Voice false wake-up processing method and device, equipment and storage medium

Abstract

The application provides a voice false wake-up processing method, device and equipment and a storage medium, and relates to the technical field of smart home. The method comprises the following steps: determining the response state of each smart device in a target space receiving a voice signal; determining a pre-wake-up signal of a smart device identified as a pre-wake-up state and a non-wake-up device identified as a non-wake-up state; determining that the voice signal in the target space is a valid wake-up voice based on the pre-wake-up signal and the non-wake-up device; and waking up a target smart device in the target space based on the valid wake-up voice. The application solves the defects of low voice recognition accuracy and high false wake-up voice frequency in the prior art, reduces the dependence on wake-up audio training in a specific space environment, and improves the voice recognition accuracy by determining the valid wake-up voice.

Description

Technical Field

[0001] This invention relates to the field of smart home technology, and in particular to a method, apparatus, device, and storage medium for handling voice-activated false wake-up. Background Technology

[0002] With the development of smart home appliance technology, more and more home electronic devices can achieve voice control functions.

[0003] Currently, when using voice control to control smart home devices, users can use voice commands to control the smart devices to perform related actions. For example, users can use the voice command "turn on the air conditioner" to make the air conditioner automatically turn on the heating or cooling function.

[0004] However, all smart devices suffer from false wake-up issues, affecting the accuracy of speech recognition. Related technologies typically improve speech recognition accuracy by training a wake-up recognition model to accurately identify wake words.

[0005] However, when the spatial environment of a smart device changes, the accuracy of the wake-up recognition model decreases, which in turn reduces the accuracy of voice recognition and leads to many false wake-up incidents, affecting the user experience. Summary of the Invention

[0006] This invention provides a method, apparatus, device, and storage medium for handling voice false wake-up, which addresses the shortcomings of low accuracy and high frequency of false wake-up voices in the prior art. It reduces the dependence on training wake-up audio in specific spatial environments and improves the accuracy of voice recognition by identifying effective wake-up voices.

[0007] In a first aspect, the present invention provides a method for handling voice-activated false wake-up, comprising the following steps.

[0008] Determine the response state of each smart device located in the target space upon receiving a voice signal; the target space is a space where at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state;

[0009] Determine the pre-wake signal of the smart device identified as being in the pre-wake state, and determine the unwake device identified as being in the unwake state;

[0010] Based on the pre-wake-up signal and the unwake-up device, the voice signal in the target space is determined to be a valid wake-up voice.

[0011] The target intelligent device in the target space is activated based on the effective wake-up voice.

[0012] Preferably, according to the voice false wake-up processing method provided by the present invention, before the step of determining the voice signal in the target space as a valid wake-up voice based on the pre-wake-up signal and the unwake-up device, the method includes:

[0013] The system counts the first number of smart devices identified as being in the pre-wake-up state, the second number of smart devices identified as being in the unwake-up state, and the target number of voice signals received by multiple smart devices located in the target space.

[0014] If the first quantity and the target quantity are equal, and the second quantity is empty, the voice signal in the target space is determined to be a valid wake-up voice.

[0015] Preferably, according to the voice false wake-up processing method provided by the present invention, the step of determining the voice signal in the target space as a valid wake-up voice based on the pre-wake-up signal and the unwake-up device includes:

[0016] The pre-wake signals of the smart devices identified as being in the pre-wake state are sent to the unwakeable devices identified as being in the unwakeable state, resulting in multiple wake-up verification results; the wake-up verification results include at least a valid wake-up result and an invalid wake-up result.

[0017] The number of valid wake-up results and the number of invalid wake-up results are counted.

[0018] Based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number, the speech signal in the target space is determined to be the valid wake-up speech.

[0019] Preferably, according to a voice false wake-up processing method provided by the present invention, determining the voice signal in the target space as the valid wake-up voice based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number includes:

[0020] The number of valid wake-up results and the first number are calculated to obtain the number of valid wake-ups;

[0021] The number of invalid wake-up results and the second number are calculated to obtain the number of invalid wake-ups;

[0022] The number of valid wake-ups is compared with a first preset threshold, and the number of invalid wake-ups is compared with a second preset threshold; the first preset threshold is greater than the second preset threshold.

[0023] If the number of valid wake-ups is greater than or equal to the first preset threshold and the number of invalid wake-ups is less than the second preset threshold, the voice signal in the target space is determined to be the valid wake-up voice.

[0024] Preferably, according to the voice false wake-up processing method provided by the present invention, after the step of comparing the number of invalid wake-ups with a second preset threshold, the method includes:

[0025] If the number of invalid wake-ups is greater than or equal to the second preset threshold, the voice signal in the target space is determined to be invalid wake-up voice.

[0026] Preferably, in the voice false wake-up processing method provided by the present invention, the step of determining the wake-up verification result includes:

[0027] The wake-up verification model of the unwake-up device is used to perform feature extraction processing on the pre-wake-up signal to obtain keywords corresponding to the pre-wake-up signal;

[0028] The keywords are matched with a preset wake-up word library to obtain keyword matching results;

[0029] The wake-up verification result is determined based on the keyword matching results.

[0030] Secondly, the present invention also provides a voice-activated false wake-up processing device, comprising:

[0031] A response state determination module is used to determine the response state of each smart device located in the target space when it receives a voice signal; the target space is a space in which at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state.

[0032] The module for identifying unwakeable devices is used to determine the pre-wake signal of a smart device identified as being in the pre-wake state, and to determine the unwakeable device identified as being in the unwake state.

[0033] A valid wake-up voice module is used to determine, based on the pre-wake-up signal and the unwake-up device, the voice signal in the target space as a valid wake-up voice;

[0034] The wake-up module is used to wake up the target smart device in the target space based on the effective wake-up voice.

[0035] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement any of the voice false wake-up processing methods described above.

[0036] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the voice false wake-up processing method as described above.

[0037] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the voice false wake-up processing method described above.

[0038] This invention provides a method, apparatus, device, and storage medium for handling voice false wake-up. It determines the response state of each smart device in a target space that receives a voice signal; the target space is a space where at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state; it determines the pre-wake-up signal of the smart device identified as being in the pre-wake-up state, and determines the non-wake-up device identified as being in the non-wake-up state; based on the pre-wake-up signal and the non-wake-up device, it determines the voice signal in the target space as a valid wake-up voice; and it wakes up the target smart device in the target space based on the valid wake-up voice. This addresses the shortcomings of existing technologies, such as low voice recognition accuracy and high frequency of false wake-up voices, reduces the reliance on training wake-up audio in a specific spatial environment, and improves voice recognition accuracy by determining valid wake-up voices. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0040] Figure 1 This is a flowchart illustrating the voice false wake-up handling method provided by the present invention.

[0041] Figure 2 This is a schematic diagram of the voice false wake-up processing device provided by the present invention.

[0042] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0044] In the relevant technologies, at least the following technical problems exist:

[0045] False wake-ups are a common problem with smart devices, affecting the accuracy of speech recognition. Related technologies typically improve speech recognition accuracy by training a wake-up recognition model to accurately identify wake words.

[0046] However, when the spatial environment of a smart device changes, the accuracy of the wake-up recognition model decreases, which in turn reduces the accuracy of voice recognition and leads to many false wake-up incidents, affecting the user experience.

[0047] The following is combined with Figures 1-3 This invention describes a method, apparatus, device, and storage medium for handling voice false wake-up, which addresses the shortcomings of low accuracy in speech recognition and high frequency of false wake-up voices in the prior art. It reduces the dependence on training wake-up audio in a specific spatial environment and improves the accuracy of speech recognition by identifying effective wake-up voices.

[0048] Figure 1 This is a flowchart illustrating a voice false wake-up handling method provided by the present invention, as shown below. Figure 1 As shown, the method may include, but is not limited to, steps S100 to S400:

[0049] S100, determine the response state of each smart device in the target space that receives a voice signal; the target space is a space in which at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state;

[0050] S200, determine the pre-wake signal of the smart device identified as the pre-wake state, and determine the unwake device identified as the unwake state;

[0051] S300, based on the pre-wake-up signal and the unwake-up device, determine that the voice signal in the target space is a valid wake-up voice;

[0052] S400, based on the effective wake-up voice, wake up the target smart device in the target space.

[0053] In step S100 of some embodiments, the response state of each smart device located in the target space that receives the voice signal is determined.

[0054] The target space is a space in which at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state.

[0055] It should be noted that the target space can also be a physical space, such as a room in a living room. This physical space houses multiple smart devices, which can connect via LAN, Bluetooth, or WiFi to form a communication network for transmitting and interacting data. Each smart device has voice recognition capabilities and is equipped with an independent voice wake-up verification model to verify the received voice signal.

[0056] Furthermore, when a voice signal is generated in the target space, the response state of each smart device in the target space upon receiving the voice signal is determined. When a smart device in the target space can accurately recognize the voice signal, it is updated to a pre-wake state.

[0057] When a smart device in the target space cannot accurately recognize a voice signal, it is updated to an unwakeable state.

[0058] Furthermore, it should be noted that smart devices include, but are not limited to, smart refrigerators, smart TVs, smart speakers, smart tablets, and smart air conditioners. Each smart device is not located in the same place; they are distributed throughout the target space, but multiple smart devices can communicate with each other.

[0059] When a voice signal is generated in the target space, such as "Xiao Lu, turn on the air conditioner".

[0060] Because the training level of the voice wake-up verification model, the accuracy of voice recognition, and the positional distance between the voice signal generator and the device that generates the voice signal vary from smart device to smart device, the voice signal received by each smart device may also be different. For example, a smart tablet may recognize "Xiao Lu, turn on the air conditioner" as "Xiao Lu, turn on the air conditioner".

[0061] However, the smart TV recognized "Xiao Lu, turn on the air conditioner," the smart air conditioner recognized "Xiao Lu, turn on the air conditioner," the smart speaker recognized "Xiao Lu, turn on the air conditioner," and the smart refrigerator did not recognize the corresponding voice signal or update its own status. Therefore, even when there is only one voice signal in a target space, the voice signals recognized by multiple smart devices are different. Thus, it is impossible to determine whether the voice signal recognized by the smart devices in a space is a valid wake-up voice or an invalid wake-up voice.

[0062] The pre-wake-up state can be indicated by the indicator light flashing, while the non-wake-up state is indicated by no response within a preset time, such as no indicator light flashing.

[0063] In step S200 of some embodiments, a pre-wake signal of a smart device identified as being in the pre-wake state is determined, and an unwake device identified as being in the unwake state is determined.

[0064] It is understandable that the pre-wake signal of a smart device identified as being in a pre-wake state is determined, such as "Xiao Lu, turn on the air conditioner" for a smart tablet, "Xiao Lu, turn on the air conditioner" for a smart TV, "Xiao Lu, turn on the air conditioner" for a smart air conditioner, and "Xiao Lu, turn on the air conditioner" for a smart speaker.

[0065] The unwake-up device identified as being in the unwake-up state is a smart refrigerator.

[0066] In some embodiments of the present invention, prior to the step of determining the speech signal in the target space as valid wake-up speech based on the pre-wake-up signal and the unwake-up device, the method includes:

[0067] The system counts the first number of smart devices identified as being in the pre-wake-up state, the second number of smart devices identified as being in the unwake-up state, and the target number of voice signals received by multiple smart devices located in the target space.

[0068] If the first quantity and the target quantity are equal, and the second quantity is empty, the voice signal in the target space is determined to be a valid wake-up voice.

[0069] It is understandable that if the first quantity is equal to the target quantity and the second quantity is empty, it means that the original voice signal in the target space can wake up all the smart devices in the target space. Therefore, the original voice signal in the target space is the effective wake-up voice.

[0070] In step S300 of some embodiments, the voice signal in the target space is determined to be a valid wake-up voice based on the pre-wake-up signal and the unwake-up device.

[0071] Further, determining the voice signal in the target space as a valid wake-up voice based on the pre-wake-up signal and the unwake-up device includes:

[0072] The pre-wake signals of the smart devices identified as being in the pre-wake state are sent to the unwakeable devices identified as being in the unwakeable state, resulting in multiple wake-up verification results.

[0073] It should be noted that the wake-up verification result includes at least a valid wake-up result and a invalid wake-up result;

[0074] The number of valid wake-up results and the number of invalid wake-up results are counted.

[0075] Based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number, the speech signal in the target space is determined to be the valid wake-up speech.

[0076] To further determine whether the original speech signal in the target space is a valid wake-up speech, cross-validation is used to determine whether the original speech signal in the target space is a valid wake-up speech.

[0077] The pre-wake signals of the smart devices identified as being in the pre-wake state are sent to the unwakeable devices identified as being in the unwakeable state, resulting in multiple wake-up verification results.

[0078] For example, the pre-wake signal of smart device 1 is M1, the pre-wake signal of smart device 2 is M2, and the pre-wake signal of smart device 3 is M3.

[0079] The unwake-up devices identified as being in the unwake-up state include at least N1, N2, and N3.

[0080] The pre-wake-up signal M1 of smart device 1, the pre-wake-up signal M2 of smart device 2, and the pre-wake-up signal M3 of smart device 3 are sent to the unwake-up device N1 respectively.

[0081] And send the pre-wake signal M1 of smart device 1, the pre-wake signal M2 of smart device 2, and the pre-wake signal M3 of smart device 3 to the unwakeable device N2 respectively.

[0082] And send the pre-wake signal M1 of smart device 1, the pre-wake signal M2 of smart device 2, and the pre-wake signal M3 of smart device 3 to the unwakeable device N3 respectively.

[0083] Taking the example of sending the pre-wake-up signal M1 of smart device 1, the pre-wake-up signal M2 of smart device 2, and the pre-wake-up signal M3 of smart device 3 to the unwakeable device N1, it can be seen that:

[0084] The wake-up verification model of the unwakeable device N1 is used to perform wake-up verification processing on the pre-wake-up signal M1 to obtain wake-up verification result P1. The wake-up verification model of the unwakeable device N1 is used to perform wake-up verification processing on the pre-wake-up signal M2 to obtain wake-up verification result P2. The wake-up verification model of the unwakeable device N1 is used to perform wake-up verification processing on the pre-wake-up signal M3 to obtain wake-up verification result P3.

[0085] For example, wake-up verification results P1 and P2 are valid wake-up results, while wake-up verification result P3 is an invalid wake-up result.

[0086] Then, count the number of valid wake-up results and the number of invalid wake-up results. Similarly, we can calculate the number of valid wake-up results and the number of invalid wake-up results determined using the unwakeable device N2, and the number of valid wake-up results and the number of invalid wake-up results determined using the unwakeable device N3.

[0087] Further, the step of determining the speech signal in the target space as the valid wake-up speech based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number specifically includes:

[0088] The number of valid wake-up results and the first number are calculated to obtain the number of valid wake-ups; the number of invalid wake-up results and the second number are calculated to obtain the number of invalid wake-ups.

[0089] The number of valid wake-ups is compared with a first preset threshold, and the number of invalid wake-ups is compared with a second preset threshold; the first preset threshold is greater than the second preset threshold.

[0090] If the number of valid wake-ups is greater than or equal to the first preset threshold and the number of invalid wake-ups is less than the second preset threshold, the voice signal in the target space is determined to be the valid wake-up voice.

[0091] The total number of valid wake-up results is obtained by adding the number of valid wake-up results determined using the unwakeable device N2, the number of valid wake-up results determined using the unwakeable device N3, the number of valid wake-up results determined using the unwakeable device N1, to the first number of smart devices identified as being in the pre-wake state during pre-wake-up.

[0092] Similarly, the total number of invalid wake-up results can be obtained by adding the number of invalid wake-up results determined by the unwakeable device N1, the number of invalid wake-up results determined by the unwakeable device N2, the number of invalid wake-up results determined by the unwakeable device N3, and the second number of smart devices identified as unwakeable during pre-wake-up.

[0093] Let the total number of valid wake-ups be K1, and the total number of invalid wake-ups be K2.

[0094] The first preset quantity threshold is A1, and the second preset quantity threshold is A2. The first preset quantity threshold is greater than the second preset quantity threshold, that is, A1 is greater than A2.

[0095] Therefore, when K1 is greater than or equal to A1 and K2 is less than A2, it means that the total number of effective wake-ups is much greater than the total number of invalid wake-ups. The invalid wake-ups may be caused by the low recognition rate of their own wake-up models. Therefore, based on the wake-up situation of most smart devices after recognizing the wake-up voice signal, it can be seen that the original voice signal in the target space is the effective wake-up voice, which can significantly improve the accuracy of the effective wake-up voice. Thus, the target smart device can be accurately woken up based on the accurately determined effective wake-up voice.

[0096] If the number of invalid wake-ups is greater than or equal to the second preset threshold, the voice signal in the target space is determined to be invalid wake-up voice.

[0097] Furthermore, if K2 is greater than or equal to A2, it is not necessary to compare the sizes of K1 and A1 to determine that there are a large number of invalid wake-ups. In this case, the original speech signal in the target space is invalid wake-up speech.

[0098] Furthermore, when K1 is less than A1 and K2 is less than A2, further cross-validation is required to determine whether the original speech signal in the target space is a valid wake-up speech.

[0099] The cross-validation steps are as follows: the pre-wake signals of the smart devices identified as being in the pre-wake state are sent to the unwakeable devices identified as being in the unwakeable state, resulting in multiple wake-up verification results; the wake-up verification results include at least valid wake-up results and invalid wake-up results; the number of valid wake-up results and the number of invalid wake-up results are counted; and the voice signal in the target space is determined to be the valid wake-up voice based on the number of valid wake-up results, the number of invalid wake-up results, the first count, and the second count. These steps will not be elaborated here.

[0100] In step S400 of some embodiments, the target smart device in the target space is woken up based on the effective wake-up voice.

[0101] For example, if the effective wake-up voice is "Turn on the air conditioner, heating mode 25 degrees Celsius", the accurately identified smart device, such as a smart speaker, will determine whether it is the target smart device, the air conditioner. If it determines that it is not the target smart device, it will broadcast the effective wake-up voice to the target smart device, the smart air conditioner, via the network. When the smart air conditioner receives the effective wake-up voice, it will automatically execute the operation corresponding to the effective wake-up voice, such as turning on the air conditioner and adjusting it to the corresponding heating mode, setting the temperature to 25 degrees Celsius. This can greatly improve the user experience and reduce the false wake-up rate of voice signals.

[0102] The embodiments provided by this invention do not rely on wake-up audio training in a specific spatial environment. Therefore, even if the spatial environment changes (such as room size, layout adjustments, etc.), it will not affect the accuracy of false wake-up detection. This is because the comparison of audio signals between different devices within the same space can reflect the true wake-up situation without being affected by environmental changes.

[0103] In some embodiments of the present invention, the step of determining the wake-up verification result includes:

[0104] The wake-up verification model of the unwake-up device is used to perform feature extraction processing on the pre-wake-up signal to obtain keywords corresponding to the pre-wake-up signal;

[0105] The keywords are matched with a preset wake-up word library to obtain keyword matching results;

[0106] The wake-up verification result is determined based on the keyword matching results.

[0107] Understandably, when the keyword matching result indicates that the keyword matches multiple preset wake-up words in the preset wake-up word library, this wake-up verification result is determined to be a valid wake-up result.

[0108] If the keyword matching result indicates that the keyword does not match any of the preset wake-up words in the preset wake-up word library, the wake-up verification result is determined to be an invalid wake-up result.

[0109] This invention provides a method, apparatus, device, and storage medium for handling voice false wake-up. It determines the response state of each smart device in a target space that receives a voice signal; the target space is a space where at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state; it determines the pre-wake-up signal of the smart device identified as being in the pre-wake-up state, and determines the non-wake-up device identified as being in the non-wake-up state; based on the pre-wake-up signal and the non-wake-up device, it determines the voice signal in the target space as a valid wake-up voice; and it wakes up the target smart device in the target space based on the valid wake-up voice. This addresses the shortcomings of existing technologies, such as low voice recognition accuracy and high frequency of false wake-up voices, reduces the reliance on training wake-up audio in a specific spatial environment, and improves voice recognition accuracy by determining valid wake-up voices.

[0110] The following describes a voice false wake-up processing device provided by the present invention. The voice false wake-up processing device described below can be referred to in correspondence with the voice false wake-up processing method described above.

[0111] like Figure 2The diagram shown is a structural schematic of the voice false wake-up processing device provided by the present invention. The voice false wake-up processing device includes the following modules:

[0112] The response state determination module 210 is used to determine the response state of each smart device located in the target space when it receives a voice signal; the target space is a space in which at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state.

[0113] The module 220 for determining unwakeable devices is used to determine the pre-wake signal of a smart device identified as being in the pre-wake state, and to determine the unwakeable device identified as being in the unwake state.

[0114] The valid wake-up voice module 230 is used to determine, based on the pre-wake-up signal and the unwake-up device, the voice signal in the target space as a valid wake-up voice;

[0115] The wake-up module 240 is used to wake up the target smart device in the target space based on the effective wake-up voice.

[0116] Preferably, the voice wake-up processing device provided by the present invention is specifically used to count the first number of smart devices identified as being in the pre-wake-up state, and to count the second number of smart devices identified as being in the unwake-up state, and to count the target number of voice signals received by multiple smart devices located in the target space.

[0117] If the first quantity and the target quantity are equal, and the second quantity is empty, the voice signal in the target space is determined to be a valid wake-up voice.

[0118] Preferably, the voice wake-up processing device provided by the present invention is specifically used to send the pre-wake-up signal of the smart device identified as being in the pre-wake-up state to the unwake-up device identified as being in the unwake-up state, thereby obtaining multiple wake-up verification results; the wake-up verification results include at least a valid wake-up result and a invalid wake-up result;

[0119] The number of valid wake-up results and the number of invalid wake-up results are counted.

[0120] Based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number, the speech signal in the target space is determined to be the valid wake-up speech.

[0121] Preferably, the voice wake-up processing device provided by the present invention is specifically used to calculate the number of valid wake-up results and the first number to obtain the number of valid wake-up results;

[0122] The number of invalid wake-up results and the second number are calculated to obtain the number of invalid wake-ups;

[0123] The number of valid wake-ups is compared with a first preset threshold, and the number of invalid wake-ups is compared with a second preset threshold; the first preset threshold is greater than the second preset threshold.

[0124] If the number of valid wake-ups is greater than or equal to the first preset threshold and the number of invalid wake-ups is less than the second preset threshold, the voice signal in the target space is determined to be the valid wake-up voice.

[0125] Preferably, the voice wake-up processing device provided by the present invention is specifically used to determine that the voice signal in the target space is an invalid wake-up voice when the number of invalid wake-ups is greater than or equal to the second preset number threshold.

[0126] Preferably, the voice wake-up processing device provided by the present invention is specifically used to perform feature extraction processing on the pre-wake-up signal using the wake-up verification model of the unwake-up device to obtain keywords corresponding to the pre-wake-up signal;

[0127] The keywords are matched with a preset wake-up word library to obtain keyword matching results;

[0128] The wake-up verification result is determined based on the keyword matching results.

[0129] This invention provides a method, apparatus, device, and storage medium for handling voice false wake-up. It determines the response state of each smart device in a target space that receives a voice signal; the target space is a space where at least two smart devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state; it determines the pre-wake-up signal of the smart device identified as being in the pre-wake-up state, and determines the non-wake-up device identified as being in the non-wake-up state; based on the pre-wake-up signal and the non-wake-up device, it determines the voice signal in the target space as a valid wake-up voice; and it wakes up the target smart device in the target space based on the valid wake-up voice. This addresses the shortcomings of existing technologies, such as low voice recognition accuracy and high frequency of false wake-up voices, reduces the reliance on training wake-up audio in a specific spatial environment, and improves voice recognition accuracy by determining valid wake-up voices.

[0130] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can invoke logical instructions in the memory 330 to execute a voice false wake-up handling method. This method includes: determining the response state of each smart device in a target space upon receiving a voice signal; the target space is a space consisting of at least two smart devices communicating with each other; the response state includes at least a pre-wake-up state and an unwake-up state; determining a pre-wake-up signal of a smart device identified as being in the pre-wake-up state, and determining an unwake-up device identified as being in the unwake-up state; based on the pre-wake-up signal and the unwake-up device, determining that the voice signal in the target space is a valid wake-up voice; and waking up the target smart device in the target space based on the valid wake-up voice.

[0131] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0132] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute a voice false wake-up processing method provided by the above methods. The method includes: determining the response state of each smart device located in a target space that receives a voice signal; the target space is a space for mutual communication composed of at least two smart devices; the response state includes at least a pre-wake-up state and a non-wake-up state; determining a pre-wake-up signal of a smart device identified as being in the pre-wake-up state, and determining a non-wake-up device identified as being in the non-wake-up state; determining, based on the pre-wake-up signal and the non-wake-up device, that the voice signal in the target space is a valid wake-up voice; and waking up the target smart device in the target space based on the valid wake-up voice.

[0133] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a voice false wake-up processing method provided by the methods described above. The method includes: determining the response state of each smart device located in a target space upon receiving a voice signal; the target space being a space consisting of at least two smart devices communicating with each other; the response state including at least a pre-wake-up state and a non-wake-up state; determining a pre-wake-up signal of a smart device identified as being in the pre-wake-up state, and determining a non-wake-up device identified as being in the non-wake-up state; determining, based on the pre-wake-up signal and the non-wake-up device, that the voice signal in the target space is a valid wake-up voice; and waking up the target smart device in the target space based on the valid wake-up voice.

[0134] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0135] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for handling voice-activated false wake-up, characterized in that, include: Determine the response status of each smart device located in the target space upon receiving a voice signal; The target space is a space in which at least two intelligent devices communicate with each other. The response state includes at least a pre-wake-up state and a non-wake-up state; Determine the pre-wake signal of the smart device identified as being in the pre-wake state, and determine the unwake device identified as being in the unwake state; The system counts the first number of smart devices identified as being in the pre-wake-up state, the second number of smart devices identified as being in the unwake-up state, and the target number of voice signals received by multiple smart devices located in the target space. When the first quantity and the target quantity are equal and the second quantity is empty, the voice signal in the target space is determined to be a valid wake-up voice. Based on the pre-wake-up signal and the unwake-up device, the voice signal in the target space is determined to be a valid wake-up voice; the determination of the voice signal in the target space as a valid wake-up voice based on the pre-wake-up signal and the unwake-up device includes: The pre-wake signals of the smart devices identified as being in the pre-wake state are sent to the unwakeable devices identified as being in the unwakeable state, resulting in multiple wake-up verification results; the wake-up verification results include at least a valid wake-up result and an invalid wake-up result. The number of valid wake-up results and the number of invalid wake-up results are counted. The speech signal in the target space is determined to be the valid wake-up speech based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number; the step of determining the speech signal in the target space as the valid wake-up speech based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number includes: The number of valid wake-up results and the first number are calculated to obtain the number of valid wake-ups; The number of invalid wake-up results and the second number are calculated to obtain the number of invalid wake-ups; The number of valid wake-ups is compared with a first preset threshold, and the number of invalid wake-ups is compared with a second preset threshold; the first preset threshold is greater than the second preset threshold. If the number of valid wake-ups is greater than or equal to the first preset threshold and the number of invalid wake-ups is less than the second preset threshold, the speech signal in the target space is determined to be the valid wake-up speech. The target intelligent device in the target space is activated based on the effective wake-up voice.

2. The voice false wake-up processing method according to claim 1, characterized in that, After the step of comparing the number of invalid wake-ups with a second preset threshold, the method includes: If the number of invalid wake-ups is greater than or equal to the second preset threshold, the voice signal in the target space is determined to be invalid wake-up voice.

3. The voice false wake-up processing method according to claim 1, characterized in that, The steps for determining the wake-up verification result include: The wake-up verification model of the unwake-up device is used to perform feature extraction processing on the pre-wake-up signal to obtain keywords corresponding to the pre-wake-up signal; The keywords are matched with a preset wake-up word library to obtain keyword matching results; The wake-up verification result is determined based on the keyword matching results.

4. A voice-activated false wake-up processing device, characterized in that, include: The response status determination module is used to determine the response status of each smart device located in the target space when it receives a voice signal; The target space is a space where at least two intelligent devices communicate with each other; the response state includes at least a pre-wake-up state and a non-wake-up state. The module for identifying unwakeable devices is used to determine the pre-wake signal of a smart device identified as being in the pre-wake state, and to determine the unwakeable device identified as being in the unwake state. A valid wake-up voice module is configured to determine that the voice signal in the target space is a valid wake-up voice based on the pre-wake-up signal and the unwakeable device. Before the step of determining that the voice signal in the target space is a valid wake-up voice based on the pre-wake-up signal and the unwakeable device, a first number of smart devices identified as being in the pre-wake-up state is counted, a second number of smart devices identified as being in the unwake-up state is counted, and a target number of voice signals received by multiple smart devices located in the target space is counted. If the first number and the target number are equal and the second number is null, the voice signal in the target space is determined to be a valid wake-up voice. The step of determining that the voice signal in the target space is a valid wake-up voice based on the pre-wake-up signal and the unwakeable device includes: sending the pre-wake-up signal of the smart device identified as being in the pre-wake-up state to the unwakeable device identified as being in the unwake-up state, obtaining multiple wake-up verification results; the wake-up verification results include at least a valid wake-up result and a invalid wake-up result; and counting the valid wake-up results. The process involves calculating the number of valid wake-up results and the number of invalid wake-up results; determining the speech signal in the target space as the valid wake-up speech based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number; and determining the speech signal in the target space as the valid wake-up speech based on the number of valid wake-up results, the number of invalid wake-up results, the first number, and the second number, which includes: calculating the number of valid wake-up results and the first number to obtain a valid wake-up count; calculating the number of invalid wake-up results and the second number to obtain an invalid wake-up count; comparing the valid wake-up count with a first preset number threshold and comparing the invalid wake-up count with a second preset number threshold; where the first preset number threshold is greater than the second preset number threshold; and where the valid wake-up count is greater than or equal to the first preset number threshold and the invalid wake-up count is less than the second preset number threshold, the speech signal in the target space is determined as the valid wake-up speech. The wake-up module is used to wake up the target smart device in the target space based on the effective wake-up voice.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the voice false wake-up processing method as described in any one of claims 1 to 3.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the voice false wake-up processing method as described in any one of claims 1 to 3.

7. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the voice false wake-up processing method as described in any one of claims 1 to 3.

Images