform sound source separation processing on the received data of each smart device when the received data is audio data for waking up the smart device, so as to obtain each pair of wake-up audio data and noise audio data corresponding to each smart device; a calculation unit, configured to obtain the audio data sum of each pair of wake-up audio data and noise audio data, and calculate a proportional coefficient of each of the wake-up audio data in the corresponding audio data sum, wherein the proportional coefficient is used to represent the anti-noise capability corresponding to each of the smart devices; The device is further used to determine the proportional coefficient as the anti-noise parameter corresponding to each smart device; obtain the diffusion ratio corresponding to the initial audio received by each smart device; and determine the direct mixing ratio corresponding to each smart device based on the anti-noise parameter and the diffusion ratio, wherein the direct mixing ratio is used to indicate the distance value between each smart device and the audio sound source that emits the initial audio; The awakening unit is used to determine the smart device with the smallest distance value from the at least two smart devices to be awakened and awaken it.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 5 when executed.

8. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 5 through the computer program.