Method, device and server for determining a responding device

Abstract

The application discloses a determination method, device and server of a response device. The method comprises the following steps: after receiving a voice control signal, obtaining a target probability value of a device as a response device according to a preset probability value strategy, wherein the probability value strategy is determined based on response state information of the device to the voice control signal; determining a target signal-to-noise ratio according to the target probability value and a signal-to-noise ratio of the received voice control signal; and determining whether the device is the response device of the voice control signal according to the target signal-to-noise ratio. The technical scheme provided by the application aims to solve the technical problem that the actual response of the device in the existing voice control scene does not match the demand.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment, and more specifically to a method, device, and server for determining a response device. Background Technology

[0002] In a smart home system, when there are multiple devices with voice response capabilities in the same space, if there is a voice wake-up command, the signal-to-noise ratio (SNR) of each device is calculated, and the device with the highest SNR is selected for wake-up.

[0003] In practical applications, there is a situation where a user plans to wake up device A, but actually wakes up device B, resulting in a discrepancy between the actual response and the expected response. Summary of the Invention

[0004] The main objective of this invention is to provide a method, device, and server for determining a response device, aiming to solve the problem in the prior art where the actual response of the device in a voice control scenario does not match the requirements.

[0005] To achieve the above objectives, the present invention provides a method for determining a response device, comprising:

[0006] After receiving a voice control signal, the device is used to obtain a target probability value for responding according to a preset probability value strategy. The probability value strategy is determined based on the device's response status information to the voice control signal.

[0007] The target signal-to-noise ratio is determined based on the target probability value and the signal-to-noise ratio of the received voice control signal;

[0008] Based on the target signal-to-noise ratio, determine whether the device should act as a response device for the voice control signal.

[0009] Preferably, the probability value strategy includes a method for determining the probability value corresponding to each time interval;

[0010] The acquisition of the target probability value of the device as a response device includes:

[0011] Obtain the reception time of the voice control signal;

[0012] The target probability value is determined according to the determination method corresponding to the time interval in which the receiving time is located.

[0013] Preferably, the method further includes:

[0014] Calculate the similarity between keywords in the voice control signal and preset standard words to obtain confidence information;

[0015] If the confidence level information is less than a preset confidence threshold, it is not allowed to determine whether the device is a response device to the voice control signal, wherein the confidence threshold is determined based on the magnitude of the target probability value.

[0016] Preferably, the method further includes:

[0017] Record the number of times the device acts as a response device to obtain response status information;

[0018] Send response status information;

[0019] Receive update information for the probability value strategy corresponding to the response status information.

[0020] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in any of the preceding descriptions.

[0021] An apparatus comprising:

[0022] A memory configured to store computer programs;

[0023] A processor configured to execute the computer program to implement the method described in any of the preceding descriptions.

[0024] A method for determining a response device, comprising:

[0025] Acquire device response status information to voice control signals, wherein the response status information includes the number of times the device acts as a responding device;

[0026] The probability value strategy of the device is determined based on the number of occurrences in the response status information;

[0027] Send the probability value strategy.

[0028] Preferably, if the response status information includes the number of times the device acts as a response device within each time interval, then the probability value strategy includes the determination method corresponding to each time interval.

[0029] Preferably, if the response status information also includes validity information of the responding device each time, then the determination method for each time interval is determined based on the number of valid times Ti and / or the number of invalid times Fi within each time interval; where i is a positive integer representing the number of the time interval.

[0030] Preferably, the probability value of the i-th time interval is determined by one or more of the following methods:

[0031] If all validity information is valid, then the maximum and / or minimum number of valid counts are obtained from the valid counts of all time intervals. At least one of the maximum and minimum number of valid counts and the number of valid counts Ti are calculated to obtain the first value, which is used as the probability value of the i-th time interval.

[0032] If all valid information is invalid, then the maximum and / or minimum invalid number of invalid times are obtained from the invalid times of all time intervals. At least one of the maximum and minimum invalid number of invalid times and the invalid number Fi are calculated to obtain a second value. Based on the difference between the preset standard value and the second value, a third value is obtained as the probability value of the i-th time interval.

[0033] If part of the validity information is valid and the other part is invalid, then the first value corresponding to the number of valid times and the second value corresponding to the number of invalid times are determined respectively. Based on the difference between the first value and the second value, a fourth value is obtained as the probability value of the i-th time interval.

[0034] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the network distribution method described in any of the preceding claims.

[0035] A server, comprising:

[0036] A memory configured to store computer programs;

[0037] A processor configured to execute the computer program to implement the method described in any of the preceding descriptions.

[0038] In the technical solution of this invention, the response device is determined under the premise of conforming to the user's operating habits, so that the determination operation can better meet the user's needs and improve the matching degree between the determination result and the user's expectations. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0040] Figure 1 A flowchart of the method for determining a response device provided in Embodiment 1 of this application;

[0041] Figure 2 A flowchart of the method for determining a response device provided in Embodiment 2 of this application;

[0042] Figure 3 A flowchart of the method for determining a response device provided in Embodiment 3 of this application;

[0043] Figure 4 A flowchart of the method for determining a response device provided in Embodiment 4 of this application;

[0044] Figure 5 This is a flowchart of a method for determining a response device provided in Embodiment 5 of this application.

[0045] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0047] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0048] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0049] In this invention, unless otherwise explicitly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0050] Furthermore, the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are feasible for those skilled in the art. If the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0051] Example 1

[0052] Figure 1 This is a flowchart illustrating the method for determining a response device provided in Embodiment 1 of this application. Figure 1 As shown, the method includes:

[0053] Step 101: If a voice control signal is received, the target probability value for determining the device as a responding device is obtained according to the pre-set probability value strategy. This probability value strategy is obtained based on the device's response status information to the voice control signal.

[0054] In one exemplary embodiment, the device has a voice response function and can be a home appliance, such as an air conditioner, television, refrigerator, and smart speaker.

[0055] In one exemplary embodiment, the voice control command may be a wake-up command, a parameter setting command, or a power on / off command.

[0056] In one exemplary embodiment, the probability value strategy can be proactively issued to the device by the cloud platform; or, the device can request the strategy from the cloud platform, and the cloud platform can issue the strategy to the device based on the request.

[0057] In one exemplary embodiment, the probability values ​​in the probability-valued strategy satisfy the following rules, including:

[0058] The more times a device is identified as a responding device, the higher the likelihood that the device will continue to act as a responding device in the current environment, and the higher the probability value of the device. Conversely, the fewer times a device is identified as a responding device, the lower the likelihood that the device will continue to act as a responding device in the current environment, and the lower the probability value of the device.

[0059] Since the probability value strategy is generated based on response status information, it can accurately reflect the user's usage habits in identifying the device as the response device during historical use. Compared with the probability values ​​in the prior art that are only applicable to the current environment, this invention, by combining the user's operating habits reflected in the probability values ​​from historical operations, can better assist in the determination of the response device.

[0060] Step 102: Determine the target signal-to-noise ratio based on the target probability value and the signal-to-noise ratio of the pre-acquired voice control signal;

[0061] The signal-to-noise ratio (SNR) is the ratio of the energy value of the received voice control signal to the energy value of the noise signal. The existing methods for calculating the SNR are all applicable to this step and will not be elaborated here.

[0062] In one exemplary embodiment, the target probability value and the signal-to-noise ratio can be multiplied to obtain the result of the multiplication operation, which is then used as the target signal-to-noise ratio.

[0063] By correcting the signal-to-noise ratio (SNR) in the current environment using the target probability value, the resulting SNR not only reflects the magnitude of the SNR but also the user's operating habits.

[0064] Step 103: Based on the target signal-to-noise ratio, determine whether the device is a responding device for the voice control signal.

[0065] Compared with the prior art that directly uses the signal-to-noise ratio in the current environment, the solution provided by the embodiments of the present invention uses the target probability value and the target signal-to-noise ratio obtained from the current signal-to-noise ratio for judgment. Since the target signal-to-noise ratio can not only reflect the magnitude of the signal-to-noise ratio, but also reflect the user's operating habits, the response device determined according to the target signal-to-noise ratio is determined under the premise of conforming to the user's operating habits, so that the determination operation can better meet the user's needs and improve the matching degree between the determination result and the user's expectations.

[0066] The method provided in Embodiment 1 of this application, if a voice control signal is received, obtains a target probability value for the device to be identified as a responding device according to a pre-set probability value strategy, determines a target signal-to-noise ratio based on the target probability value and the signal-to-noise ratio of the received voice control signal, and determines whether the device is a responding device based on the target signal-to-noise ratio. The method determines the responding device under the premise of conforming to the user's operating habits, so that the determination operation can better meet the user's needs and improve the matching degree between the determination result and the user's expectations.

[0067] Example 2

[0068] Figure 2 This is a flowchart illustrating the method for determining a response device provided in Embodiment 2 of this application. Figure 2 As shown, the method includes:

[0069] Step 201: If a voice control signal is received, the target probability value for determining the device as a responding device is obtained according to the pre-set probability value strategy. This probability value strategy is obtained based on the device's response status information to the voice control signal.

[0070] In one exemplary embodiment, the device has a voice response function and can be a home appliance, such as an air conditioner, television, refrigerator, and smart speaker.

[0071] In one exemplary embodiment, the voice control command may be a wake-up command, a parameter setting command, or a power on / off command.

[0072] In one exemplary embodiment, the probability value strategy can be proactively issued to the device by the cloud platform; or, the device can request the strategy from the cloud platform, and the cloud platform can issue the strategy to the device based on the request.

[0073] In one exemplary embodiment, the probability values ​​in the probability-valued strategy satisfy the following rules, including:

[0074] The more times a device is identified as a responding device, the higher the likelihood that the device will continue to act as a responding device in the current environment, and the higher the probability value of the device. Conversely, the fewer times a device is identified as a responding device, the lower the likelihood that the device will continue to act as a responding device in the current environment, and the lower the probability value of the device.

[0075] Since the probability value strategy is generated based on response status information, it can accurately reflect the user's usage habits in identifying the device as the response device during historical use. Compared with the probability values ​​in the prior art that are only applicable to the current environment, this invention, by combining the user's operating habits reflected in the probability values ​​from historical operations, can better assist in the determination of the response device.

[0076] In one exemplary embodiment, the probability value strategy can be the method by which the device determines the probability value corresponding to each time interval;

[0077] When determining the target probability value of a device being identified as a responding device, the time when the device receives the voice control signal can be obtained, and the target probability value can be obtained according to the determination method used for the target time interval corresponding to that time.

[0078] The duration of the time interval can be set as needed. Taking 1 hour as an example, if the receiving time is 11:30, the target time interval is 11:00 to 12:00. The target probability value is obtained according to the determination method corresponding to this target time interval.

[0079] By setting the probability value strategy for devices, including the determination method of different time intervals, the probability information of the same device acting as a response device in different time intervals can be reflected. This can more accurately reflect the user's operating habits of the device at different times, making the target probability value more accurate and providing support for more accurate determination of the response device.

[0080] Step 202: Calculate the similarity between the keywords and the pre-set standard words to obtain confidence information, where the keywords come from the voice control signal;

[0081] Taking voice control signals as wake-up commands as an example, the confidence level information can be obtained by calculating the similarity between the keywords in the voice control signals and the standard words corresponding to the wake-up commands.

[0082] Step 203: Determine whether the confidence level information is greater than the preset confidence level threshold;

[0083] The confidence threshold is determined based on the magnitude of the target probability value. The higher the target probability value, the higher the probability that the device is identified as a response device, and the lower the confidence value. Conversely, the lower the target probability value, the lower the probability that the device is identified as a response device, and the higher the confidence value.

[0084] Taking voice control signals as wake-up commands as an example, the higher the target probability value, the higher the likelihood of the device being woken up. To ensure the device can be woken up smoothly in different environments, the confidence level is set to be low to improve the success rate of waking up the device. Conversely, the lower the target probability value, the lower the likelihood of the device being woken up. To reduce the likelihood of the device being woken up by incorrect environmental sounds, the confidence level is set to be high to appropriately control the success rate of waking up the device.

[0085] As can be seen from the above examples, setting a confidence threshold based on the target probability value can support the determination of whether a device is a responding device.

[0086] If the confidence level is greater than the confidence threshold, proceed to step 204; otherwise, the process ends.

[0087] Step 204: Determine the target signal-to-noise ratio based on the target probability value and the signal-to-noise ratio of the pre-acquired voice control signal;

[0088] The signal-to-noise ratio (SNR) is the ratio of the energy value of the received voice control signal to the energy value of the noise signal. The existing methods for calculating the SNR are all applicable to this step and will not be elaborated here.

[0089] In one exemplary embodiment, the target probability value and the signal-to-noise ratio can be multiplied to obtain the result of the multiplication operation, which is then used as the target signal-to-noise ratio.

[0090] By correcting the signal-to-noise ratio (SNR) in the current environment using the target probability value, the resulting SNR not only reflects the magnitude of the SNR but also the user's operating habits.

[0091] Step 205: Based on the target signal-to-noise ratio, determine whether the device is a responding device for the voice control signal.

[0092] Compared with the prior art that directly uses the signal-to-noise ratio in the current environment, the solution provided by the embodiments of the present invention uses the target probability value and the target signal-to-noise ratio obtained from the current signal-to-noise ratio for judgment. Since the target signal-to-noise ratio can not only reflect the magnitude of the signal-to-noise ratio, but also reflect the user's operating habits, the response device determined according to the target signal-to-noise ratio is determined under the premise of conforming to the user's operating habits, so that the determination operation can better meet the user's needs and improve the matching degree between the determination result and the user's expectations.

[0093] The method provided in Embodiment 2 of this application, upon receiving a voice control signal, obtains a target probability value for the device to be identified as a responding device according to a pre-set probability value strategy. Based on the target probability value and the signal-to-noise ratio (SNR) of the received voice control signal, a target SNR is determined. Based on the target SNR, the method determines whether the device is a responding device for the voice control signal. This method identifies the responding device while adhering to user operating habits, making the determination operation more aligned with user needs and improving the match between the determination result and user expectations. By setting a probability value strategy for the device that includes different time intervals, the method can more accurately reflect the user's operating habits on the device at different times, making the target probability value more accurate and supporting more accurate identification of the responding device. Setting a confidence threshold based on the target probability value further supports the determination that the device is a responding device.

[0094] Example 3

[0095] Figure 3 This is a flowchart illustrating the method for determining a response device provided in Embodiment 3 of this application. Figure 3 As shown, the method includes:

[0096] Step 301: If a voice control signal is received, the target probability value for determining the device as a responding device is obtained according to the pre-set probability value strategy. This probability value strategy is obtained based on the device's response status information to the voice control signal.

[0097] In one exemplary embodiment, the device has a voice response function and can be a home appliance, such as an air conditioner, television, refrigerator, and smart speaker.

[0098] In one exemplary embodiment, the voice control command may be a wake-up command, a parameter setting command, or a power on / off command.

[0099] In one exemplary embodiment, the probability value strategy can be proactively issued to the device by the cloud platform; or, the device can request the strategy from the cloud platform, and the cloud platform can issue the strategy to the device based on the request.

[0100] In one exemplary embodiment, the probability values ​​in the probability-valued strategy satisfy the following rules, including:

[0101] The more times a device is identified as a responding device, the higher the likelihood that the device will continue to act as a responding device in the current environment, and the higher the probability value of the device. Conversely, the fewer times a device is identified as a responding device, the lower the likelihood that the device will continue to act as a responding device in the current environment, and the lower the probability value of the device.

[0102] Since the probability value strategy is generated based on response status information, it can accurately reflect the user's usage habits in identifying the device as the response device during historical use. Compared with the probability values ​​in the prior art that are only applicable to the current environment, this invention, by combining the user's operating habits reflected in the probability values ​​from historical operations, can better assist in the determination of the response device.

[0103] In one exemplary embodiment, the probability value strategy can be the method by which the device determines the probability value corresponding to each time interval;

[0104] When determining the target probability value of a device being identified as a responding device, the reception time of the voice control signal can be obtained, and the target probability value can be obtained according to the determination method corresponding to the target time interval in which the reception time is located.

[0105] The duration of the time interval can be set as needed. Taking 1 hour as an example, if the receiving time is 11:30, the target time interval is 11:00 to 12:00. The target probability value is obtained according to the determination method corresponding to this target time interval.

[0106] By setting the probability value strategy for devices, including the determination method of different time intervals, the probability information of the same device acting as a response device in different time intervals can be reflected. This can more accurately reflect the user's operating habits of the device at different times, making the target probability rate value more accurate and providing support for more accurate determination of the response device.

[0107] Step 302: Calculate the similarity between the keywords and the pre-set standard words to obtain confidence information, where the keywords come from the voice control signal;

[0108] Taking voice control signals as wake-up commands as an example, the confidence level information can be obtained by calculating the similarity between the keywords in the voice control signals and the standard words corresponding to the wake-up commands.

[0109] Step 303: Determine whether the confidence level information is greater than the preset confidence level threshold;

[0110] The confidence threshold is determined based on the magnitude of the target probability value. The higher the target probability value, the higher the probability that the device is identified as a response device, and the lower the confidence value. Conversely, the lower the target probability value, the lower the probability that the device is identified as a response device, and the higher the confidence value.

[0111] Taking voice control signals as wake-up commands as an example, the higher the target probability value, the higher the likelihood of the device being woken up. To ensure the device can be woken up smoothly in different environments, the confidence level is set to be low to improve the success rate of waking up the device. Conversely, the lower the target probability value, the lower the likelihood of the device being woken up. To reduce the likelihood of the device being woken up by incorrect environmental sounds, the confidence level is set to be high to appropriately control the success rate of waking up the device.

[0112] If the confidence level is greater than the confidence threshold, proceed to step 303; otherwise, the process ends.

[0113] Step 304: Determine the target signal-to-noise ratio based on the target probability value and the signal-to-noise ratio of the pre-acquired voice control signal;

[0114] The signal-to-noise ratio (SNR) is the ratio of the energy value of the received voice control signal to the energy value of the noise signal. The existing methods for calculating the SNR are all applicable to this step and will not be elaborated here.

[0115] In one exemplary embodiment, the target probability value and the signal-to-noise ratio can be multiplied to obtain the result of the multiplication operation, which is then used as the target signal-to-noise ratio.

[0116] By correcting the signal-to-noise ratio (SNR) in the current environment using the target probability value, the resulting SNR not only reflects the magnitude of the SNR but also the user's operating habits.

[0117] Step 305: Based on the target signal-to-noise ratio, determine whether the device is a responding device for the voice control signal.

[0118] Compared with the prior art that directly uses the signal-to-noise ratio in the current environment, the solution provided by the embodiments of the present invention uses the target probability value and the target signal-to-noise ratio obtained from the current signal-to-noise ratio for judgment. Since the target signal-to-noise ratio can not only reflect the magnitude of the signal-to-noise ratio, but also reflect the user's operating habits, the response device determined according to the target signal-to-noise ratio is determined under the premise of conforming to the user's operating habits, so that the determination operation can better meet the user's needs and improve the matching degree between the determination result and the user's expectations.

[0119] Step 306: Record the response status information, which includes the number of times the device has been identified as a responding device;

[0120] Step 307: Send response status information;

[0121] The triggering condition for the above sending operation can be that the sending is initiated after the recording time reaches a preset time threshold, or that the sending is initiated after receiving a request from the cloud platform.

[0122] Step 308: Receive the update information of the probability value strategy corresponding to the response status information;

[0123] By receiving updates to the probability value strategy, the strategy can be updated in a timely manner, better matching the user's operating habits and further improving the match between the determined result and the user's expectations.

[0124] The method provided in this application, upon receiving a voice control signal, obtains a target probability value for the device to be identified as a responding device according to a pre-set probability value strategy. Based on the target probability value and the signal-to-noise ratio (SNR) of the received voice control signal, a target SNR is determined. Based on the target SNR, the method determines whether the device is a responding device for the voice control signal. This method determines the responding device while conforming to user operating habits, making the determination operation more aligned with user needs and improving the matching degree between the determination result and user expectations. By setting the device probability value strategy, including different time interval determination methods, it can more accurately reflect the user's operating habits on the device at different times, making the target probability value more accurate and supporting more accurate identification of the responding device. Setting a confidence threshold based on the target probability value can support the determination that the device is identified as a responding device. By updating the probability value strategy, it can better match user operating habits, further improving the matching degree between the determination result and user expectations.

[0125] Example 4

[0126] Figure 4 This is a flowchart illustrating the method for determining a response device provided in Embodiment 4 of this application. Figure 4 As shown, the method includes:

[0127] Step 401: Obtain the device's response status information, which records the number of times the device has responded to the voice control signal;

[0128] The triggering condition for the above acquisition operation can be that the device actively sends the request after the recording time reaches a preset time threshold, or it can be received after the cloud platform sends an acquisition request.

[0129] Step 402: Determine the probability value strategy of the device based on the frequency information in the response status information;

[0130] In one exemplary embodiment, the probability values ​​in the probability-valued strategy satisfy the following rules, including:

[0131] The more times a device is identified as a responding device, the higher the likelihood that the device will continue to act as a responding device in the current environment, and the higher the probability value of the device. Conversely, the fewer times a device is identified as a responding device, the lower the likelihood that the device will continue to act as a responding device in the current environment, and the lower the probability value of the device.

[0132] Since the probability value strategy is generated based on response status information, it can accurately reflect the user's usage habits in identifying the device as the response device during historical use. Compared with the probability values ​​in the prior art that are only applicable to the current environment, this invention, by combining the user's operating habits reflected in the probability values ​​from historical operations, can better assist in the determination of the response device.

[0133] Step 403: Send probability value strategy.

[0134] By sending a probability-based strategy that reflects the user's operating habits to the device, the device can determine whether to respond to voice control signals based on the probability-based strategy.

[0135] The method provided in Embodiment 4 of this application obtains response status information from the historical data of the device, determines the probability value strategy of the device based on the frequency information in the response status information, and sends the probability value strategy to provide data support for determining whether the operation can meet the user's needs.

[0136] In the above embodiment four, the probability value strategy can be determined by the following means, including:

[0137] Application Scenario 1:

[0138] The response status information includes the number of times the device was identified as a responding device in each time interval;

[0139] By utilizing the number of times a device is identified as a responding device in each time interval, the probability value of a device being identified as a responding device in each time interval can be determined, thereby obtaining a probability value strategy for the device.

[0140] By setting the probability value strategy for devices, including the determination method of different time intervals, the probability information of the same device acting as a response device in different time intervals can be reflected. This can more accurately reflect the user's operating habits of the device at different times, making the target probability value more accurate and providing support for more accurate determination of the response device.

[0141] Application Scenario 2

[0142] Response status information includes:

[0143] Information on the number of times a device is identified as a responding device in each time interval;

[0144] The device provides validity information for each response.

[0145] Once a device is identified as a responding device, if the user confirms that the device is the correct responding device, then the validity information is valid.

[0146] If the device is determined to be a false wake-up caused by ambient sound, or if the user confirms that the device is an incorrect responding device, then the validity information is invalid.

[0147] If a device is identified as a response device, it is considered a correct response device if it is the same as the object controlled by the user's requirements; otherwise, it is considered an incorrect response device.

[0148] Based on the above response status information, the number of valid times Ti and / or the number of invalid times Fi that a device is identified as a responding device can be determined for each time interval, where i represents the time interval number and is a positive integer; then, based on the number of valid times Ti and / or the number of invalid times Fi corresponding to each time interval, the method for determining the probability value corresponding to each time interval is determined, thereby obtaining the probability value calculation strategy.

[0149] The probability value corresponding to the i-th time interval is determined in the following way, including:

[0150] 1) All validity information in the i-th time interval is valid;

[0151] Count the number of times the device responds to the device in each time interval, select the number of times with the largest value to obtain the maximum number of times Tmax; and / or select the number of times with the smallest value to obtain the minimum number of times Tmin;

[0152] The first value is obtained by performing calculations on at least one of the minimum effective count Tmin and the maximum effective count Tmax, and the effective count Ti, and is used as the probability value corresponding to the i-th time interval.

[0153] In the above method, since at least one of the minimum effective count Tmin and the maximum effective count Tmax can be used, the first value can be determined in three ways, including:

[0154] Method 1: Only use the maximum number of valid counts, Tmax

[0155] Calculate the ratio of the number of valid counts Ti to the maximum number of valid counts Tmax, and use this ratio as the first value;

[0156] Method 2: Use only the minimum effective count, Tmin;

[0157] First, calculate the difference between the effective number of times Ti and the minimum effective number of times Tmin to obtain the first calculated value. Then, calculate the ratio between the first calculated value and the minimum effective number of times Tmin.

[0158] Method 3: Use the maximum number of valid counts Tmax and the minimum number of valid counts Tmin

[0159] The difference between the effective number of times Ti and the minimum effective number of times Tmin is calculated to obtain the first calculated value; and the difference between the maximum effective number of times Tmax and the minimum effective number of times Tmin is calculated to obtain the second calculated value.

[0160] The first numerical value can be obtained by calculating the first and second calculated values.

[0161] The first value can be determined by calculating the first calculated value and the second calculated value.

[0162] 2) All validity information in the i-th time interval is invalid;

[0163] Count the number of invalid responses from the device in each time interval, select the invalid response with the largest value to obtain the maximum invalid response value Fmax; and / or select the invalid response with the smallest value to obtain the minimum invalid response value Fmin;

[0164] The second value is obtained by performing a calculation operation on at least one of the minimum invalid number Fmin and the maximum invalid number Fmax, and the invalid number Fi, which is used as the probability value corresponding to the i-th time interval.

[0165] In the above method, since at least one of the minimum invalid count Fmin and the maximum invalid count Fmax can be used, the second value can be determined in three ways, including:

[0166] Method 1: Only use the maximum invalid count Fmax

[0167] Calculate the ratio between the number of invalid instances Fi and the maximum number of invalid instances Fmax, and use this ratio as the first value;

[0168] Method 2: Use only the minimum number of invalid attempts, Fmin;

[0169] First, calculate the difference between the number of invalid occurrences Fi and the minimum number of invalid occurrences Fmin to obtain the third calculated value. Then, calculate the ratio between the first calculated value and the minimum number of invalid occurrences Fmin.

[0170] Method 3: Use the maximum number of invalid attempts Fmax and the minimum number of invalid attempts Fmin

[0171] The difference between the number of invalid occurrences Fi and the minimum number of invalid occurrences Fmin is calculated to obtain the third calculated value; and the difference between the maximum number of invalid occurrences Fmax and the minimum number of invalid occurrences Fmin is calculated to obtain the fourth calculated value.

[0172] By calculating the third and fourth calculated values, the second value can be obtained;

[0173] The second value can be determined by calculating the third and fourth calculated values.

[0174] 3) In the i-th time interval, part of the validity information is valid and the other part is invalid.

[0175] The method described above is used to obtain the first value corresponding to valid validity information and the second value corresponding to invalid validity information;

[0176] The fourth value is calculated using the difference between the first and second values, and is used as the probability value of the device at the i-th time interval.

[0177] The probability values ​​determined using the above method are more accurate and thus better suited to users' operating habits.

[0178] Example 5

[0179] Figure 5 A flowchart of the method for determining a response device provided in Embodiment 5 of this application is shown below. Figure 5 As shown, Figure 5 The method shown takes voice control signal as wake-up command as an example. The method includes:

[0180] Step 501: After the device has been connected to the server for one week, the server generates a wake-up threshold adjustment table (corresponding to the probability value strategy above) based on the device's usage records over the week and sends it to the device.

[0181] The wake-up threshold adjustment table is obtained through the following methods:

[0182] Taking a time interval of 1 hour as an example, the number of effective wake-ups (T) and the number of invalid wake-ups (F) of the device within each hour are counted, and the maximum effective wake-up (Tmax), the maximum invalid wake-up (Fmax), the minimum effective wake-up (Tmin), and the minimum invalid wake-up (Fmin) are obtained from all time intervals.

[0183] Among them, a valid wake-up means that a valid intention is recognized within 1 minute before and after the wake-up; an invalid wake-up means that an invalid wake-up includes a false wake-up caused by environmental sound and a situation where the user wants to wake up other devices but is misjudged as waking up the current device. The judgment condition is that there is no valid intention within 1 minute before and after the wake-up.

[0184] Among them, the single-hour wake-up threshold condition percentage P = (T - Tmin) / (Tmax - Tmin) – (F - Fmin) / (Fmax - Fmin) (0% < P < 100%). The content of the adjustment table is the P value for each hour.

[0185] Among them, the adjustable range of the threshold is the threshold range corresponding to the wake-up rate of 80% to 95%.

[0186] Step 502: After the device receives the wake-up threshold adjustment table, set the wake-up threshold according to the preset adjustable range of the threshold and combine the percentage in the adjustment table, and calculate the threshold once per hour;

[0187] The wake-up confidence indicates that the wake-up model will output the wake-up confidence after processing the input audio, and this value is the similarity between the audio and the wake-up word.

[0188] Wake-up threshold: It is a fixed value obtained through test verification. Among them, if the wake-up confidence is higher than this value, it is determined that the wake-up word is hit.

[0189] Step 503: When the user wakes up the device, calculate the signal-to-noise ratio SNR value, multiply SNR by the current percentage in the wake-up threshold adjustment table to obtain the snr value that conforms to the user's usage habit;

[0190] Step 504: When multiple devices are woken up simultaneously, compare the values calculated in step 503 and select the device with the largest value to be woken up;

[0191] Step 505: The server regularly updates the wake-up threshold adjustment table.

[0192] The method provided in the fifth embodiment of this application can improve the success rate of the user waking up the target device by statistically analyzing the user's usage habits per hour and adjusting the wake-up threshold and the SNR value after wake-up. Moreover, when the user uses it less frequently, reducing the wake-up threshold can reduce the probability of false wake-up by environmental sound.

[0193] The embodiment of the present invention provides a storage medium storing a computer program, which can execute any of the above methods when the computer program is run. Figures 1 to 3 any of the methods

[0194] The embodiment of the present invention provides a storage medium storing a computer program, which can execute the method Figure 4 above

[0195] This invention provides a device including a memory and a processor. The memory stores a computer program, and the processor, when running the computer program in the memory, is capable of executing the above-mentioned... Figures 1 to 3 Any method.

[0196] The processor in the aforementioned device can be set as a separate module in the device, or integrated into the device's processor; it is used to determine whether the device is a responsive device when a voice control signal is received.

[0197] This invention provides a server, including a memory and a processor. The memory stores a computer program, and the processor, when running the computer program in the memory, is capable of executing the above-mentioned... Figure 4 The method.

[0198] The processor in the aforementioned server can be set up as an independent module in the device, or integrated into the processor of the server; it is used to generate probability value strategies for the device so that the device can determine the response device based on the probability value strategies.

[0199] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural transformations made under the concept of the present invention using the description and drawings of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A method for determining a response device, characterized in that, include: After receiving a voice control signal, the device is used to obtain a target probability value for responding according to a preset probability value strategy. The probability value strategy is determined based on the device's response status information to the voice control signal. The similarity between keywords in the voice control signal and pre-set standard words is calculated to obtain confidence information. If the confidence information is greater than a preset confidence threshold, the target probability value and the signal-to-noise ratio of the received voice control signal are multiplied together to obtain the target signal-to-noise ratio. Based on the target signal-to-noise ratio, determine whether the device should act as a response device for the voice control signal; The confidence threshold is determined based on the target probability value. The higher the target probability value, the lower the confidence threshold; conversely, the lower the target probability value, the higher the confidence threshold. The probability value strategy includes the method for determining the probability value corresponding to each time interval. The acquisition of the target probability value of the device as a response device includes: Obtain the reception time of the voice control signal; The target probability value is determined according to the determination method corresponding to the time interval in which the receiving time is located; The probability value strategy is generated by the server based on the device's usage records during the preset duration after the device has been connected to the server for a preset duration. The expression for calculating the probability value Pi of the i-th time interval in the probability value strategy is as follows: Pi=(Ti–Tmin) / (Tmax–Tmin)–(Fi–Fmin) / (Fmax–Fmin); in: Ti represents the number of effective wake-ups of the device within the i-th time interval; Fi represents the number of invalid wake-ups of the device within the i-th time interval; Tmax represents the maximum number of valid wake-ups in each time interval across all time intervals; Fmax represents the maximum number of invalid wake-ups in each time interval across all time intervals; Tmin represents the minimum number of valid wake-ups in each time interval across all time intervals; Fmin represents the minimum number of invalid wake-ups in each time interval across all time intervals; Where i is a positive integer.

2. The method as described in claim 1, characterized in that, The method further includes: Record the number of times the device is woken up as a responding device to obtain response status information; Send response status information; Receive update information for the probability value strategy corresponding to the response status information.

3. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in claim 1 or 2.

4. A device, characterized in that, include: A memory configured to store computer programs; A processor configured to execute the computer program to implement the method as claimed in claim 1 or 2.

Images