Voice wake-up method, device and equipment

Abstract

The invention provides a voice wake-up method, device and equipment, and belongs to the field of data processing, and the method comprises the steps: receiving first audio information; determining whether the first device is a first reference device according to the first audio information and the target wake-up word, wherein a preset wake-up word of the first reference device is matched with the target wake-up word; if the device is the first reference device, outputting first response information; receiving second audio information; acquiring a first wake-up parameter of each awakened device, wherein the first wake-up parameter is acquired based on the second audio information; determining whether the first reference device is a target device according to the target wake-up word and the first wake-up parameter; and if the device is the target device, outputting second response information. Therefore, the selection accuracy of the response device can be improved, and the power consumption of the device is reduced.

Description

Technical Field

[0001] This application belongs to the field of data processing, and specifically relates to a voice wake-up method, apparatus, and device. Background Technology

[0002] As intelligent voice technology matures and smart devices become increasingly prevalent in homes, voice interaction with these devices is gaining acceptance. Currently, when multiple devices with voice wake-up capabilities coexist, a user uttering a wake-up word can simultaneously wake up all devices. To avoid this confusion, current voice wake-up methods keep all devices in a listening state and then select the appropriate device based on time differences. This not only causes all devices to continuously consume power but also results in low accuracy in selecting the correct device to respond to user requests in complex environments. Summary of the Invention

[0003] This application provides a voice wake-up method, apparatus, and device to improve the accuracy of selecting devices that respond to user needs, reduce device power consumption, and increase device battery life.

[0004] In a first aspect, embodiments of this application provide a voice wake-up method, applied to a first device in a working state within a sound pickup system, the sound pickup system including multiple devices with audio acquisition capabilities, the method comprising: Receive first audio information, the first audio information including a target wake-up word; Based on the first audio information and the target wake word, it is determined whether the first device is a first reference device, and the preset wake word of the first reference device matches the target wake word; If it is the first reference device, then output the first response information in response to the first audio information; Receive second audio information while in wake-up state; Obtain the first wake-up parameter for each woke-up device in the set of woke-up devices. The first wake-up parameter is obtained based on the second audio information. The set of woke-up devices includes the first reference device and a second device that has been woken up from the sleep state and whose preset wake-up word matches the target wake-up word. Determine whether the first device is the target device based on the target wake-up word and the first wake-up parameter; If it is the target device, then output the second response information in response to the second audio information.

[0005] According to the voice wake-up method provided in this application, determining whether the first device is a first reference device based on the first audio information and the target wake-up word includes: obtaining a second wake-up parameter for each first device in a first device set, wherein the second wake-up parameter is obtained based on the first audio information, and the first device set includes at least one first device that receives the first audio information in a working state; scoring each second wake-up parameter to obtain a first score value; and determining the first device as a first reference device when the first score value of the first device is the maximum value among the first score values ​​of the first matching device, wherein the first matching device is the first device in the first device set whose preset wake-up word matches the target wake-up word.

[0006] According to the voice wake-up method provided in this application, the step of determining whether the first reference device is the target device based on the target wake-up word and the first wake-up parameter includes: scoring each first wake-up parameter to obtain a second score value; and determining the first device as the target device when the second score value of the first device is the maximum value among the second score values ​​of the second matching device, and the second matching device being the wake-up device in the set of wake-up devices whose preset wake-up word matches the target wake-up word.

[0007] According to the voice wake-up method provided in this application, the method further includes: determining whether the first device is a second reference device based on the first audio information and the target wake-up word; if it is the second reference device, sending wake-up information to the second device, wherein the second device is a device in a sleep state in the sound pickup system.

[0008] According to the voice wake-up method provided in this application, determining whether the first device is a second reference device based on the first audio information and the target wake-up word includes: obtaining a second wake-up parameter for each first device in a first device set, wherein the second wake-up parameter is obtained based on the first audio information, and the first device set includes at least one first device that receives the first audio information in a working state; scoring each second wake-up parameter to obtain a first score value; and determining the first device as a second reference device when the first score value of the first device is the maximum value of the first score values ​​of the first devices included in the first device set.

[0009] According to the voice wake-up method provided in this application, each wake-up parameter is scored, including: obtaining the average energy, signal-to-noise ratio, and distortion of the audio information according to each wake-up parameter; and scoring each wake-up parameter according to the average energy, signal-to-noise ratio, and distortion.

[0010] According to the voice wake-up method provided in this application, sending wake-up information to the second device includes: determining the user's movement trajectory based on the first audio information; obtaining the location information of the second device; and sequentially sending wake-up information to the second device based on the movement trajectory and the location information.

[0011] Secondly, this application also provides a voice wake-up method, applied to a second device in a sleep state within a sound pickup system, wherein the sound pickup system includes multiple devices with audio acquisition capabilities, and the method includes: Receive wake-up information and adjust the sleep state to wake-up state. The wake-up information is sent by a second reference device, which is a device that receives the first audio information in the working state. Receive second audio information in the wake-up state; The first wake-up parameter of each woke-up device in the set of woke-up devices is obtained. The first wake-up parameter is obtained based on the second audio information. The set of woke-up devices includes a first reference device and a second device that is woken up from a sleep state by the second reference device and whose preset wake-up word matches the target wake-up word. The first reference device is a device that receives the first audio information in a working state and is used to respond to the first audio information. The preset wake-up word of the first reference device matches the target wake-up word. The target wake-up word is included in the first audio information. Determine whether the second device is the target device based on the target wake-up word and the first wake-up parameter; If it is the target device, then output the second response information in response to the second audio information.

[0012] According to the voice wake-up method provided in this application, before obtaining the first wake-up parameter of each wake-up device in the set of wake-up devices, the method further includes: The target wake-up word is determined based on the wake-up information; Determine whether the preset wake word of the second device matches the target wake word; If there is no match, return to sleep mode; If a match is found, the first wake-up parameter is obtained for each of the wake-up devices in the set of wake-up devices.

[0013] Thirdly, this application also provides a voice wake-up device, applied to a first device in a sound pickup system in a working state, the sound pickup system including multiple devices with audio acquisition capabilities, including: The first receiving unit is configured to receive first audio information, wherein the first audio information includes a target wake-up word; The first determining unit is configured to determine whether the first device is a first reference device based on the first audio information and the target wake-up word, wherein the preset wake-up word of the first reference device matches the target wake-up word. The first output unit is configured to output first response information for the first audio information if it is the first reference device. The second receiving unit is used to receive second audio information in the wake-up state; The acquisition unit is used to acquire a first wake-up parameter for each wake-up device in the set of wake-up devices. The first wake-up parameter is acquired based on the second audio information. The set of wake-up devices includes a first reference device and a second device that has been woken up from a sleep state and whose preset wake-up word matches the target wake-up word. The second determining unit is used to determine whether the first device is a target device based on the target wake-up word and the first wake-up parameter. The second output unit is configured to output second response information in response to the second audio information if it is the target device.

[0014] Fourthly, this application also provides a voice wake-up device, applied to a second device in a dormant state within a sound pickup system, wherein the sound pickup system includes multiple devices with audio acquisition capabilities, including: The first receiving unit is used to receive wake-up information and adjust the sleep state to the wake-up state. The wake-up information is sent by the second reference device, which is a device that receives the first audio information in the working state. The second receiving unit is used to receive second audio information in the wake-up state; The acquisition unit is used to acquire a first wake-up parameter for each of the wake-up devices in the set of wake-up devices. The first wake-up parameter is acquired based on the second audio information. The set of wake-up devices includes a first reference device and a second device that is woken from a sleep state by the second reference device and whose preset wake-up word matches the target wake-up word. The first reference device is a device that receives the first audio information in a working state and is used to respond to the first audio information. The preset wake-up word of the first reference device matches the target wake-up word, and the target wake-up word is included in the first audio information. A determining unit is configured to determine whether the second device is a target device based on the target wake-up word and the first wake-up parameter; The output unit is configured to output second response information in response to the second audio information if the target device is present.

[0015] Fifthly, embodiments of this application provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the steps of the voice wake-up method described in the first or second aspect.

[0016] In a sixth aspect, embodiments of this application provide a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the voice wake-up method described in the first or second aspect.

[0017] In a seventh aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the voice wake-up method described in the first or second aspect.

[0018] The voice wake-up method, apparatus, and device provided in this application firstly receive first audio information, which includes a target wake-up word. Then, based on the first audio information and the target wake-up word, it determines whether the first device is a first reference device. If it is the first reference device, it outputs first response information for the first audio information. Then, in the wake-up state, it receives second audio information and obtains first wake-up parameters for each woken-up device in the set of woken-up devices. Simultaneously, the second device receives wake-up information and obtains the first wake-up parameters for each woken-up device in the set of woken-up devices. Then, based on the second audio information, it determines a target second wake-up parameter. Finally, based on the target wake-up word and the first wake-up parameter, it determines a target device, and the target device outputs second response information for the second audio information. Therefore, in this solution, the wake-up process relies on the woken-up device to wake up the dormant device and put it into operation through signal transmission. Compared to existing device technologies where the microphone picks up sound continuously, the average power consumption is lower, significantly extending the standby time of offline powered devices. Furthermore, by combining wake-up parameters to determine the target device, the accuracy of selecting the device that responds to user needs can be improved. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application 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 application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is one of the flowcharts of the voice wake-up method provided in this application.

[0021] Figure 2 This is the second flowchart of the voice wake-up method provided in this application.

[0022] Figure 3 This is the third flowchart of the voice wake-up method provided in this application.

[0023] Figure 4 This is one of the functional unit block diagrams of the voice wake-up device provided in this application.

[0024] Figure 5 This is the second block diagram of the functional units of the voice wake-up device provided in this application.

[0025] Figure 6 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] Current voice wake-up methods involve all devices being in a listening state, and then selecting the device to respond to the user's request based on time differences. This not only causes all devices to continuously consume power, but also results in low accuracy in selecting the device to respond to the user's request in complex environments.

[0030] To address the aforementioned problems, this application provides a voice wake-up method, apparatus, and device. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0031] The audio pickup system includes multiple electronic devices with audio acquisition capabilities. These devices can be categorized into primary devices and secondary devices based on their current state. Specifically, the electronic device currently in operation is the primary device, and the electronic device currently in sleep mode is the secondary device. Primary devices can receive wake-up words while in operation. Secondary devices can be woken up by other primary devices while in sleep mode.

[0032] Please see Figure 1 , Figure 1 This is one of the flowcharts illustrating the voice wake-up method provided in this application. The voice wake-up method is applied to a first device in a voice pickup system that is in operation, and specifically includes the following steps.

[0033] S101, receive the first audio information.

[0034] The first audio information includes a target wake-up word. This target wake-up word is the preset wake-up word corresponding to the target device. It should be noted that multiple devices in the voice system can simultaneously have the target wake-up word as their preset wake-up word. As long as the devices in the sound pickup system are operational, they can receive this first audio information. Specifically, electronic devices with audio acquisition capabilities in the sound pickup system can simultaneously include multiple wake-up words. For example, a multi-wake-word device can be determined based on the operational status of each device; the preset wake-up word of this multi-wake-word device simultaneously includes the wake-up words of all electronic devices in the sound pickup system. For example, if a monitoring device is continuously operational, it can be identified as a multi-wake-word device.

[0035] S102, determine whether the first device is the first reference device based on the first audio information and the target wake-up word.

[0036] The first reference device has a preset wake-up word that matches the target wake-up word, used to respond to the first audio information. The preset wake-up word of the first reference device is the target wake-up word, while the preset wake-up word of the second device may or may not be the target wake-up word.

[0037] S103, if it is the first reference device, then output the first response information for the first audio information.

[0038] The first response information can be a response to a target wake word. For example, if the target wake word is "puppy, puppy," the corresponding first response information is "I am here." After entering the wake-up state, the first electronic device can continue to receive the second audio information. If the first electronic device is not the first reference device and / or the second reference device, then the first electronic device remains in a working state and does not receive subsequent second audio information, or receives the second audio information but does not process it.

[0039] S104, receives second audio information in the wake-up state.

[0040] In specific implementation, before determining the first reference device and / or the second reference device, if the preset wake-up word of the first device is the same as the target wake-up word, the first device enters the wake-up state and sends wake-up information to other first devices to wake them up. If the preset wake-up word of the first device does not match the target wake-up word, the first audio information is stored. When the first device is determined to be the first reference device and / or the second reference device, the first device enters the wake-up state.

[0041] S105, obtain the first wake-up parameter of each wake-up device in the set of wake-up devices.

[0042] The first wake-up parameter is obtained based on the second audio information. The set of woken-up devices includes the first reference device and second devices that have been woken up from sleep mode and whose preset wake-up words match the target wake-up words. That is, the set of woken-up devices can include all devices in a wake-up state, including the first reference device and all second devices woken up by the second reference device. Specifically, matching can also be performed based on the preset wake-up words and target wake-up words of each electronic device with sound pickup function. If the preset wake-up word of the second reference device matches the target wake-up word, the second reference device can be added to the set of woken-up devices. In other words, the set of woken-up devices includes all electronic devices in a wake-up state whose preset wake-up words match the target wake-up words.

[0043] In practice, the first wake-up parameter can be obtained by the first electronic device by analyzing the second audio information locally, or the first electronic device can send the second audio information to the cloud server for analysis and acquisition.

[0044] S106, determine whether the first device is the target device based on the target wake-up word and the first wake-up parameter.

[0045] In this process, after acquiring the first wake-up parameter, the first device sends it to other woken devices in the set of woken devices. At the same time, these other woken devices also acquire the first wake-up parameter based on the second audio information and send it to all other woken devices in the set except themselves. The first reference device can then determine whether it is the target device based on the acquired first wake-up parameter. Similarly, other woken devices can determine whether they are the target device based on the acquired first wake-up parameter and the target wake-up word. In a specific implementation, a first time window can be set for each woken device to acquire the first wake-up parameter. After this first time window has passed, the device determines whether it is the target device based on all acquired first wake-up parameters.

[0046] S107, if it is the target device, then output the second response information for the second audio information.

[0047] The second response information can be the execution result of the second audio information. For example, if the audio information is "play a song", then the second response information is the music played.

[0048] As can be seen, in this embodiment, the device is in a sleep state and does not perform audio wake-up. The first device in the working state responds promptly. In the wake-up process, the sleep device is awakened by the wake-up device through signal transmission and enters the working state. Compared with the existing device technology solution where the microphone picks up sound for a long time, the average power consumption is lower, which can significantly extend the standby time of the offline power supply device.

[0049] In one possible embodiment, determining whether the first device is a first reference device based on the first audio information and the target wake-up word includes: obtaining a second wake-up parameter for each first device in a first device set, the second wake-up parameter being obtained based on the first audio information, the first device set including at least one first device receiving the first audio information in a working state; scoring each second wake-up parameter to obtain a first score value; and determining the first device as a first reference device when the first score value of the first device is the maximum value among the first score values ​​of the first matching device, the first matching device being the first device in the first device set whose preset wake-up word matches the target wake-up word.

[0050] This implementation allows multiple electronic devices in operational states to simultaneously receive the first audio information, thus potentially including a first device set. Upon receiving the first audio information, each first device can analyze the information locally to obtain a second wake-up parameter, or it can obtain the second wake-up parameter from a cloud server. After obtaining the first audio information, the first electronic device can send the second wake-up parameter to each of the other first devices in the first device set, and can also receive the second wake-up parameter from other first electronic devices. In other words, each first electronic device, upon receiving the first audio information, can obtain the second wake-up parameter for the first audio information using the aforementioned method, and then send the obtained second wake-up parameter to each of the other first electronic devices in the first device set. Specifically, a second time window can be set for each first device to obtain the second wake-up parameter. After the second time window has passed, the device is then determined whether it is a first reference device and / or a second reference device based on all the obtained second wake-up parameters.

[0051] For example, there are three first electronic devices currently in operation, namely, the first device set includes three first electronic devices: first electronic device 1, first electronic device 2, and first electronic device 3. At this time, each of these three first electronic devices obtains a second wake-up parameter based on the first audio information. Taking first electronic device 1 as an example, first electronic device 1 obtains the second wake-up parameter based on the first audio information and sends the second wake-up parameter to first electronic device 2 and first electronic device 3 respectively. Simultaneously, it receives second wake-up parameters from both first electronic device 2 and first electronic device 3. At this point, first electronic device 1 can determine whether it is the first reference device based on the target wake-up word and multiple second wake-up parameters.

[0052] In the specific implementation, when scoring the wake-up parameters, the scoring can be performed locally on the first electronic device, or the received first audio information can be uploaded to a cloud server for scoring and then distributed to the first electronic device. When determining whether a device is the first reference device and / or the second reference device, the score values ​​can be sorted from high to low. Combined with the wake-up words supported by the device, the device with the highest score among all devices matching the target wake-up word is designated as the first reference device A2. Then, A2 is used as the initial wake-up device for this round, and other first devices return to the working state, waiting for the next round of wake-up.

[0053] As can be seen, in this embodiment, scoring is performed based on wake-up parameters, and then the first electronic device with the highest score matching the wake-up word is used as the first reference device to respond to the first audio information. This can reduce the overall device power consumption in the sound pickup system, while also waking up the target device in a timely manner and improving the user experience.

[0054] In one possible embodiment, determining whether the first reference device is the target device based on the target wake-up word and the first wake-up parameter includes: scoring each first wake-up parameter to obtain a second score value; and determining the first device as the target device when the second score value of the first device is the maximum value among the second score values ​​of the second matching device, wherein the second matching device is a wake-up device in the set of wake-up devices whose preset wake-up word matches the target wake-up word.

[0055] In determining whether a device is the target device, each woken-up device in the set of woken-up devices can obtain a first wake-up parameter based on the second audio information. The obtained first wake-up parameter is then sent to each of the other second devices in the set of woken-up devices. Each woken-up device can then obtain a score and ranking based on the obtained first wake-up parameter. The device with the highest score that supports the target wake-up word is selected as the target response device C1 and proceeds with subsequent audio pickup and response. Non-target response devices end their audio pickup state and return to sleep or working state, awaiting wake-up voice activation.

[0056] As can be seen, in this embodiment, scoring the wake-up parameters obtained based on audio information and determining the target response device based on the score and the target wake-up word can improve the accuracy of confirming the device that responds to the user's needs.

[0057] In one possible embodiment, the method further includes: determining whether the first device is a second reference device based on the first audio information and the target wake-up word; if it is the second reference device, sending wake-up information to the second device, wherein the second device is a device in a sleep state in the sound pickup system.

[0058] If the first device is determined to be the second reference device, then the first device sends a wake-up message to each of the second devices in a sleep state. This wake-up message may include a target wake-up word. After a second reference device sends a wake-up message, if its preset wake-up word is not the target wake-up word, it can return to its working state and stop receiving second audio information.

[0059] It should be noted that the first device can be both the first reference device and the second reference device.

[0060] In one possible embodiment, determining whether the first device is a second reference device based on the first audio information and the target wake-up word includes: obtaining a second wake-up parameter for each first device in a first device set, the second wake-up parameter being obtained based on the first audio information, the first device set including at least one first device that receives the first audio information in a working state; scoring each second wake-up parameter to obtain a first score value; and determining the first device as a second reference device if the first score value of the first device is the maximum value of the first score values ​​of the first devices included in the first device set.

[0061] The calculation method for the first score value can refer to the content of the aforementioned embodiment, and will not be repeated here. The first device with the highest first score value can be recorded as the second reference device A1. Then, A1 is synchronously used as the initial wake-up device for this wake-up, and the other first devices return to the working state, waiting for the next round of wake-up. That is, A1 and A2 are synchronously woken up at this time.

[0062] As can be seen, in this embodiment, the second reference device wakes up the second device from the sleep state, and the first reference device responds to the first audio information, which can ensure the response speed, improve the user experience, and reduce the overall power consumption of the system.

[0063] In one possible embodiment, scoring each wake-up parameter includes: obtaining the average energy, signal-to-noise ratio, and distortion of the audio information based on each wake-up parameter; and scoring each wake-up parameter based on the average energy, signal-to-noise ratio, and distortion.

[0064] The score, whether the first or second score is obtained, can be calculated using the following formula: Score=α×Norm(SNR)+β×Norm(Energy)-γ×Norm(Distortion) Where Norm(x) is the normalization function, used to normalize each indicator to a range of 0 to 1. α, β, and γ are weight parameters, representing the importance of each indicator, and are not zero. Energy is the mean energy, representing the volume intensity, indirectly reflecting distance. SNR is the signal-to-noise ratio, representing the signal purity, and Distortion is the distortion, representing the speech clarity.

[0065] The wake-up parameters include the average energy, signal-to-noise ratio, and distortion of the audio information acquired by each device. A score is then calculated for each wake-up parameter based on this information. For example, when determining whether a first device is a first reference device and / or a second reference device, a score is calculated for each acquired first wake-up parameter based on the above formula to obtain a first score value. When determining a target response device, a score is calculated for each acquired second wake-up parameter based on the above formula to obtain a second score value.

[0066] As can be seen, in this embodiment, by combining volume intensity, signal-to-noise ratio, and distortion, and normalizing them before calculating the score using a linear equation, the accuracy of selecting the first reference device, the second reference device, and the target response device can be improved.

[0067] In one possible embodiment, sending a wake-up signal to the second device includes: determining the user's movement trajectory based on the first audio information; obtaining the location information of the second device; and sequentially sending a wake-up signal to the second device based on the movement trajectory and the location information.

[0068] The second reference device can analyze the first audio information to determine its intensity and direction. Based on this, it can then determine the user's location and movement trajectory, and gradually activate nearby devices as the user moves. Alternatively, the second reference device can acquire video surveillance footage and determine the user's movement trajectory and location based on it. If a second device exists that does not match the user's trajectory, it will not be activated.

[0069] As can be seen, in this embodiment, gradually waking up nearby second devices based on the user's movement can improve the accuracy and responsiveness of voice interaction in mobile scenarios and avoid resource consumption caused by unrelated devices being accidentally woken up.

[0070] Please see Figure 2 , Figure 2 This is the second flowchart of the voice wake-up method provided in this application. The voice wake-up method is applied to a second device in a sleep state in the sound pickup system, and specifically includes the following steps.

[0071] S201, Receive wake-up information and adjust the sleep state to wake-up state.

[0072] The wake-up message is sent by a second reference device, which is the device that receives the first audio information while in working mode. This wake-up message may include a target wake-up word and a first wake-up parameter from the second reference device. Specifically, if the preset wake-up word of the second reference device does not match the target wake-up word, the first wake-up parameter may not be included in the wake-up message. That is, after receiving the wake-up message from the second reference device, the second device, which is in sleep mode, enters wake-up mode from sleep mode.

[0073] S202, receiving second audio information in the wake-up state.

[0074] The second audio information may include user requests.

[0075] S203, obtain the first wake-up parameter of each wake-up device in the set of wake-up devices.

[0076] Wherein, the first wake-up parameter is obtained based on the second audio information, the set of woken-up devices includes a first reference device and a second device that is woken up from the sleep state by the second reference device and whose preset wake-up word matches the target wake-up word, the first reference device is a device that receives the first audio information in the working state and is used to respond to the first audio information, the preset wake-up word of the first reference device matches the target wake-up word, and the target wake-up word is included in the first audio information.

[0077] In a specific implementation, the set of awakened devices may include a first reference device, a second reference device, and all second devices awakened by the second reference device. Alternatively, the set of awakened devices may include a first reference device, a second reference device whose preset wake word matches the target wake word, and other second devices.

[0078] In a specific implementation, each woken-up device in the set of woken-up devices can obtain the first wake-up parameter based on the second audio information, and send the first wake-up parameter to other woken-up devices in the set of woken-up devices. The method of obtaining the first wake-up parameter can be the same as that in the aforementioned embodiments.

[0079] S204, determine whether the second device is the target device based on the target wake-up word and the first wake-up parameter.

[0080] In this case, the target device can also be determined based on the first wake-up parameter to determine the first score value, and then the wake-up device with the highest score value and whose preset wake-up word matches the target wake-up word is determined as the target response device.

[0081] S205, if it is the target device, then output the second response information for the second audio information.

[0082] As can be seen, in this embodiment, the second device is in a sleep state and does not perform audio wake-up. Instead, it is woken up by sending a signal after another device is woken up by audio, and then the target response device is determined by subsequent voice. Compared with existing device technologies where the microphone picks up sound continuously, the average power consumption is lower, which can significantly extend the standby time of offline powered devices.

[0083] In one possible embodiment, before obtaining the first wake-up parameter of each woken-up device in the set of woken-up devices, the method further includes: determining the target wake-up word based on the wake-up information; determining whether the preset wake-up word of the second device matches the target wake-up word; if they do not match, returning to a sleep state; if they match, obtaining the first wake-up parameter of each woken-up device in the set of woken-up devices.

[0084] The wake-up parameters may include a target wake-up word. If the preset wake-up word of the second device does not match the target wake-up word, the second wake-up parameters can return to the sleep state. If the preset wake-up word of the second device matches the target wake-up word, then the second device obtains the first wake-up parameters based on the second audio information.

[0085] As can be seen, in this embodiment, the first wake-up parameter is only obtained when the preset wake-up word matches the target wake-up word, which can reduce device power consumption.

[0086] The following is combined Figure 3 The overall process of this application is described below.

[0087] For example, the sound pickup system includes three electronic devices with sound pickup capabilities: a battery-powered toy, a television, and a speaker. The preset wake-up word for the battery-powered toy is "puppy puppy," the preset wake-up word for the television is "television television," and the preset wake-up words for the speaker include both "television television" and "puppy puppy." At this time, the battery-powered toy is in sleep mode, i.e., the battery-powered toy is the second device, while the television and speaker are both in working mode, i.e., the television and speaker are the first devices.

[0088] First, the user outputs the wake-up voice "puppy, puppy." At this point, both the TV and the speaker, which are already in operation, receive the wake-up voice, i.e., the first audio information. Since the TV's default wake-up word is "TV, TV," it doesn't respond to "puppy, puppy," and stores the audio. Second, because the speaker's wake-up word contains "puppy, puppy," it is awakened and enters the wake-up state. Then, the speaker generates a second wake-up parameter based on the first audio information, which includes the target wake-up word, average energy, signal-to-noise ratio, and distortion information, and sends this wake-up packet to the TV within time window T1. Upon receiving the wake-up packet, the TV is awakened and generates a wake-up packet containing the same information based on the stored audio, simultaneously sending it to the speaker. After time window T1 ends, the TV and speaker score the received wake-up packets, obtaining a first score value for each packet, and then sorting the scores. This is the pre-wake-up process. If, based on the score value sorting, the TV is determined to be the device with the highest score among all devices (i.e., the second reference device A1), and the speaker is determined to be the device with the highest matching wake-up device score (i.e., the first reference device A2), then the speaker is awakened.

[0089] The wake-up process then begins. First, the speaker, acting as the first reference device, outputs a first response message, such as "I'm here." Simultaneously, the TV, acting as the second reference device, sends an activation command (wake-up information) to the battery toy. The battery toy then enters a pickup state, receiving the second audio message, the command voice, such as "Tell a story." Then, within time window T2, the woken-up devices—the battery toy, TV, and speaker—each acquire the first wake-up parameters, i.e., the wake-up packet, and send these parameters to other woken-up devices. This continues until time window T2 ends, where each woken-up device is scored and ranked, ultimately determining the battery toy as the target response device. The battery toy then outputs the second response message, i.e., the answer audio, and subsequent interactions are also based on the user's voice commands.

[0090] The voice wake-up device provided in the embodiments of this application is described below. The voice wake-up device described below can be referred to in correspondence with the voice wake-up method described above.

[0091] Please see Figure 4A voice wake-up device is applied to a first device in a voice pickup system that is in operation. The voice pickup system includes multiple devices with audio acquisition capabilities, including: a first receiving unit, configured to determine whether the first device is a first reference device based on the first audio information and the target wake-up word, wherein the preset wake-up word of the first reference device matches the target wake-up word; a first output unit, configured to output first response information for the first audio information if the device is the first reference device; a second receiving unit, configured to receive second audio information in the wake-up state; an acquisition unit, configured to acquire a first wake-up parameter for each wake-up device in the set of wake-up devices, wherein the first wake-up parameter is acquired based on the second audio information, and the set of wake-up devices includes the first reference device and a second device that has been woken up from a sleep state and whose preset wake-up word matches the target wake-up word; a second determining unit, configured to determine whether the first device is a target device based on the target wake-up word and the first wake-up parameter; and a second output unit, configured to output second response information for the second audio information if the device is the target device.

[0092] In one possible embodiment, in determining whether the first device is a first reference device based on the first audio information and the target wake-up word, the first determining unit is specifically configured to: obtain a second wake-up parameter for each first device in the first device set, wherein the second wake-up parameter is obtained by scoring each second wake-up parameter based on the first audio information, wherein at least one first device in the first device set that receives the first audio information in a working state is scored to obtain a first score value; and if the first score value of the first device is the maximum value among the first score values ​​of the first matching device, determine that the first device is a first reference device, and the first matching device is a first device in the first device set whose preset wake-up word matches the target wake-up word.

[0093] In one possible embodiment, in determining whether the first reference device is a target device based on the target wake-up word and the first wake-up parameter, the second determining unit is specifically configured to: score each first wake-up parameter to obtain a second score value; and determine the first device as a target device when the second score value of the first device is the maximum value among the second score values ​​of the second matching device, and the second matching device is an already woken device in the set of already woken devices whose preset wake-up word matches the target wake-up word.

[0094] In one possible embodiment, the voice wake-up device further includes a sending unit, which is specifically used to: determine whether the first device is a second reference device based on the first audio information and the target wake-up word; if it is the second reference device, send wake-up information to the second device, which is a device in a sleep state in the sound pickup system.

[0095] In one possible embodiment, in determining whether the first device is a second reference device based on the first audio information and the target wake-up word, the first determining unit is specifically configured to: obtain a second wake-up parameter for each first device in a first device set, the second wake-up parameter being obtained based on the first audio information, the first device set including at least one first device that receives the first audio information in a working state; score each second wake-up parameter to obtain a first score value; and determine the first device as a second reference device if the first score value of the first device is the maximum value of the first score values ​​of the first devices included in the first device set.

[0096] In one possible embodiment, in scoring each wake-up parameter, the first determining unit or the second determining unit is specifically configured to: obtain the average energy, signal-to-noise ratio, and distortion of the audio information according to each wake-up parameter; and score each wake-up parameter according to the average energy, signal-to-noise ratio, and distortion.

[0097] In one possible embodiment, in sending wake-up information to the second device, the sending unit is specifically configured to: determine the user's movement trajectory based on the first audio information; obtain the location information of the second device; and sequentially send wake-up information to the second device based on the movement trajectory and the location information.

[0098] Please see Figure 5This application also provides a voice wake-up device applied to a second device in a sleep state within a sound pickup system. The sound pickup system includes multiple devices with audio acquisition capabilities, comprising: a first receiving unit for receiving wake-up information and adjusting the sleep state to a wake-up state, wherein the wake-up information is sent by a second reference device, which is a device that receives first audio information in a working state; a second receiving unit for receiving second audio information in the wake-up state; an acquisition unit for acquiring a first wake-up parameter for each woke-up device in a set of woke-up devices, wherein the first wake-up parameter is acquired based on the second audio information, and the set of woke-up devices includes a first reference device and a second device that is woken from a sleep state by the second reference device and whose preset wake-up word matches a target wake-up word, wherein the first reference device is a device that receives the first audio information in a working state and is used to respond to the first audio information, and whose preset wake-up word matches the target wake-up word, which is contained in the first audio information; a determining unit for determining whether the second device is a target device based on the target wake-up word and the first wake-up parameter; and an output unit for outputting second response information for the second audio information if the second device is the target device.

[0099] In one possible embodiment, before acquiring the first wake-up parameter of each woken-up device in the set of woken-up devices, the acquisition unit is further configured to: determine the target wake-up word based on the wake-up information; determine whether the preset wake-up word of the second device matches the target wake-up word; if they do not match, return to the sleep state; if they match, acquire the first wake-up parameter of each woken-up device in the set of woken-up devices.

[0100] Please see Figure 6 , Figure 6A schematic diagram of the physical structure of an electronic device is provided. This electronic device may include a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other via the communication bus 640. Specifically, this electronic device may be a first device. The processor 610 can invoke a computer program in the memory 630 to execute a voice wake-up method, the method comprising: receiving first audio information, the first audio information including a target wake-up word; determining whether a first device is a first reference device based on the first audio information and the target wake-up word, wherein a preset wake-up word of the first reference device matches the target wake-up word; if it is the first reference device, outputting first response information for the first audio information; receiving second audio information in a wake-up state; acquiring a first wake-up parameter for each wake-up device in a set of wake-up devices, the first wake-up parameter being acquired based on the second audio information, wherein the set of wake-up devices includes the first reference device and a second device whose preset wake-up word matches the target wake-up word and has been woken up from a sleep state; determining whether the first device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0101] Alternatively, the electronic device may be a second device. The processor 610 may invoke a computer program in the memory 630 to execute a voice wake-up method, the method comprising: receiving wake-up information and adjusting the sleep state to an awake state, the wake-up information being sent by a second reference device, the second reference device being a device that receives first audio information in an operating state; receiving second audio information in the awake state; acquiring a first wake-up parameter for each awakened device in a set of awakened devices, the first wake-up parameter being acquired based on the second audio information, the set of awakened devices including a first reference device and a second device awakened from a sleep state by the second reference device by a preset wake-up word matching a target wake-up word, the first reference device being a device that receives the first audio information in an operating state and is used to respond to the first audio information, the preset wake-up word of the first reference device matching the target wake-up word, the target wake-up word being included in the first audio information; determining whether the second device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0102] Furthermore, the logical instructions in the aforementioned memory 630 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 this application, in essence, 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 this application. 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.

[0103] On the other hand, this application 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 can execute the voice wake-up method provided in the above embodiments. The method includes: receiving first audio information, the first audio information including a target wake-up word; determining whether a first device is a first reference device based on the first audio information and the target wake-up word, wherein a preset wake-up word of the first reference device matches the target wake-up word; if it is the first reference device, outputting first response information for the first audio information; receiving second audio information in a wake-up state; obtaining a first wake-up parameter for each wake-up device in a set of wake-up devices, the first wake-up parameter being obtained based on the second audio information, wherein the set of wake-up devices includes the first reference device and a second device whose preset wake-up word matches the target wake-up word and has been woken up from a sleep state; determining whether the first device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0104] Alternatively, the computer can execute the voice wake-up method provided in the above embodiments, the method comprising: receiving wake-up information and adjusting the sleep state to a wake-up state, the wake-up information being sent by a second reference device, the second reference device being a device that receives first audio information in a working state; receiving second audio information in the wake-up state; acquiring a first wake-up parameter for each woke-up device in a set of woke-up devices, the first wake-up parameter being acquired based on the second audio information, the set of woke-up devices including a first reference device and a second device that is woken from a sleep state by the second reference device and whose preset wake-up word matches a target wake-up word, the first reference device being a device that receives the first audio information in a working state and is used to respond to the first audio information, the preset wake-up word of the first reference device matching the target wake-up word, the target wake-up word being included in the first audio information; determining whether the second device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0105] On the other hand, embodiments of this application also provide a processor-readable storage medium storing a computer program for causing a processor to execute the voice wake-up method provided in the above embodiments. The method includes: receiving first audio information, the first audio information including a target wake-up word; determining whether a first device is a first reference device based on the first audio information and the target wake-up word, wherein a preset wake-up word of the first reference device matches the target wake-up word; if it is the first reference device, outputting first response information for the first audio information; receiving second audio information in a wake-up state; obtaining a first wake-up parameter for each wake-up device in a set of wake-up devices, the first wake-up parameter being obtained based on the second audio information, the set of wake-up devices including the first reference device and a second device whose preset wake-up word matches the target wake-up word and has been woken up from a sleep state; determining whether the first device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0106] Alternatively, the computer program can be used to cause the processor to execute the voice wake-up method provided in the above embodiments. The method includes: receiving wake-up information and adjusting the sleep state to a wake-up state, the wake-up information being sent by a second reference device, the second reference device being a device that receives first audio information in a working state; receiving second audio information in the wake-up state; obtaining a first wake-up parameter for each woken-up device in a set of woken-up devices, the first wake-up parameter being obtained based on the second audio information, the set of woken-up devices including a first reference device and a second device whose preset wake-up word matches a target wake-up word and is woken from a sleep state by the second reference device, the first reference device being a device that receives the first audio information in a working state and is used to respond to the first audio information, the preset wake-up word of the first reference device matching the target wake-up word, the target wake-up word being included in the first audio information; determining whether the second device is a target device based on the target wake-up word and the first wake-up parameter; if it is the target device, outputting second response information for the second audio information.

[0107] The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO)), optical memory (e.g., CD, DVD, BD, HVD), and semiconductor memory (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)).

[0108] 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.

[0109] 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.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such 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 this application.

Claims

1. A voice wake-up method, characterized in that, The method involves applying a first device in a working state within a sound pickup system, wherein the sound pickup system includes multiple devices with audio acquisition capabilities, and includes: Receive first audio information, the first audio information including a target wake-up word; Based on the first audio information and the target wake word, it is determined whether the first device is a first reference device, and the preset wake word of the first reference device matches the target wake word; If it is the first reference device, then output the first response information in response to the first audio information; Receive second audio information while in wake-up state; Obtain the first wake-up parameter for each woke-up device in the set of woke-up devices. The first wake-up parameter is obtained based on the second audio information. The set of woke-up devices includes the first reference device and a second device that has been woken up from the sleep state and whose preset wake-up word matches the target wake-up word. Determine whether the first device is the target device based on the target wake-up word and the first wake-up parameter; If it is the target device, then output the second response information in response to the second audio information.

2. The method according to claim 1, characterized in that, Determining whether the first device is the first reference device based on the first audio information and the target wake-up word includes: Obtain a second wake-up parameter for each first device in the first device set, the second wake-up parameter being obtained based on the first audio information, the first device set including at least one first device that receives the first audio information in a working state; Each second wake-up parameter is scored to obtain a first score value; If the first score value of the first device is the maximum value among the first score values ​​of the first matching device, the first device is determined to be the first reference device, and the first matching device is the first device in the first device set whose preset wake word matches the target wake word.

3. The method according to claim 1, characterized in that, Determining whether the first reference device is the target device based on the target wake-up word and the first wake-up parameter includes: Each first wake-up parameter is scored to obtain a second score value; If the second score of the first device is the maximum value among the second score values ​​of the second matching device, the first device is determined to be the target device, and the second matching device is the woken-up device in the set of woken-up devices whose preset wake-up word matches the target wake-up word.

4. The method according to claim 1, characterized in that, The method further includes: Determine whether the first device is the second reference device based on the first audio information and the target wake word; If it is the second reference device, then a wake-up message is sent to the second device, which is a device in the sound pickup system that is in a sleep state.

5. The method according to claim 4, characterized in that, The step of determining whether the first device is the second reference device based on the first audio information and the target wake-up word includes: Obtain a second wake-up parameter for each first device in the first device set, the second wake-up parameter being obtained based on the first audio information, the first device set including at least one first device that receives the first audio information in a working state; Each second wake-up parameter is scored to obtain a first score value; If the first device's first score is the maximum first score of the first devices included in the first device set, then the first device is determined to be the second reference device.

6. The method according to any one of claims 3-5, characterized in that, Each wake-up parameter is scored, including: The average energy, signal-to-noise ratio, and distortion of the audio information are obtained based on each wake-up parameter. Each wake-up parameter is scored based on its mean energy, signal-to-noise ratio, and distortion.

7. The method according to claim 4, characterized in that, Sending wake-up information to the second device includes: Determine the user's movement trajectory based on the first audio information; Obtain the location information of the second device; Wake-up information is sent to the second device sequentially based on the movement trajectory and the location information.

8. A voice wake-up method, characterized in that, A second device in a dormant state is applied to a sound pickup system, which includes multiple devices with audio acquisition capabilities; the method includes: Receive wake-up information and adjust the sleep state to wake-up state. The wake-up information is sent by a second reference device, which is a device that receives the first audio information in the working state. Receive second audio information in the wake-up state; The first wake-up parameter of each woke-up device in the set of woke-up devices is obtained. The first wake-up parameter is obtained based on the second audio information. The set of woke-up devices includes a first reference device and a second device that is woken up from a sleep state by the second reference device and whose preset wake-up word matches the target wake-up word. The first reference device is a device that receives the first audio information in a working state and is used to respond to the first audio information. The preset wake-up word of the first reference device matches the target wake-up word. The target wake-up word is included in the first audio information. Determine whether the second device is the target device based on the target wake-up word and the first wake-up parameter; If it is the target device, then output the second response information in response to the second audio information.

9. The method according to claim 8, characterized in that, Before obtaining the first wake-up parameter of each wake-up device in the set of wake-up devices, the method further includes: The target wake-up word is determined based on the wake-up information; Determine whether the preset wake word of the second device matches the target wake word; If there is no match, return to sleep mode; If a match is found, the first wake-up parameter is obtained for each of the wake-up devices in the set of wake-up devices.

10. A voice wake-up device, characterized in that, A first device used in a sound pickup system in a working state, the sound pickup system including multiple devices with audio acquisition capabilities, including: The first receiving unit is configured to receive first audio information, wherein the first audio information includes a target wake-up word; The first determining unit is configured to determine whether the first device is a first reference device based on the first audio information and the target wake-up word, wherein the preset wake-up word of the first reference device matches the target wake-up word. The first output unit is configured to output first response information for the first audio information if it is the first reference device. The second receiving unit is used to receive second audio information in the wake-up state; The acquisition unit is used to acquire a first wake-up parameter for each wake-up device in the set of wake-up devices. The first wake-up parameter is acquired based on the second audio information. The set of wake-up devices includes a first reference device and a second device that has been woken up from a sleep state and whose preset wake-up word matches the target wake-up word. The second determining unit is used to determine whether the first device is a target device based on the target wake-up word and the first wake-up parameter. The second output unit is configured to output second response information in response to the second audio information if it is the target device.

11. A voice wake-up device, characterized in that, A second device applied in a dormant state within a sound pickup system, the sound pickup system comprising multiple devices with audio acquisition capabilities, including: The first receiving unit is used to receive wake-up information and adjust the sleep state to the wake-up state. The wake-up information is sent by the second reference device, which is a device that receives the first audio information in the working state. The second receiving unit is used to receive second audio information in the wake-up state; The acquisition unit is used to acquire a first wake-up parameter for each of the wake-up devices in the set of wake-up devices. The first wake-up parameter is acquired based on the second audio information. The set of wake-up devices includes a first reference device and a second device that is woken from a sleep state by the second reference device and whose preset wake-up word matches the target wake-up word. The first reference device is a device that receives the first audio information in a working state and is used to respond to the first audio information. The preset wake-up word of the first reference device matches the target wake-up word, and the target wake-up word is included in the first audio information. A determining unit is configured to determine whether the second device is a target device based on the target wake-up word and the first wake-up parameter; The output unit is configured to output second response information in response to the second audio information if the target device is present.

12. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the voice wake-up method according to any one of claims 1 to 9.

13. 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 steps of the voice wake-up method according to any one of claims 1 to 9.

14. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the voice wake-up method according to any one of claims 1 to 9.

Images