Control method of Internet of Things equipment and Internet of Things equipment

Abstract

The invention discloses a control method of Internet of Things equipment and the Internet of Things equipment, and belongs to the technical field of communication. The control method of the Internet of Things equipment is applied to the Internet of Things equipment, the Internet of Things equipment comprises a main processor, a coprocessor and a voice wake-up module, the power consumption of the voice wake-up module in a first time period is smaller than a first threshold value, the voice wake-up module is in communication connection with the coprocessor, and the coprocessor is in communication connection with the main processor; the method comprises the steps that the voice wake-up module acquires first audio data; the voice wake-up module obtains a first keyword in the first audio data and type indication information of the first keyword; and the voice awakening module awakens the coprocessor in the dormant state and / or the main processor in the dormant state according to the type indication information of the first keyword. Therefore, according to the embodiment of the invention, the ultra-low standby power consumption can be realized, and meanwhile, the user can be ensured to carry out non-inductive interaction with the equipment through voice at any time.

Description

Technical Field

[0001] This application belongs to the field of communication technology, specifically relating to a control method for an Internet of Things (IoT) device and the IoT device itself. Background Technology

[0002] As IoT devices become increasingly complex, power consumption has gradually become a key factor limiting their battery life. To balance performance and battery life, the industry generally adopts a dual-system architecture: a high-performance main processor (such as an application processor (AP)) handles complex computing tasks, while a low-power coprocessor (such as a microcontroller unit (MCU)) is responsible for sensor data acquisition and device status monitoring, and controls the main processor to power down when appropriate to reduce power consumption.

[0003] Currently, the main processor's power-off is controlled by sensor status. However, tests have shown that once the device enters deep standby mode (i.e., both the main processor and coprocessor are asleep), the voice wake-up function immediately fails. If the coprocessor or part of the main processor continues to work to maintain voice wake-up capability, standby power consumption increases significantly, making it difficult to achieve the ideal "long standby" goal.

[0004] Therefore, it is clear that how to achieve ultra-low standby power consumption while still ensuring that users can seamlessly interact with devices via voice at any time is an urgent problem that needs to be solved. Summary of the Invention

[0005] The purpose of this application is to provide a control method and an IoT device that can achieve ultra-low standby power consumption while ensuring that users can interact with the device seamlessly via voice at any time.

[0006] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, embodiments of this application provide a control method for an Internet of Things (IoT) device, which is applied to an IoT device. The IoT device includes a main processor, a coprocessor, and a voice wake-up module. The power consumption of the voice wake-up module is less than a first threshold during a first time period. The voice wake-up module is communicatively connected to the coprocessor, and the coprocessor is communicatively connected to the main processor. The method includes: The voice wake-up module acquires the first audio data; The voice wake-up module acquires the first keyword in the first audio data and the type indication information of the first keyword; The voice wake-up module wakes up the coprocessor and / or the main processor that are in a dormant state according to the type indication information of the first keyword.

[0007] Therefore, in the embodiments of this application, a voice wake-up module with power consumption less than a first threshold during a first time period can be set in the IoT device, and the voice wake-up module can be connected to the coprocessor of the IoT device, and the coprocessor can be connected to the main processor. In this way, the voice wake-up module can obtain the first audio data, thereby obtaining the first keyword and the type indication information of the first keyword in the first audio data, and then the voice wake-up module can wake up the coprocessor and / or the main processor in a dormant state according to the type indication information of the first keyword.

[0008] As can be seen, in the embodiments of this application, a dedicated hardware unit (i.e., a voice wake-up module) is added to the IoT device. This transforms the voice monitoring entity from the main processor or coprocessor in sleep mode into a dedicated low-power voice wake-up module, resulting in a significant leap in power consumption. Furthermore, after the voice wake-up module acquires the audio data, it first identifies the type of keywords in the audio data and then determines the wake-up target based on the keyword type, achieving refined power consumption management. Therefore, the embodiments of this application can achieve ultra-low standby power consumption while ensuring that users can seamlessly interact with the device via voice at any time.

[0009] Secondly, this application provides an Internet of Things (IoT) device, which includes a main processor, a coprocessor, and a voice wake-up module. The power consumption of the voice wake-up module is less than a first threshold during a first time period. The voice wake-up module is communicatively connected to the coprocessor, and the coprocessor is communicatively connected to the main processor. The voice wake-up module is used for: Obtain the first audio data; Obtain the first keyword and the type indication information of the first keyword from the first audio data; Based on the type indication information of the first keyword, wake up the coprocessor and / or the main processor that are in a dormant state.

[0010] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0011] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the application embodiments, the drawings used in the description of the application embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart illustrating a control method for an Internet of Things (IoT) device provided in an embodiment of this application; Figure 2 This is a structural block diagram of an IoT device in a specific embodiment provided in this application. Figure 3 This is a flowchart illustrating an application scenario of an IoT device in a specific implementation provided in this application. Figure 4 This is a schematic diagram of the processing flow of an IoT device in an application scenario in a specific embodiment provided in this application; Figure 5 This is a structural block diagram of the IoT device provided in the embodiments of this application; Figure 6 A structural block diagram of the electronic device provided in the application embodiment. Detailed Implementation

[0014] In the application embodiments, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0015] In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0017] In a first aspect, embodiments of this application provide a control method for an Internet of Things (IoT) device, which is applied to an IoT device including a main processor, a coprocessor, and a voice wake-up module. The power consumption of the voice wake-up module is less than a first threshold during a first time period. The voice wake-up module is communicatively connected to the coprocessor, and the coprocessor is communicatively connected to the main processor.

[0018] The main processor undertakes complex computational tasks, such as handling various application logics and user interactions; the main processor may be, for example, an application processing unit (AP).

[0019] The coprocessor can be responsible for sensor data acquisition and equipment status monitoring, and can control the main processor to power down when appropriate; the coprocessor can be, for example, an MCU. MCU can also be called a single-chip microcomputer or a microcontroller.

[0020] The voice wake-up module can be a voice activity detection (VAD) chip, where voice activity detection, also known as speech activity detection or speech detection, is a technology that detects whether human speech is present.

[0021] Furthermore, in existing technologies, standby voice solutions for IoT devices often rely on the low-power island of the main processor or a general-purpose MCU for monitoring, resulting in high system coupling and difficulty in achieving extreme power consumption optimization. In this application, in some embodiments, the voice wake-up module can be independent of the main processor and coprocessor, i.e., an "external" voice wake-up module dedicated to voice monitoring. This evolves the IoT device from a traditional "dual system of main processor + coprocessor" to a "triple system of main processor + coprocessor + voice wake-up module," achieving functional decoupling at the hardware level. Moreover, in this process of transforming from a dual system to a triple system, only the power supply to this wake-up module needs to be considered, without requiring significant modifications to the original working logic of the dual systems. This simplifies power management design and enables fundamental optimization at the system architecture level.

[0022] Alternatively, in some embodiments, the voice wake-up module can be integrated into a coprocessor. For example, the coprocessor has a coprocessor domain whose power consumption is less than a first threshold during a first time period and is always on, specifically for running the voice wake-up module, with an effect similar to the "plug-in" scenario described above.

[0023] In addition, the IoT devices can be smart headphones, smart remote controls, smart home control screens, smartwatches, smart glasses, etc.

[0024] like Figure 1 As shown, the control method for this IoT device may include the following steps 101 to 103: Step 101: The voice wake-up module acquires the first audio data.

[0025] The voice wake-up module can acquire the first audio data when neither the main processor nor the coprocessor is in sleep mode; alternatively, it can acquire the first audio data when at least one of the main processor and the coprocessor is in sleep mode. Therefore, in some embodiments, a dedicated hardware unit (i.e., a voice wake-up module) can be added to the IoT device. This transforms the voice monitoring entity from the main processor or coprocessor in sleep mode into a dedicated low-power voice wake-up module, resulting in a significant leap in power efficiency.

[0026] In addition, the main processor being in a hibernation state can refer to the main processor being completely powered down or in the deepest hibernation state, with extremely low current consumption (e.g., below the second threshold). In the deepest hibernation state, the processor will shut down unnecessary power supplies (e.g., turn off the core clock) and significantly reduce the voltage, retaining only critical state information to achieve rapid wake-up.

[0027] When a coprocessor is in a sleep state, it can mean that the coprocessor is in a deep sleep state, retaining only the data in the random access memory (RAM), turning off most of the clock, and waiting for an external interrupt to wake it up.

[0028] Optionally, the IoT device is equipped with a microphone, and the voice wake-up module is communicatively connected to the microphone; the voice wake-up module acquires first audio data, including: The voice wake-up module acquires the first audio data collected through the microphone.

[0029] For example, the voice wake-up module can use a pulse density modulation (PDM) interface to communicate with the microphone.

[0030] As can be seen, in some embodiments, the voice wake-up module can communicate with a microphone to obtain the first audio data collected by the microphone.

[0031] Step 102: The voice wake-up module obtains the first keyword and the type indication information of the first keyword in the first audio data.

[0032] The voice wake-up module can preload a neural network model so that after the voice wake-up module obtains the first audio data, it can use the neural network model to obtain the first keyword and the type indication information of the first keyword in the first audio data.

[0033] In addition, the type indication information of the first keyword may include at least one of the type label of the first keyword and the keyword identifier (ID).

[0034] Step 103: The voice wake-up module wakes up the coprocessor and / or the main processor that are in a dormant state according to the type indication information of the first keyword.

[0035] Different types of keywords can correspond to different wake-up objects. For example, the wake-up object corresponding to the first type of keyword includes the coprocessor. In this case, if the first type of keyword exists in the first audio data obtained by the voice wake-up module, only the coprocessor needs to be woken up. The wake-up object corresponding to the second type of keyword includes both the coprocessor and the main processor. In this case, if the second type of keyword exists in the first audio data obtained by the voice wake-up module, both the coprocessor and the main processor need to be woken up.

[0036] As can be seen from steps 101 to 103 above, in the embodiments of this application, a voice wake-up module with power consumption less than a first threshold during a first time period can be set in the IoT device, and the voice wake-up module is communicatively connected to the coprocessor of the IoT device, and the coprocessor is communicatively connected to the main processor; in this way, the voice wake-up module can obtain the first audio data, thereby obtaining the first keyword and the type indication information of the first keyword in the first audio data, and then the voice wake-up module can wake up the coprocessor and / or the main processor in a dormant state according to the type indication information of the first keyword.

[0037] As can be seen, in the embodiments of this application, a dedicated hardware unit (i.e., a voice wake-up module) is added to the IoT device. This transforms the voice monitoring entity from the main processor or coprocessor in sleep mode into a dedicated low-power voice wake-up module, resulting in a significant leap in power consumption. Furthermore, after the voice wake-up module acquires the audio data, it first identifies the type of keywords in the audio data and then determines the wake-up target based on the keyword type, achieving refined power consumption management. Therefore, the embodiments of this application can achieve ultra-low standby power consumption while ensuring that users can seamlessly interact with the device via voice at any time.

[0038] For example, when the IoT device is smart glasses, a voice wake-up module (such as a VAD chip) can be added to the smart glasses. This voice wake-up module is then connected to a coprocessor within the smart glasses, which in turn is connected to the main processor. In this way, the voice wake-up module can acquire first audio data, thereby obtaining a first keyword and its type indication information. Based on the keyword's type indication information, it can then wake up the main processor and / or the coprocessor from their dormant states. This achieves ultra-low standby power consumption for the smart glasses while ensuring seamless user interaction via voice at any time.

[0039] Optionally, in step 103 above, the voice wake-up module wakes up the coprocessor and / or the main processor in a dormant state according to the type indication information of the first keyword, including one of the following steps A-1 to A-3: Step A-1: ​​When the coprocessor is in a sleep state and the voice wake-up module determines that the first keyword includes a keyword of the first type according to the type indication information, the voice wake-up module sends a first wake-up signal to the coprocessor, and the coprocessor responds to the first wake-up signal to exit the sleep state; The first keyword includes a first type of keyword, indicating that the first type of keyword exists in the first audio data. Therefore, in some embodiments, if the first type of keyword exists in the first audio data obtained by the voice wake-up module when the coprocessor is in a sleep state, only the coprocessor can be woken up.

[0040] In addition, the first type of keyword refers to the keyword of the instruction corresponding to the operation that the coprocessor can execute.

[0041] Step A-2: When both the coprocessor and the main processor are in sleep mode, and the voice wake-up module determines that the first keyword includes a keyword of the second type based on the type indication information, the voice wake-up module sends the first wake-up signal and first information to the coprocessor. The coprocessor responds to the first wake-up signal to exit the sleep mode and sends a second wake-up signal to the main processor based on the first information. The main processor responds to the second wake-up signal to exit the sleep mode. The first information is used to instruct the main processor to wake up. The first keyword includes a second type of keyword, indicating that the first audio data contains a second type of keyword. Therefore, in some embodiments, if the voice wake-up module obtains the first audio data containing a second type of keyword when both the coprocessor and the main processor are in a sleep state, it may be necessary to wake up the coprocessor and the main processor.

[0042] Furthermore, the second type of keyword refers to keywords corresponding to instructions that the main processor can execute. Therefore, if the first audio data acquired by the voice wake-up module contains a keyword of the second type, it indicates that the main processor needs to perform the corresponding operation, thus requiring the main processor to be woken up. Since the voice wake-up module communicates with the coprocessor, and the coprocessor communicates with the main processor, when the main processor needs to be woken up, the coprocessor must first be woken up. Therefore, in step A-2, the voice wake-up module needs to wake up the coprocessor, and then the coprocessor wakes up the main processor.

[0043] Step A-3: When the coprocessor is not in a sleep state and the main processor is in a sleep state, and the voice wake-up module determines that the first keyword includes the second type of keyword according to the type indication information, the voice wake-up module sends the first information to the coprocessor, the coprocessor sends the second wake-up signal to the main processor according to the first information, and the main processor responds to the second wake-up signal to exit the sleep state.

[0044] In step A-3, the coprocessor is not in a sleep state, only the main processor is in a sleep state. Therefore, if the first audio data obtained by the voice wake-up module contains keywords of the second type, the voice wake-up module can notify the coprocessor to wake up the main processor.

[0045] Optionally, the first information includes type indication information of the first keyword; in step A-2 or A-3 above, the coprocessor sends a second wake-up signal to the main processor based on the first information, including: If the coprocessor determines, based on the type indication information, that the first keyword includes a keyword of the second type, the coprocessor sends a second wake-up signal to the main processor.

[0046] As can be seen, in some embodiments, if the voice wake-up module needs to notify the coprocessor to wake up the main processor, the voice wake-up module can send the type indication information of the first keyword it has obtained to the coprocessor. The coprocessor can then determine whether there is a second type of keyword in the first keyword based on the type indication information. If there is, the coprocessor can send a second wake-up signal to the main processor.

[0047] Optionally, the first information includes first indication information; the first indication information is used to indicate that: there is a second type of keyword in the first keyword, or to indicate: wake up the main processor; in the above steps A-2 or A-3, the coprocessor sends a second wake-up signal to the main processor according to the first information, including: the coprocessor sends a second wake-up signal to the main processor according to the first indication information.

[0048] As can be seen, in some embodiments, if the voice wake-up module needs to notify the coprocessor to wake up the main processor, the voice wake-up module can directly send the aforementioned first instruction information to the coprocessor to inform the coprocessor that "a second type of keyword exists in the first keyword", or directly notify the coprocessor to wake up the main processor, so that the coprocessor can send a second wake-up signal to the main processor.

[0049] Optionally, if the first keyword includes keywords of the first type, the method further includes: The voice wake-up module sends the first audio data to the coprocessor; The coprocessor receives the first audio data and generates a first operation instruction based on the first audio data; The coprocessor responds to the first operation instruction and executes the operation corresponding to the first operation instruction.

[0050] In the case where the voice wake-up module and the coprocessor are connected via a first data bus, the voice wake-up module sends first audio data to the coprocessor via the first data bus. The first data bus can be one of the following: a Serial Peripheral Interface (SPI) bus, an Inter-IC Sound (I2S) bus, or a PDM.

[0051] Therefore, in some embodiments, if the first keyword includes a keyword of the first type, the voice wake-up module can also send the acquired first audio data to the coprocessor, which then generates a corresponding first operation instruction based on the first audio data and executes the operation corresponding to the first operation instruction. Here, the voice wake-up module does not only send the acquired first keyword to the coprocessor, but sends the complete first audio data to the coprocessor. In this way, the coprocessor can obtain a more accurate first operation instruction based on the complete first audio data, preventing false triggering.

[0052] Optionally, if the first keyword includes keywords of the second type, the method further includes: The voice wake-up module sends the first audio data to the coprocessor; The coprocessor receives the first audio data and sends the first audio data to the main processor; The main processor receives the first audio data and generates a second operation instruction based on the first audio data; The main processor responds to the second operation instruction and executes the operation corresponding to the second operation instruction.

[0053] In the case where the coprocessor and the main processor are connected via a second data bus, the coprocessor sends the first audio data to the main processor through the second data bus. The second data bus can be a high-speed data bus.

[0054] Therefore, in some embodiments, if the first keyword includes a second type of keyword, the voice wake-up module can also send the acquired first audio data to the coprocessor, which then sends the first audio data to the main processor. The main processor then generates a corresponding second operation instruction based on the first audio data and executes the operation corresponding to the second operation instruction. Here, instead of just sending the first keyword to the main processor, the complete first audio data is sent to the main processor. This allows the main processor to obtain a more accurate second operation instruction based on the complete first audio data, preventing accidental triggering.

[0055] Optionally, the voice wake-up module is provided with a first pin, and the coprocessor is provided with a second pin, the first pin and the second pin being electrically connected; the voice wake-up module sends a first wake-up signal to the coprocessor, including: The voice wake-up module sets the first pin to a first preset level; wherein, the signal at the first preset level serves as the first wake-up signal; The coprocessor responds to the first wake-up signal and exits the sleep state, including: The coprocessor detects the first preset level on the second pin and exits the sleep state.

[0056] Therefore, in some embodiments, the first pin of the voice wake-up module can be electrically connected to the second pin of the coprocessor. When the voice wake-up module sets the first pin to a first preset level, the coprocessor can detect the first preset level of its second pin and thus exit the sleep state. In this way, waking up the coprocessor through a hardware-level interrupt mechanism has extremely fast response characteristics and does not require upper-level software intervention, significantly reducing wake-up latency.

[0057] The first and second pins mentioned above can both be general-purpose input / output (GPIO) pins.

[0058] Optionally, the coprocessor is provided with a third pin, and the main processor is provided with a fourth pin, the third pin and the fourth pin being electrically connected; the coprocessor sends a second wake-up signal to the main processor, including: The coprocessor sets the third pin to a second preset level; wherein the signal of the second preset level serves as the second wake-up signal; The main processor responds to the second wake-up signal and exits the sleep state, including: The main processor detects the second preset level of the fourth pin and exits the sleep state.

[0059] Therefore, in some embodiments, the third pin of the coprocessor can be electrically connected to the fourth pin of the main processor. When the coprocessor sets the first pin to a second preset level, the main processor can detect the second preset level of its fourth pin and thus exit the sleep state. In this way, waking up the main processor through a hardware-level interrupt mechanism provides extremely fast response characteristics without the need for upper-level software intervention, significantly reducing wake-up latency.

[0060] The third and fourth pins mentioned above can both be GPIO pins.

[0061] Optionally, the method further includes: If the voice wake-up module determines that the first keyword includes a keyword of the third type according to the type indication information, but the first keyword does not include keywords of the first type and the second type, and no second audio data is obtained within a predetermined time period after the first moment, the voice wake-up module ignores the keyword of the third type. Wherein, the first moment is the moment when the voice wake-up module acquires the first audio data.

[0062] The first keyword includes a third type of keyword, indicating that the first audio data obtained by the voice wake-up module contains a third type of keyword.

[0063] As can be seen, in some embodiments, if the first audio data obtained by the voice wake-up module contains a third type of keyword, but not the first or second type of keyword, and the second audio data is not obtained within a predetermined time period after the voice wake-up module obtains the first audio data, then it means that it is not necessary to wake up either the main processor or the coprocessor; in this case, the third type of keyword can be understood as a wake-up word, indicating that it is necessary to start listening to the words of subsequent keywords.

[0064] Optionally, the voice wake-up module acquires a first keyword and type indication information of the first keyword from the first audio data, including: The voice wake-up module obtains the first keyword and the type indication information based on the predetermined second information; wherein, the second information includes at least one keyword and the type indication information of the keyword.

[0065] The keywords included in the second information can be divided into different lists according to type, or the keywords included in the second information can all be in the same list.

[0066] As can be seen, in some embodiments, at least one keyword and type indication information of these keywords can be predetermined and stored in the voice wake-up module; in this way, after the voice wake-up module obtains the first audio data, it can identify the first keyword and the type indication information of the first keyword in the first audio data based on the stored keywords and the corresponding type indication information.

[0067] In addition, in some embodiments, the voice wake-up module can convert the first audio data into text information through a preloaded neural network model, and then perform word segmentation on the text information, converting each word into a corresponding feature vector, and then comparing it with the feature vectors of each keyword in the second information, for example, calculating the distance between the feature vectors. The closer the distance, the better the match, and then outputting the matched first keyword and the type indication information of the first keyword.

[0068] Optionally, the method further includes: When the IoT device is detected to be in a first predetermined state, the voice wake-up module disables the second type of keywords in the second information; When the IoT device is detected to be in a second predetermined state, the voice wake-up module restores the second type of keywords in the second information.

[0069] As can be seen, in some embodiments, the state of the IoT device can be used to determine whether to disable the second type of keywords in the second information. For example, in the first predetermined state, the second type of keywords in the second information can be disabled, so that the voice wake-up module will not recognize the second type of keywords in the audio data it acquires, and thus will not wake up the main processor in the sleep state. For example, in the second predetermined state, the second type of keywords in the second information can be restored, so that the voice wake-up module can recognize the second type of keywords in the audio data it acquires, and thus can wake up the main processor in the sleep state.

[0070] It should be noted that in both the first and second predetermined states, the first type of keyword in the second information is always available. Thus, regardless of whether it is in the first or second predetermined state, the voice wake-up module can always recognize the first type of keyword in the audio data it has acquired, thereby waking up the coprocessor that is in a dormant state.

[0071] For example, when the IoT device is smart glasses, if the wear sensor detects that the smart glasses have been removed, the second type of keywords in the second information can be disabled, but the first type of keywords in the second information can be retained. In this way, when the smart glasses are removed, only the coprocessor in a dormant state can be woken up. When the wear sensor detects that the smart glasses have been worn, the second type of keywords in the second information can be restored. In this way, when the smart glasses are worn, the coprocessor and main processor in a dormant state can be woken up. Thus, further power consumption optimization can be achieved without affecting the core experience.

[0072] The following describes two specific methods for disabling and restoring keywords of the second type in the second information: Method 1: When the IoT device is detected to be in a first predetermined state, the keywords of the second type in the second information are moved to the first storage area; When the IoT device is detected to be in a second predetermined state, the keywords of the second type stored in the first storage area are added to the first information.

[0073] Method 2: Each keyword in the second information has a flag. When a keyword is disabled, its corresponding flag can be set to "disabled"; when a keyword is restored, its corresponding flag can be set to "available".

[0074] To facilitate understanding of the control method for the IoT device in this application embodiment, a smart glasses example is used to describe the specific implementation of the control method for the IoT device in this application embodiment: The smart glasses adopt a three-system hardware architecture, the core of which lies in "functional decoupling" and "power domain independence"; the three main computing units of this architecture and their roles are as follows: AP main core: As the "brain", it is usually a high-performance application processor; it runs a complete operating system and complex applications, but consumes the most power; when in standby mode, most of its functional modules are completely powered down, and only a very small amount of logic is retained for being woken up.

[0075] MCU chip: As the "nerve center", it is a low-power microcontroller; it manages Bluetooth connection, performs preliminary fusion of data from multiple sensors, and has simple offline voice recognition capabilities; it has its own independent memory and program storage, and can run independently when the AP main core is powered off; when in standby mode, it is in a deep sleep state.

[0076] The VAD chip, acting as a "dedicated ear," is an ultra-low-power application-specific integrated circuit or digital signal processor (DSP). It has a built-in small keyword list, including third-type keywords (such as "Hello, XX"), first-type keywords (such as keywords for simple commands: "next song", "play"), second-type keywords (such as keywords for complex commands: "today's weather"), and type labels for each keyword. It can also preload neural network models. In addition, it is usually powered by an always-on power domain to ensure that it can continue to work even if all other systems are powered off.

[0077] The connection relationship between the above three computing units can be described as follows: Figure 2 As shown: The VAD chip communicates with the microphone of the smart glasses; The GPIP pin of the VAD chip is electrically connected to one GPIO pin of the MCU chip, and the other GPIO pin of the MCU chip is electrically connected to the GPIO pin of the AP main core. The MCU chip and the AP main core communicate with each other via a Universal Asynchronous Receiver / Transmitter (UART) and an Inter-IC Sound (I2S) bus. The AP core can also communicate with cloud servers.

[0078] Depend on Figure 2As can be seen, the three-system hardware architecture of the smart glasses achieves efficient collaboration through the following three key signal paths: Regarding the wake-up signal line, the VAD chip is directly connected to the external wake-up pin of the MCU chip via GPIO pins, and the MCU chip's GPIO pins are directly connected to the external wake-up pin of the AP main core, forming a hardware-level interrupt mechanism. This design features extremely fast response characteristics and requires no upper-layer software intervention, significantly reducing wake-up latency. The audio data channel uses a PDM interface to directly connect to the digital microphone. When keyword recognition is triggered, the VAD chip seamlessly transmits the cached keywords and related audio data streams to the woken-up system through a shared memory area, ensuring the integrity and continuity of subsequent voice processing. The control bus uses UART to realize the transmission of configuration information and command interaction between the three parties, completing the collaborative control between the systems. These three paths together constitute a clearly hierarchical and functionally decoupled hardware collaboration architecture.

[0079] In addition, the VAD chip can perform voice detection and identify keywords in the audio data collected by the microphone.

[0080] The MCU chip also communicates with the microphone and speaker of the smart glasses, enabling music and other audio data to be played through the speaker and Bluetooth calls to be made through the microphone and speaker; in addition, the MCU chip can also perform noise reduction / echo cancellation processing on the audio segments captured by the microphone.

[0081] The AP main core can work with other modules or chips in smart glasses to achieve AI functions, take photos, issue commands, etc.

[0082] The following is combined with Figure 3 This section introduces two typical application scenarios for the aforementioned smart glasses: Scenario 1: Offline control with ultimate energy efficiency; User experience: The user takes off the smart glasses and places them on the table. The screen is off and the smart glasses are in standby mode. A few minutes later, the user wants to change the song and says "next song". The user observes that the indicator light on the smart glasses flashes slightly, and the song on the phone changes immediately, but the AP main core of the smart glasses does not start.

[0083] User experience value: extremely fast response (no noticeable delay), seamless operation (no need to touch the device or wait for the screen to light up), and this operation is very power-efficient.

[0084] Scenario 2: A seamlessly integrated smart assistant; User experience: In the same standby state, the user directly asks "Today's weather"; the smart glasses immediately broadcast the current weather information and future forecast.

[0085] User experience value: The experience is seamless and natural, as if the device is always attentively waiting for instructions, breaking the traditional perception that "standby = no response".

[0086] The following section details the specific processing flow of smart glasses for the two scenarios mentioned above. Figure 4 As shown: (1) Standby state (starting point): AP main core: Completely powered down or in the deepest sleep state, with extremely low current consumption.

[0087] MCU chip: In deep sleep, only RAM data is retained, most clocks are turned off, and it is waiting for an external interrupt to wake it up.

[0088] VAD chip: running at full power.

[0089] (2) Keyword identification and classification: The VAD chip contains not just a single wake word, but a small list of keywords, and each keyword has a type tag (e.g., TYPE_OFFLINE_CONTROL, TYPE_CLOUD_QUERY).

[0090] The internal workflow of the VAD chip is a loop of "acquisition -> feature extraction -> model inference". That is, it continuously acquires audio data from the microphone, converts it into feature vectors through a preloaded neural network model, and matches them with the feature vectors of keywords in its internally stored keyword list, thereby outputting the matched keywords and their corresponding type labels.

[0091] As can be seen, when audio data matches a keyword in the keyword list, the VAD chip outputs not only "event detected", but also the type label of that keyword.

[0092] (3) Hierarchical wake-up and data transfer (i.e., collaborative working mechanism): In Scenario 1: When the VAD chip recognizes "Next Song" (type TYPE_OFFLINE_CONTROL), it executes the following two processes: "Trigger" and "Transfer": Trigger: Pull the GPIO interrupt line connected to the MCU chip; that is, set the GPIO pin of the VAD chip to the first preset level so that the GPIO pin of the MCU chip can detect the first preset level, thereby exiting the sleep state; Transmission: The complete audio data packet containing the "next song" instruction is sent to the MCU chip via SPI.

[0093] When the MCU chip is woken up, it loads a lightweight speech recognition engine (such as one based on Dynamic Time Warping (DTW) or a small-scale neural network) from memory to perform secondary parsing and confirmation of the audio data, preventing false triggering. After confirmation, it sends remote control commands (i.e., next track) to the connected mobile phone via Bluetooth. During this process, the AP main core remains in sleep mode.

[0094] In scenario two: if the VAD chip recognizes "Today's Weather" (type TYPE_CLOUD_QUERY), it will execute the following "triggering" and "transmission" process: Trigger 1: Pull the GPIO interrupt line connected to the MCU chip; that is, set the GPIO pin of the VAD chip to the first preset level so that the GPIO pin of the MCU chip can detect the first preset level, thereby exiting the sleep state; Transmission 1: Send the complete audio data packet containing the "Today's Weather" command to the MCU chip via SPI; Trigger 2: Pull the GPIO interrupt line connected to the AP main core (or wake up via PMIC); that is, set the GPIO pin of the MCU chip to the second preset level so that the GPIO pin of the AP main core can detect the second preset level, thereby exiting the sleep state; Transmission 2: The complete audio data containing the "Today's Weather" command will be sent to the AP main core via a high-speed interface.

[0095] In this process, the AP main core powers on and boots up, the operating system loads, and the full-featured voice assistant is activated, receiving audio data for cloud communication and natural language processing. Although this path consumes a lot of power, it is used to handle complex tasks, making the best use of resources.

[0096] In addition, the voice wake-up mechanism under the three-system architecture in this embodiment can also be combined with the operating strategy of dual-system sensors (e.g., power off after 5 / 20 minutes of removing the glasses). For example, when the wear sensor detects that the smart glasses have been removed, the system can dynamically adjust the keyword list of the VAD chip, such as disabling keywords corresponding to instructions that need to be processed by the AP main core and only retaining keywords corresponding to instructions that need to be processed by the MCU chip, thereby achieving further power consumption optimization without affecting the core experience.

[0097] In addition, in existing technologies, the following three solutions are mainly used to reduce the power consumption of smart glasses: 1. It uses a low-power "small core (i.e., coprocessor)" system to handle standby tasks, sensor data, and simple communication, and wakes up a high-performance "big core (i.e., main processor)" system when needed; its power-saving logic is mostly based on physical sensors (such as wear detection, gyroscope, and light sensor).

[0098] 2. Low-power voice wake-up based on the main processor: This involves creating a low-power island on the application processor (or using a coprocessor) to run the voice detection algorithm at a lower frequency and voltage during standby. This method is still within the power domain of the application processor, and the power consumption reduction is limited.

[0099] 3. Voice wake-up based on general-purpose MCU: This means that a separate, low-power MCU is responsible for voice monitoring and wake-up. This MCU usually also performs other tasks (such as sensor hub).

[0100] It is evident that current smart glasses technologies largely rely on dual-system solutions for standby power consumption optimization. Their standby voice wake-up is either slow to respond (requiring a cold start from deep sleep) or consumes a lot of power (the main processor is not completely powered off), resulting in a compromise between their advertised standby time and instant response capability.

[0101] In this embodiment, a third ultra-low-power dedicated voice wake-up chip (VAD) is innovatively introduced to construct a three-system collaborative architecture. This VAD chip acts as a dedicated "ear" for the device in standby mode, continuously listening to voice commands even when the main processor and coprocessor are completely asleep. It possesses preliminary voice recognition and intelligent routing capabilities, accurately determining the target to be woken up. This design completely decouples "continuous sensing" from "high-power operation" at the system level, achieving extremely low standby power consumption while ensuring immediate response to voice interaction.

[0102] As can be seen, the three-system architecture of this application embodiment can achieve the following breakthroughs through hardware division of labor and intelligent routing mechanism: (1) Reduce standby power consumption: The power consumption of VAD chips (usually in the milliwatt or even microwatt level) is much lower than that of an ARM Cortex-M series MCU in the listening state (tens of milliwatts level); among them, the ARM Cortex-M series is a low-power, high-efficiency microcontroller core designed for embedded systems and IoT devices.

[0103] (2) Improved response latency: In standby mode, both the main processor and coprocessor are in hibernation mode to save power; at this time, the device cannot respond to the user's voice commands, thus losing the interactive convenience of smart glasses. If voice is required to respond at any time, some functions of the main processor and coprocessor must be kept running, which will lead to increased standby power consumption, which contradicts the original intention of saving power; however, in the embodiments of this application, dedicated hardware is used to process voice wake-up, which is faster, and the hierarchical wake-up mechanism avoids the waste of power by "using a sledgehammer to crack a nut".

[0104] (3) Optimize user experience: Users can interact with the device instantly through voice commands at any time (even after taking off their glasses). This feeling is "always online", which is better than the experience of "may need to wait" or "sometimes unresponsive".

[0105] (4) Existing standby voice solutions mostly rely on the low-power island of the main processor or a general-purpose MCU for listening, resulting in high system coupling and difficulty in achieving extreme power consumption optimization. In the embodiments of this application, the "external VAD chip + intelligent router" architecture not only completely decouples the voice wake-up function from the main system, but also simplifies the power management design, achieving fundamental optimization at the system architecture level.

[0106] The control method for the IoT device provided in the application embodiments has been described above. The IoT device provided in the application embodiments will be described below with reference to the accompanying drawings.

[0107] Secondly, the application embodiments also provide an Internet of Things (IoT) device, such as... Figure 5 As shown, the IoT device includes: a main processor 501, a coprocessor 502, and a voice wake-up module 503. The power consumption of the voice wake-up module 503 is less than a first threshold during a first time period. The voice wake-up module 503 is communicatively connected to the coprocessor 502, and the coprocessor 502 is communicatively connected to the main processor 501. The voice wake-up module 503 is used for: Obtain the first audio data; Obtain the first keyword and the type indication information of the first keyword from the first audio data; Based on the type indication information of the first keyword, wake up the coprocessor 502 and / or the main processor 501 that are in a dormant state.

[0108] Optionally, when the voice wake-up module 503 wakes up the coprocessor and / or the main processor in a dormant state according to the type indication information of the first keyword, it is specifically used for: When the coprocessor 502 is in a sleep state, and the voice wake-up module 503 determines, based on the type indication information, that the first keyword includes a keyword of the first type, the voice wake-up module 503 sends a first wake-up signal to the coprocessor 502, and the coprocessor 502 responds to the first wake-up signal to exit the sleep state; or... When both the coprocessor 502 and the main processor 501 are in a sleep state, and the voice wake-up module 503 determines, based on the type indication information, that the first keyword includes a keyword of the second type, the voice wake-up module 503 sends a first wake-up signal and first information to the coprocessor 502. The coprocessor 502 responds to the first wake-up signal and exits the sleep state, and sends a second wake-up signal to the main processor 501 based on the first information. The main processor 501 responds to the second wake-up signal and exits the sleep state. The first information is used to instruct the main processor 501 to be woken up, or... When the coprocessor 502 is not in a sleep state, the main processor 501 is in a sleep state, and the voice wake-up module 503 determines that the first keyword includes the keyword of the second type according to the type indication information, the voice wake-up module 503 sends the first information to the coprocessor 502, the coprocessor 502 sends the second wake-up signal to the main processor 501 according to the first information, and the main processor 501 responds to the second wake-up signal to exit the sleep state.

[0109] Optionally, if the first keyword includes a keyword of the first type, the voice wake-up module 503 is further configured to: send the first audio data to the coprocessor 502; The coprocessor 502 is further configured to: receive the first audio data and generate a first operation instruction based on the first audio data; The coprocessor 502 is further configured to: respond to the first operation instruction and execute the operation corresponding to the first operation instruction.

[0110] Optionally, if the first keyword includes a keyword of the second type, the voice wake-up module 503 is further configured to: send the first audio data to the coprocessor 502; The coprocessor 502 is further configured to: receive the first audio data and send the first audio data to the main processor 501; The main processor 501 is further configured to: receive the first audio data and generate a second operation instruction based on the first audio data; The main processor 501 is further configured to: respond to the second operation instruction and execute the operation corresponding to the second operation instruction.

[0111] Optionally, the voice wake-up module 503 is provided with a first pin, and the coprocessor 502 is provided with a second pin, wherein the first pin and the second pin are electrically connected; When the voice wake-up module 503 sends the first wake-up signal to the coprocessor 502, it is specifically used for: The voice wake-up module 503 sets the first pin to a first preset level; wherein, the signal of the first preset level serves as the first wake-up signal; When the coprocessor 502 responds to the first wake-up signal and exits the sleep state, it is specifically used for: The coprocessor 502 detects the first preset level of the second pin and exits the sleep state.

[0112] Optionally, the coprocessor 502 is provided with a third pin, and the main processor 501 is provided with a fourth pin, wherein the third pin and the fourth pin are electrically connected; When the coprocessor 502 sends the second wake-up signal to the main processor 501, it is specifically used for: The coprocessor 502 sets the third pin to a second preset level; wherein the signal of the second preset level serves as the second wake-up signal; When the main processor 501 responds to the second wake-up signal and exits the sleep state, it is specifically used for: The main processor 501 detects the second preset level of the fourth pin and exits the sleep state.

[0113] Optionally, if the voice wake-up module 503 determines that the first keyword includes a third type of keyword based on the type indication information, but the first keyword does not include the first type and the second type of keyword, and no second audio data is acquired within a predetermined time period after the first moment, the voice wake-up module 503 is further configured to: ignore the third type of keyword; The first moment is the moment when the voice wake-up module 503 acquires the first audio data.

[0114] Optionally, when the voice wake-up module 503 obtains the first keyword and the type indication information of the first keyword in the first audio data, it is specifically used for: The voice wake-up module 503 obtains the first keyword and the type indication information according to the predetermined second information; wherein, the second information includes at least one keyword and the type indication information of the keyword.

[0115] Optionally, the voice wake-up module 503 is further configured to: If the IoT device is detected to be in a first predetermined state, the second type of keyword in the second information is disabled; If the IoT device is detected to be in a second predetermined state, the second type of keyword in the second information is restored.

[0116] Optionally, the IoT device is equipped with a microphone, and the voice wake-up module 503 is communicatively connected to the microphone; when the voice wake-up module 503 acquires the first audio data, it is specifically used to: acquire the first audio data collected through the microphone.

[0117] It should be noted that the module division in this embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of this application can be integrated into one module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0118] If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a processor-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 all or 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.) or processor 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.

[0119] It should be noted that the apparatus provided in the application embodiments can implement all the method steps implemented in the above method embodiments and can achieve the same technical effect. Therefore, the parts that are the same as those in the method embodiments and the beneficial effects will not be described in detail here.

[0120] Embodiments of the present invention also provide an electronic device, such as... Figure 6As shown, the electronic device includes a memory 620, a transceiver 610, and a processor 600; Memory 620 is used to store computer programs; Transceiver 610 is used to receive and send data under the control of processor 600; The processor 600 is used to read the computer program in the memory 620 and execute the control method of the Internet of Things device described in the first aspect above.

[0121] Among them, Figure 6 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 610) and memory (memory 620). The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 610 can be multiple elements, including transmitters and receivers, providing units for communicating with various other devices over transmission media, including wireless channels, wired channels, optical fibers, etc. The processor 600 is responsible for managing the bus architecture and general processing, and the memory 620 can store data used by the processor 600 during operation.

[0122] The processor 600 can be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor 600 can also adopt a multi-core architecture.

[0123] It should be noted that the apparatus provided in the application embodiments can implement all the method steps implemented in the above method embodiments and can achieve the same technical effect. Therefore, the parts that are the same as those in the method embodiments and the beneficial effects will not be described in detail here.

[0124] Embodiments of the present invention also provide a processor-readable storage medium storing a computer program for causing the processor to execute the control method for the Internet of Things device described above.

[0125] 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)).

[0126] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0127] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0128] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the processor-readable memory produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These processors can execute instructions that can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0130] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A control method for an Internet of Things (IoT) device, characterized in that, The device is applied to Internet of Things (IoT) devices, which include a main processor, a coprocessor, and a voice wake-up module. The power consumption of the voice wake-up module is less than a first threshold during a first time period. The voice wake-up module is communicatively connected to the coprocessor, and the coprocessor is communicatively connected to the main processor. The method includes: The voice wake-up module acquires the first audio data; The voice wake-up module acquires the first keyword in the first audio data and the type indication information of the first keyword; The voice wake-up module wakes up the coprocessor and / or the main processor that are in a dormant state according to the type indication information of the first keyword.

2. The method according to claim 1, characterized in that, The voice wake-up module wakes up the coprocessor and / or the main processor, which are in a dormant state, according to the type indication information of the first keyword, including: When the coprocessor is in sleep mode and the voice wake-up module determines, based on the type indication information, that the first keyword includes a keyword of the first type, the voice wake-up module sends a first wake-up signal to the coprocessor, and the coprocessor responds to the first wake-up signal to exit sleep mode; or... When both the coprocessor and the main processor are in sleep mode, and the voice wake-up module determines, based on the type indication information, that the first keyword includes a keyword of the second type, the voice wake-up module sends a first wake-up signal and first information to the coprocessor. The coprocessor responds to the first wake-up signal to exit sleep mode and sends a second wake-up signal to the main processor based on the first information. The main processor responds to the second wake-up signal to exit sleep mode. The first information is used to instruct the main processor to wake up, or... When the coprocessor is not in a sleep state and the main processor is in a sleep state, and the voice wake-up module determines that the first keyword includes a keyword of the second type based on the type indication information, the voice wake-up module sends the first information to the coprocessor, the coprocessor sends the second wake-up signal to the main processor based on the first information, and the main processor responds to the second wake-up signal to exit the sleep state.

3. The method according to claim 2, characterized in that, If the first keyword includes a keyword of the first type, the method further includes: The voice wake-up module sends the first audio data to the coprocessor; The coprocessor receives the first audio data and generates a first operation instruction based on the first audio data; The coprocessor responds to the first operation instruction and executes the operation corresponding to the first operation instruction.

4. The method according to claim 2, characterized in that, If the first keyword includes keywords of the second type, the method further includes: The voice wake-up module sends the first audio data to the coprocessor; The coprocessor receives the first audio data and sends the first audio data to the main processor; The main processor receives the first audio data and generates a second operation instruction based on the first audio data; The main processor responds to the second operation instruction and executes the operation corresponding to the second operation instruction.

5. The method according to claim 2, characterized in that, The voice wake-up module is provided with a first pin, and the coprocessor is provided with a second pin. The first pin and the second pin are electrically connected. The voice wake-up module sends a first wake-up signal to the coprocessor, including: The voice wake-up module sets the first pin to a first preset level; wherein, the signal at the first preset level serves as the first wake-up signal; The coprocessor responds to the first wake-up signal and exits the sleep state, including: The coprocessor detects the first preset level on the second pin and exits the sleep state.

6. The method according to claim 2, characterized in that, The coprocessor is provided with a third pin, and the main processor is provided with a fourth pin. The third pin and the fourth pin are electrically connected. The coprocessor sends a second wake-up signal to the main processor, including: The coprocessor sets the third pin to a second preset level; wherein the signal of the second preset level serves as the second wake-up signal; The main processor responds to the second wake-up signal and exits the sleep state, including: The main processor detects the second preset level of the fourth pin and exits the sleep state.

7. The method according to claim 2, characterized in that, The method further includes: If the voice wake-up module determines that the first keyword includes a keyword of the third type according to the type indication information, but the first keyword does not include keywords of the first type and the second type, and no second audio data is obtained within a predetermined time period after the first moment, the voice wake-up module ignores the keyword of the third type. Wherein, the first moment is the moment when the voice wake-up module acquires the first audio data.

8. The method according to any one of claims 1 to 7, characterized in that, The voice wake-up module acquires the first keyword and the type indication information of the first keyword from the first audio data, including: The voice wake-up module obtains the first keyword and the type indication information based on the predetermined second information; wherein, the second information includes at least one keyword and the type indication information of the keyword.

9. The method according to claim 8, characterized in that, The method further includes: When the IoT device is detected to be in a first predetermined state, the voice wake-up module disables the second type of keywords in the second information; When the IoT device is detected to be in a second predetermined state, the voice wake-up module restores the second type of keywords in the second information.

10. An Internet of Things (IoT) device, characterized in that, The IoT device includes a main processor, a coprocessor, and a voice wake-up module. The power consumption of the voice wake-up module is less than a first threshold during a first time period. The voice wake-up module is communicatively connected to the coprocessor, and the coprocessor is communicatively connected to the main processor. The voice wake-up module is used for: Obtain the first audio data; Obtain the first keyword and the type indication information of the first keyword from the first audio data; Based on the type indication information of the first keyword, wake up the coprocessor and / or the main processor that are in a dormant state.

Images