Accurate voice wake-up control method and system
By using distributed networking and cloud platform collaborative processing, the system identifies nearby devices and performs voice recognition and command matching, solving the problems of false wake-up and recognition errors of smart home appliances, and improving the collaborative work between devices and the user experience.
Patent Information
- Application Number
- CN202411030355.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-26
AI Technical Summary
Existing smart home appliance voice wake-up technology is prone to false wake-ups and device recognition errors, especially in multi-device environments where it is difficult to accurately identify the device the user intends to wake up.
By using distributed networking to identify the nearest device, the main device and cloud platform are used to compare the signal-to-noise ratio and perform voice recognition to ensure that only the nearest device is woken up and listens for user commands, while other devices are in standby mode. The cloud platform performs semantic parsing and action command matching to ensure that commands are accurately transmitted to the target device.
It improves the collaborative working ability and user experience among smart home appliances, reduces false wake-ups and recognition errors, and enhances the accuracy and intelligence of voice control.
Smart Images

Figure CN119028360A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of smart home appliances, and more specifically, to an accurate voice wake-up control method and system. Background Art
[0002] With the development of technology, the technology of voice control of home appliances has been widely used in the field of smart home appliances, providing users with efficient and convenient human-computer interaction experience. Existing smart home appliances need to be awakened before collecting audio data. However, there may be the following usage defects in the actual use of the nearby awakening technology: 1. When the awakening position is close to the distance and direction of the two devices, it will inevitably cause false awakening; 2. The user does not intend to wake up the device closest to him. If the awakening position remains unchanged, the distant device may not be awakened; 3. Based on the current cross-quality control solution, users need to shout out relatively long control commands (such as "lower the wind speed" needs to be shouted as "lower the wind speed of the air conditioner in the living room"), and longer control commands are more likely to cause misrecognition. Summary of the invention
[0003] The present invention overcomes the disadvantage of easy false awakening in the prior art, and provides an accurate voice awakening control method and system, in the hope of solving the problems in the prior art.
[0004] In order to solve the above technical problems, one aspect of the present invention provides an accurate voice wake-up control method:
[0005] An accurate voice wake-up control method comprises the following steps:
[0006] The user speaks the wake-up content, and multiple smart home appliances receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio;
[0007] The multiple smart home appliances send the signal-to-noise ratio and identification code of the wake-up content audio to a master device in the smart home appliances;
[0008] The main device compares the signal-to-noise ratios of multiple smart home appliances and determines the device with a higher signal-to-noise ratio as the nearest device;
[0009] The master device sends a wake-up command to the nearby device, and the nearby device is awakened to listen to the control command issued by the user. The master device sends a standby command to other devices, and the other devices remain in the standby state.
[0010] Multiple smart home appliances are distributedly networked. After the networking is completed, a master device is designated. The signal-to-noise ratio of the master device is compared with the signal-to-noise ratio sent by the slave device on the master device to determine which device is the nearest device. If the master device is the nearest device, the master device enters the wake-up state to listen to the control command, and the slave device remains in the waiting state to be awakened; if the slave device is the nearest device, the master device sends a voice wake-up mark to the slave device through networking communication to let the slave device enter the wake-up state to listen to the control command, and the master device remains in the waiting state to be awakened.
[0011] A further technical solution is that the identification code is an SN code.
[0012] A further technical solution is that the nearby device monitors the control command issued by the user and then executes the following steps:
[0013] The nearby device sends the control command issued by the user to the cloud platform, and the cloud platform uses voice recognition technology to recognize the audio data of the control command as text information;
[0014] The cloud platform performs semantic analysis on the text information and parses out action instructions;
[0015] The cloud platform compares the parsed action instructions with the action instruction library of the smart home appliance. After finding a matching action instruction in the action instruction library, the action instruction is sent to the matching smart home appliance.
[0016] The matching smart home appliances execute the matching action instructions.
[0017] Since the nearest device is not necessarily the device the user actually wants to operate, when the nearest wake-up occurs by mistake, the user can continue to speak the control command. The mistakenly awakened device will send the control command audio to the cloud for identification and analysis. The cloud will send the analysis result to the device corresponding to the skill hit by the control command to realize voice control.
[0018] A further technical solution is that, while the nearby device sends the control command issued by the user to the cloud platform, the main device sends all smart home appliance identification codes that receive the wake-up content to the cloud platform;
[0019] The action instruction library of the smart home appliance is specifically an action instruction library of all smart home appliances that receive wake-up content.
[0020] If the action command library covers too many home appliances, even if the action command library is matched, it will be impossible to determine the device that the user actually wants to control because the action commands of multiple smart home appliances are the same. By limiting the action command library to the action command library of smart home appliances that receive the wake-up content, the range of devices that the user wants to control can be narrowed down, reducing the probability of misoperation.
[0021] A further technical solution is that after the cloud platform compares the parsed action instructions with the contents of the action instruction library of the smart home appliances, there are multiple smart home appliances in the action instruction library that match the action instructions;
[0022] The specific steps of sending the action command to the matching smart home appliance are as follows:
[0023] The cloud platform sends the action instructions and matching smart appliance information to the main device;
[0024] After receiving the action command and the matching smart home appliance information, the main device sends the action command to the device with the highest audio signal-to-noise ratio of the wake-up content among the matching smart home appliances.
[0025] If there are still multiple devices that match the action command among the devices that receive the wake-up content, the matching device closest to the user issuing the control command will be selected for control.
[0026] A further technical solution is that when the cloud platform compares the parsed action instructions with the content of the action instruction library of the smart home appliance, the parsed action instructions are first compared with the content of the action instruction library of the nearest device.
[0027] Another aspect of the present invention further provides an accurate voice wake-up control system, comprising a smart home appliance, wherein the smart home appliance comprises a main device and other devices;
[0028] The master device is used to receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio; receive the signal-to-noise ratio and identification code sent by other devices; compare the audio signal-to-noise ratio of the master device with the signal-to-noise ratio sent by other devices, determine the device with a higher signal-to-noise ratio as a nearby device, and send a wake-up instruction to the nearby device;
[0029] The other device is used to receive the wake-up content audio, calculate the signal-to-noise ratio of the wake-up content audio, and send the signal-to-noise ratio and the identification code to the main device.
[0030] A further technical solution is that the voice wake-up control system further includes a cloud platform;
[0031] The cloud platform is used to receive user control instructions issued by a nearby device, recognize the audio data of the control instructions as text information; perform semantic analysis on the text information to parse out action instructions;
[0032] The cloud platform is used to compare the parsed action instructions with the contents of the action instruction library of the smart home appliance. After finding a matching action instruction in the action instruction library, the action instruction is sent to the matching smart home appliance.
[0033] Compared with the prior art, the present invention has at least the following beneficial effects: the present invention networks multiple smart home appliances, designates a master device after the networking is completed, compares the signal-to-noise ratio of the master device with the signal-to-noise ratio sent from the slave device on the master device, and determines which device is the nearest device. The wake-up and waiting-to-be-awakened states of the device are determined by determining the nearest device, and further identification and matching are performed through the cloud platform to improve the control effect of the control command. The intelligence level and user experience of the smart home system are further improved, and the coordination and intelligent identification capabilities between devices are strengthened. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a schematic diagram of steps S1-S4 in Example 1;
[0035] Figure 2 This is a schematic diagram of steps S5-S8 in Example 1;
[0036] Figure 3 This is a schematic diagram of the action instruction library content comparison. DETAILED DESCRIPTION
[0037] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0038] Embodiment 1
[0039] An accurate voice wake-up control method, see Figure 1 , including the following steps:
[0040] S1: The user speaks the wake-up content, and multiple smart home appliances receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio;
[0041] In this embodiment, the plurality of smart home appliances include device 1 (a television) and device 2 (an air conditioner);
[0042] Voice collection is the device collecting the user's wake-up audio and control command audio through the microphone. The device calculates the signal-to-noise ratio (SNR) of the wake-up audio through voice signal processing and voice signal feature extraction (SNR refers to the power ratio between the voice signal and the noise signal, which is used to measure the clarity and recognizability of the voice signal).
[0043] Among them, all devices 1 (TV) and device 2 (air conditioner) are distributedly networked through MQTT technology, and each device can communicate with each other.
[0044] Device 1 (TV) and device 2 (air conditioner) in the distributed network use the same wake-up command.
[0045] S2: multiple smart home appliances send the signal-to-noise ratio and identification code of the wake-up content audio to a master device in the smart home appliances. In this embodiment, the master device is device 1 (TV), and the identification code is an SN code;
[0046] S3: The master device compares the signal-to-noise ratios of multiple smart home appliances and determines the device with a higher signal-to-noise ratio as the nearest device;
[0047] In this embodiment, the signal-to-noise ratio of device 1 (television) is higher than that of device 2 (air conditioner), so device 1 (television) is regarded as the nearest device.
[0048] S4: The master device sends a wake-up command to the nearby device, and the nearby device is awakened to listen to the control command issued by the user. The master device sends a standby command to other devices, and the other devices remain in the standby state.
[0049] That is, device 1 (TV) is awakened, while device 2 (air conditioner) remains in the waiting state;
[0050] In fact, after completing this step, the problem of false wake-up has been solved.
[0051] For next steps, see Figure 2 .
[0052] S5: The nearby device sends the control command issued by the user to the cloud platform, and the cloud platform uses speech recognition technology to recognize the audio data of the control command as text information;
[0053] In this embodiment, distributed networking is achieved between all devices and the cloud through MQTT technology. The cloud acts as an MQTT proxy server and the device acts as a client, so that each device and the cloud can communicate with each other.
[0054] Device 1 (TV) is awakened and enters the state of monitoring control commands. The user shouts the control command, and the TV receives the control command audio information and sends the control command issued by the user to the cloud platform. The data format is similar to: "voice": "audio data in digital form";
[0055] In a preferred embodiment, the main device, i.e., the TV, sends all smart home appliance identification codes that receive the wake-up content to the cloud platform, and the data format is similar to: "sns": ["TV SN", "air conditioner SN"].
[0056] The cloud platform uses speech recognition technology to recognize the digital audio data as text information.
[0057] S6: The cloud platform performs semantic analysis on the text information to parse out action instructions;
[0058] The text information is subjected to NLP semantic analysis, and the NLP semantic analysis converts the text information into an action instruction representation that can be understood and operated by a device.
[0059] For example, "It's too cold, turn down the air conditioner speed", will be converted into: turn down the fan speed;
[0060] S7: The cloud platform compares the parsed action command with the content of the action command library of the smart home appliance, and after finding a matching action command in the action command library, sends the action command to the matching smart home appliance;
[0061] The action instruction library is the operable instructions (skill) of the devices in the distributed network, see Figure 3 For example, the operable instructions of device 1 (TV) include: brightness adjustment intent; channel adjustment intent, etc.;
[0062] In this embodiment, since the action instruction "lower the wind speed" is associated with the wind volume adjustment intent in the operable instruction of device 2 (air conditioner), the action instruction "lower the wind speed" matches device 2 (air conditioner), and the cloud platform sends the action instruction "lower the wind speed" to device 2 (air conditioner).
[0063] In a preferred embodiment, the action instruction library of the smart home appliance is specifically an action instruction library of all smart home appliances that receive the wake-up content.
[0064] That is, only the action instruction library of all smart home appliances that receive the wake-up content in the distributed network is matched;
[0065] After the matching is completed, the cloud platform sends the action instructions to the matching smart home appliances;
[0066] Further preferably, when the cloud platform compares the parsed action instructions with the content of the action instruction library of the smart home appliance, the parsed action instructions are first compared with the content of the action instruction library of the nearest device.
[0067] Considering that users usually get close to the device when controlling it, starting the comparison from the action instruction library of the nearest device can increase the comparison speed;
[0068] It should be noted that, considering that there may be multiple identical devices in a distributed network, when there are multiple smart home appliances matching the action command in the action command library;
[0069] The cloud platform sends the action instructions and matching smart appliance information to the main device;
[0070] For example, if there are two air conditioners, the action command and the matching information of the two air conditioners are sent to the master device;
[0071] After receiving the action command and the matching smart home appliance information, the main device sends the action command to the device with the highest audio signal-to-noise ratio of the wake-up content among the matching smart home appliances.
[0072] S8: The matched smart home appliance executes the matched action instruction.
[0073] Embodiment 2
[0074] An accurate voice wake-up control system, comprising a smart home appliance and a cloud platform, wherein the smart home appliance comprises a main device and other devices;
[0075] The master device is used to receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio; receive the signal-to-noise ratio and identification code sent by other devices; compare the audio signal-to-noise ratio of the master device with the signal-to-noise ratio sent by other devices, determine the device with a higher signal-to-noise ratio as a nearby device, and send a wake-up instruction to the nearby device;
[0076] The other device is used to receive the wake-up content audio, calculate the signal-to-noise ratio of the wake-up content audio, and send the signal-to-noise ratio and the identification code to the main device.
[0077] The cloud platform is used to receive user control instructions issued by a nearby device, recognize the audio data of the control instructions as text information; perform semantic analysis on the text information to parse out action instructions;
[0078] The cloud platform is used to compare the parsed action instructions with the contents of the action instruction library of the smart home appliance. After finding a matching action instruction in the action instruction library, the action instruction is sent to the matching smart home appliance.
[0079] Although the present invention is described herein with reference to the illustrative embodiments of the present invention, it should be understood that those skilled in the art can design many other modifications and implementations, which will fall within the scope and spirit of the principles disclosed in the present application. More specifically, within the scope disclosed in the present application, multiple variations and improvements can be made to the components and / or layout of the subject combination layout. In addition to the variations and improvements made to the components and / or layout, other uses will also be apparent to those skilled in the art.
Claims
1. An accurate voice wake-up control method, characterized in that: The following steps are involved: The user speaks the wake-up content, and multiple smart home appliances receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio; The multiple smart home appliances send the signal-to-noise ratio and identification code of the wake-up content audio to a master device in the smart home appliances; The main device compares the signal-to-noise ratios of multiple smart home appliances and determines the device with a higher signal-to-noise ratio as the nearest device; The master device sends a wake-up command to the nearby device, and the nearby device is awakened to listen to the control command issued by the user. The master device sends a standby command to other devices, and the other devices remain in the standby state.
2. The accurate voice wake-up control method according to claim 1, characterized in that: The identification code is an SN code.
3. The accurate voice wake-up control method according to claim 1, characterized in that: The nearby device listens to the control command issued by the user and performs the following steps: The nearby device sends the control command issued by the user to the cloud platform, and the cloud platform uses voice recognition technology to recognize the audio data of the control command as text information; The cloud platform performs semantic analysis on the text information and parses out action instructions; The cloud platform compares the parsed action instructions with the action instruction library of the smart home appliance. After finding a matching action instruction in the action instruction library, the action instruction is sent to the matching smart home appliance. The matching smart home appliances execute the matching action instructions.
4. The accurate voice wake-up control method according to claim 3, characterized in that: While the nearby device sends the control command issued by the user to the cloud platform, the main device sends all the smart home appliance identification codes received by the wake-up content to the cloud platform; The action instruction library of the smart home appliance is specifically an action instruction library of all smart home appliances that receive wake-up content.
5. The accurate voice wake-up control method according to claim 4, characterized in that: After the cloud platform compares the parsed action instructions with the contents of the action instruction library of the smart home appliances, there are multiple smart home appliances in the action instruction library that match the action instructions; The specific steps of sending the action command to the matching smart home appliance are as follows: The cloud platform sends the action instructions and matching smart appliance information to the main device; After receiving the action command and the matching smart home appliance information, the main device sends the action command to the device with the highest audio signal-to-noise ratio of the wake-up content among the matching smart home appliances.
6. The accurate voice wake-up control method according to claim 3, characterized in that: When the cloud platform compares the parsed action command with the content of the action command library of the smart home appliance, the parsed action command is first compared with the content of the action command library of the nearest device.
7. An accurate voice wake-up control system, comprising: a smart home appliance, wherein the smart home appliance comprises a main device and other devices; The master device is used to receive the wake-up content audio and calculate the signal-to-noise ratio of the wake-up content audio; receive the signal-to-noise ratio and identification code sent by other devices; compare the audio signal-to-noise ratio of the master device with the signal-to-noise ratio sent by other devices, determine the device with a higher signal-to-noise ratio as a nearby device, and send a wake-up instruction to the nearby device; The other device is used to receive the wake-up content audio, calculate the signal-to-noise ratio of the wake-up content audio, and send the signal-to-noise ratio and the identification code to the main device.
8. The accurate voice wake-up control system according to claim 7, characterized in that: Also includes, cloud platform; The cloud platform is used to receive user control instructions issued by a nearby device, recognize the audio data of the control instructions as text information; perform semantic analysis on the text information to parse out action instructions; The cloud platform is used to compare the parsed action instructions with the contents of the action instruction library of the smart home appliance. After finding a matching action instruction in the action instruction library, the action instruction is sent to the matching smart home appliance.
