home appliance

Abstract

The application discloses a kind of household appliances, comprising: voice module, for receiving the voice wake-up signal provided by user;Acquisition module, for obtaining the attribute information that influences the propagation of the voice wake-up signal to the household appliance;Control module, the control module is connected with the voice module, the acquisition module, the control module is specifically configured as: according to the attribute information determines wake-up weight coefficient;According to the voice wake-up signal and the wake-up weight coefficient determines the wake-up state of the household appliance.The household appliance can be awakened by dynamically updating the wake-up weight coefficient to determine the wake-up household appliance, thereby avoiding the problem that the household appliance is miswakened or cannot be awakened due to the change of attribute information affecting the propagation effect of the voice wake-up signal, so that the awakened household appliance is more in line with the actual intention of user, and improves user experience.

Description

Technical Field

[0001] This invention relates to the field of home appliance technology, and in particular to a home appliance. Background Technology

[0002] In related technologies, fixed weight coefficients are set based on experience to determine whether a home appliance is a user's voice wake-up device. However, the weight coefficients do not take into account that the propagation effect of the voice wake-up signal will be affected by different parameters. As a result, a single weight strategy is difficult to adapt to the recognition of home appliances as wake-up devices under different parameter changes. It can also lead to the problem that different parameter changes affect the propagation effect of the voice wake-up signal, resulting in home appliances being falsely woken up or unable to be woken up, thus reducing the user experience. Summary of the Invention

[0003] This invention aims to at least solve one of the technical problems existing in the prior art. Therefore, one objective of this invention is to provide a home appliance that can determine the wake-up time by dynamically updating a wake-up weight coefficient. This avoids the problem of the home appliance being mistakenly woken up or unable to be woken up due to changes in attribute information affecting the propagation effect of the voice wake-up signal, thus making the woken-up home appliance more in line with the user's actual wishes and improving the user experience.

[0004] A first aspect of the present invention provides a home appliance, comprising: a voice module for receiving a voice wake-up signal provided by a user; an acquisition module for acquiring attribute information affecting the propagation of the voice wake-up signal to the home appliance; and a control module connected to the voice module and the acquisition module, wherein the control module is specifically configured to: determine a wake-up weight coefficient based on the attribute information; and determine the wake-up state of the home appliance based on the voice wake-up signal and the wake-up weight coefficient.

[0005] According to embodiments of the present invention, the wake-up weight coefficient of the home appliance is dynamically updated by changes in attribute information. The dynamically updated wake-up weight coefficient optimizes the matching degree of the voice wake-up signal with the user's expected wake-up signal, thereby eliminating the influence of attribute information on the propagation of the voice wake-up signal. Therefore, compared with the prior art that uses a single weight to determine the wake-up state of the wake-up device, the present application uses a dynamically changing wake-up weight coefficient to determine the wake-up state of the home appliance. This allows the dynamically changing wake-up weight system to adapt to the recognition of the wake-up device under different attribute information. Furthermore, the wake-up weight coefficient is dynamically updated as the attribute information of the home appliance changes, thereby avoiding the problem of the home appliance being mistakenly woken up or unable to be woken up due to the influence of attribute information changes on the propagation effect of the voice wake-up signal. This makes the woken-up home appliance more in line with the user's actual wishes and improves the user experience.

[0006] In some embodiments, the wake-up weight coefficient includes a wake-up static weight coefficient and a wake-up dynamic weight coefficient. For determining the wake-up static weight coefficient based on the attribute information, the control module is specifically configured to: determine the wake-up static weight coefficient and / or the wake-up dynamic weight coefficient based on the attribute information.

[0007] The above technical solution has the following advantages or beneficial effects: it dynamically updates the wake-up static weight coefficient and wake-up dynamic weight coefficient according to the changes in attribute information, and can update the wake-up static weight coefficient and wake-up dynamic weight coefficient corresponding to different types of attribute information according to the degree of influence of different types of attribute information on the voice wake-up signal, so that the wake-up static weight coefficient and wake-up dynamic weight coefficient can accurately reflect the influence of different types of attribute information on the propagation of the voice wake-up signal.

[0008] In some embodiments, the attribute information includes environmental state information of the home appliance, performance information of the home appliance, and user behavior information. A wake-up static weight coefficient is determined based on the attribute information. The control module is specifically configured to determine the wake-up static weight coefficient based on one or more of the environmental state information, the performance information, and the user behavior information.

[0009] The above technical solution has the following advantages or beneficial effects: when determining the wake-up state of home appliances by using the wake-up static weight coefficient and voice wake-up signal, it can adapt to the usage needs of different users for home appliances with the same wake-up word, and can also avoid the problem of home appliances being accidentally woken up due to changes in performance information and environmental status information. Moreover, the wake-up static weight coefficient is optimized based on user behavior information, thereby avoiding the problem of rigid wake-up of home appliances, so that the woken-up home appliances are more in line with the actual wishes of users and improve the user experience.

[0010] In some embodiments, the attribute information includes environmental state information of the home appliance, performance information of the home appliance, and user behavior information. The wake-up dynamic weight coefficient is determined based on the attribute information. The control module is specifically configured to determine the wake-up dynamic weight coefficient based on one or more of the environmental state information, the performance information, and the user behavior information.

[0011] The above technical solution has the following advantages or beneficial effects: when determining the wake-up state of home appliances by using the wake-up dynamic weight coefficient and voice wake-up signal, it can adapt to the usage needs of different users for home appliances with the same wake-up word, and can also avoid the problem of home appliances being accidentally woken up due to changes in performance information and environmental status information. Moreover, the wake-up dynamic weight coefficient is optimized based on user behavior information, thereby avoiding the problem of rigid wake-up of home appliances, so that the woken-up home appliances are more in line with the actual wishes of users and improve the user experience.

[0012] In some embodiments, a wake-up weight coefficient is determined based on the attribute information. Specifically, the control module is configured to determine the wake-up weight coefficient based on the attribute information using an adaptive model, wherein the adaptive model is trained on multiple attribute information.

[0013] The above technical solution has the following advantages or beneficial effects: the adaptive model can continuously learn the attribute information that changes in real time, which greatly improves the accuracy of the wake-up weight coefficient output by the adaptive model. Moreover, by continuously optimizing the recognition of wake-up devices through the adaptive model, it is possible to more flexibly and accurately determine the unique home appliance to be woken up from multiple home appliances.

[0014] In some embodiments, the home appliance further includes a communication module connected to the control module and a user terminal device, used to acquire static wake-up feature information about the home appliance provided by the user terminal device; for determining the wake-up state of the home appliance based on the voice wake-up signal and the wake-up weight coefficient, the control module is specifically configured to: extract dynamic wake-up feature information based on the voice wake-up signal; determine the wake-up score of the home appliance based on the dynamic wake-up feature information, the static wake-up feature information, and the wake-up weight coefficient; and determine the wake-up state of the home appliance based on the static wake-up feature information and the wake-up score.

[0015] The above technical solution has the following advantages or beneficial effects: by influencing the attribute information of voice signal propagation, it avoids the problem of home appliances being mistakenly woken up or unable to be woken up due to changes in attribute information affecting the propagation effect of voice wake-up signal, so that the woken-up home appliances are more in line with the user's actual wishes and improve the user experience.

[0016] In some embodiments, the dynamic wake-up feature information includes one or more of the noise intensity, signal strength, signal arrival time, and direct mixing ratio of the voice wake-up signal, and the static wake-up feature information includes one or more of the wake-up sensitivity and wake-up priority level of the home appliance.

[0017] The above technical solution has the following advantages or beneficial effects: by determining the wake-up status of home appliances through static wake-up feature information and wake-up score, the accuracy of the wake-up status judgment of home appliances can be improved, and the problem of the currently wake-up device being falsely woken up can be avoided.

[0018] In some embodiments, the wake-up weight coefficient includes a static wake-up weight coefficient and a dynamic wake-up weight coefficient. For determining the wake-up score of the home appliance based on the dynamic wake-up feature information, the static wake-up feature information, and the wake-up weight coefficient, the control module is specifically configured to: determine a first product value of the static wake-up feature information and the static wake-up weight coefficient; determine a second product value of the dynamic wake-up feature information and the dynamic wake-up weight coefficient; and determine the wake-up score based on the first product value and the second product value.

[0019] The above technical solution has the following advantages or beneficial effects: the wake-up score of home appliances is dynamically determined based on dynamic wake-up feature information, static wake-up feature information and wake-up weight coefficient, which can more flexibly and accurately identify the unique home appliance to be woken up from multiple home appliances.

[0020] In some embodiments, the communication module is further configured to receive reference wake-up information from other home appliances. Specifically, for determining the wake-up state of the home appliance based on the static wake-up feature information and the wake-up score, the control module is configured to: acquire the reference wake-up information from the other home appliances; and determine the wake-up state of the home appliance based on the static wake-up feature information, the wake-up score, and the reference wake-up information.

[0021] The above technical solution has the following advantages or beneficial effects: by determining the wake-up status of home appliances through static wake-up feature information, wake-up score and reference wake-up information, the accuracy of the current wake-up status judgment of home appliances can be further improved, and the problem of the current wake-up device being falsely woken up can be avoided.

[0022] In some embodiments, the static wake-up feature information includes the wake-up priority level of the home appliance, and the reference wake-up information includes the reference wake-up score and reference wake-up priority level of other home appliances. The wake-up state of the home appliance is determined based on the static wake-up feature information, the wake-up score, and the reference wake-up information. The control module is specifically configured to: if the wake-up score is determined to be lower than the reference wake-up score, then the wake-up state of the home appliance is determined to be not woken up; if the wake-up score is determined to be higher than the reference wake-up score and the wake-up priority level of the home appliance is higher than the reference wake-up priority level, then the wake-up state of the home appliance is determined to be woken up.

[0023] The above technical solution has the following advantages or beneficial effects: When determining whether the wake-up device is the user's wake-up target, this application can effectively improve the accuracy of the wake-up status judgment of the wake-up device by using the reference wake-up information of other home appliances, thereby avoiding the wake-up device being woken up by mistake and improving the user experience.

[0024] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a structural block diagram of a household appliance according to an embodiment of the present invention; Figure 2 This is a flowchart of the control module control process according to an embodiment of the present invention; Figure 3 This is a flowchart of the control module control process according to another embodiment of the present invention; Figure 4 This is a structural block diagram of a household appliance according to another embodiment of the present invention; Figure 5 This is a flowchart of the control module control process according to another embodiment of the present invention; Figure 6 This is a schematic diagram of the principle of voice wake-up of a home appliance according to an embodiment of the present invention; Figure 7 This is a flowchart of the control module control process according to another embodiment of the present invention; Figure 8 This is a flowchart of the control module control process according to another embodiment of the present invention; Figure 9 This is a flowchart of the control module control process according to another embodiment of the present invention; Figure 10 This is a flowchart of the control module control process according to another embodiment of the present invention.

[0026] Figure label: 10 home appliances; 20 user terminal devices; Voice module 1; Acquisition module 2; Control module 3; Memory module 4; Communication module 7; Intelligent decision-making module 31. Detailed Implementation

[0027] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.

[0028] With the rapid development and advancement of science and technology, especially the Internet of Things (IoT), and people's pursuit of comfort and demand for smart living, home appliances have gradually become the new favorites among electronic products. More and more home appliances are entering households and playing an irreplaceable role. In particular, voice control of home appliances makes the home environment more intelligent, but it also brings problems: when users control home appliances by voice, they usually need to wake the appliance with a wake-up word. Appliances using the same wake-up word in the same household may be mistakenly woken up, but users only want to wake up the closest appliance. This problem not only affects the user experience, but also increases the power consumption of multiple appliances being woken up simultaneously, leading to resource waste and interference.

[0029] To address the aforementioned issues, the first aspect of this invention proposes a home appliance that uses a dynamically updated wake-up weight coefficient to determine when to wake up the appliance. This avoids the problem of the appliance being mistakenly woken up or unable to be woken up due to changes in attribute information affecting the propagation effect of the voice wake-up signal. As a result, the woken-up appliance is more in line with the user's actual wishes, thus improving the user experience.

[0030] The following is for reference. Figure 1 Home appliances that describe embodiments of the present invention, such as Figure 1 As shown, the home appliance 10 includes a voice module 1, an acquisition module 2, and a control module 3.

[0031] The system comprises: a voice module 1 for receiving voice wake-up signals provided by the user; an acquisition module 2 for acquiring attribute information that affects the transmission of the voice wake-up signal to the home appliance; and a control module 3 connected to the voice module 1 and the acquisition module 2. Home appliances refer to devices with voice wake-up functionality, such as stereos, televisions, and air conditioners—small appliances with voice wake-up capabilities.

[0032] Based on the architecture of the aforementioned home appliances, and referring to Figure 2 As shown, the control module is specifically configured to execute steps S1-S2, and the specific steps are as follows.

[0033] Step S1: Determine the wake-up weight coefficient based on the attribute information.

[0034] Among these, attribute information refers to information that influences the propagation of the voice wake-up signal to home appliances. The wake-up weight coefficient can be understood as a quantification of the degree of influence of attribute information on the voice wake-up signal. Since attribute information is a variable value, the wake-up weight coefficient determined based on the attribute information is also a variable value.

[0035] Specifically, the transmission effect of the voice wake-up signal deteriorates due to factors such as changes in the distance between the speaker and the appliance, the user's location, environmental noise, and the physical state of the air medium, during its propagation to the appliance. Furthermore, these factors are not static. Therefore, this application dynamically determines the wake-up weight coefficient based on these attributes. For example, the wake-up weight coefficient corresponding to a given attribute can be determined through a mapping relationship between different attributes and the corresponding coefficient. Thus, compared to the fixed wake-up weight coefficients set empirically in existing technologies, this application dynamically updates the wake-up weight coefficient based on changes in the attribute information.

[0036] Step S2: Determine the wake-up status of the home appliance based on the voice wake-up signal and the wake-up weight coefficient.

[0037] Voice wake-up signals can be understood as specific keywords or sound signals contained in the user's voice audio that trigger home appliances to enter working mode from sleep mode. In the wake-up state, the voice module inside the home appliance is activated and prepares to receive the voice wake-up signal, but the appliance itself is not running. For example, if the appliance is an air conditioner, the motor will not operate. False wake-ups are possible in the wake-up state. This means that due to various factors such as the distance between the user and the device, pronunciation standards, and noise, the voice wake-up process may result in the home appliance being woken up even when there is no audio stream or the audio stream does not contain the characteristics or events required for wake-up. In other words, the user only wants to wake up the nearest home appliance, but multiple devices in the same household using the same wake-up word may respond and play the same message. In this case, multiple appliances besides the nearest one are falsely woken up, affecting the user experience.

[0038] Specifically, since attribute information affects the propagation effect of voice wake-up signals—for example, attribute information affects the acoustic characteristics of voice wake-up signals such as timbre, pitch, and intensity—this application considers that changes in the attribute information of home appliances can affect the propagation of voice wake-up signals, potentially lowering the matching degree of the voice wake-up signal as the user's expected wake-up signal. This could lead to misjudgments when determining the wake-up status of home appliances using voice wake-up signals. Therefore, the wake-up weight coefficient is dynamically updated based on changes in the attribute information of home appliances, so that the wake-up weight coefficient can reflect the degree of influence of changed attribute information on the voice wake-up signal. This optimizes the matching degree of the voice wake-up signal as the user's expected wake-up signal by optimizing the wake-up weight coefficient, thereby eliminating or weakening the influence of attribute information on the propagation of the voice wake-up signal. Finally, the wake-up status of the home appliance is determined as awake or not awake based on the optimized matching degree. Therefore, compared to the existing technology that uses fixed weight coefficients to determine the wake-up state of home appliances, this application uses dynamically changing wake-up weight coefficients to determine the wake-up state of home appliances. The wake-up weight coefficients are dynamically updated as the attribute information of home appliances changes, allowing the dynamically changing wake-up weight system to adapt to the recognition of wake-up devices under different attribute information. Furthermore, by optimizing the matching degree of the voice wake-up signal to the user's expected wake-up device signal through dynamically updated wake-up weight coefficients, the influence of attribute information on the propagation of the voice wake-up signal is eliminated, which helps to improve the accuracy of voice wake-up. This avoids the problem of home appliances being mistakenly woken up or unable to be woken up due to the influence of attribute information changes on the propagation effect of the voice wake-up signal. As a result, the woken-up home appliances are more in line with the user's actual wishes, improving the user experience. Moreover, there is no need to increase the cost of home appliances or manually intervene in the selection of wake-up devices, reducing the power consumption of all home appliances and avoiding resource waste and interference.

[0039] According to embodiments of the present invention, the wake-up weight coefficient of the home appliance is dynamically updated by changes in attribute information. The dynamically updated wake-up weight coefficient optimizes the matching degree of the voice wake-up signal with the user's expected wake-up signal, thereby eliminating the influence of attribute information on the propagation of the voice wake-up signal. Therefore, compared with the prior art that uses a single weight to determine the wake-up state of the wake-up device, the present application uses a dynamically changing wake-up weight coefficient to determine the wake-up state of the home appliance. This allows the dynamically changing wake-up weight system to adapt to the recognition of the wake-up device under different attribute information. Furthermore, the wake-up weight coefficient is dynamically updated as the attribute information of the home appliance changes, thereby avoiding the problem of the home appliance being mistakenly woken up or unable to be woken up due to the influence of attribute information changes on the propagation effect of the voice wake-up signal. This makes the woken-up home appliance more in line with the user's actual wishes and improves the user experience.

[0040] In some embodiments, the wake-up weight coefficient includes a wake-up static weight coefficient and a wake-up dynamic weight coefficient. For determining the wake-up static weight coefficient based on attribute information, the control module is specifically configured to: determine the wake-up static weight coefficient and / or the wake-up dynamic weight coefficient based on attribute information.

[0041] Among them, the static wake-up weighting coefficient and the dynamic wake-up weighting coefficient take into account the degree of influence of different attribute information on the voice wake-up signal. That is, the static wake-up weighting coefficient and the dynamic wake-up weighting coefficient quantify the degree of influence of different types of attribute information on the voice wake-up signal.

[0042] Specifically, the transmission effect of the voice wake-up signal deteriorates due to various factors, including changes in the distance between the speaker and the appliance, the user's location, environmental noise, and the physical state of the air medium, during its propagation to the home appliance. Therefore, this application dynamically determines the static and dynamic wake-up weight coefficients corresponding to different types of attribute information based on these coefficients. Thus, compared to the fixed wake-up weight coefficients set empirically in existing technologies, this application dynamically updates the static and dynamic wake-up weight coefficients based on changes in attribute information. Furthermore, it updates the static and dynamic wake-up weight coefficients for different types of attribute information based on their respective degrees of influence on the voice wake-up signal, ensuring that these coefficients accurately reflect the impact of different types of attribute information on the propagation of the voice wake-up signal.

[0043] Based on this, the static and dynamic wake-up weight coefficients corresponding to different types of attribute information are dynamically updated according to changes in the attribute information of home appliances. This ensures that the static and dynamic wake-up weight coefficients reflect the degree of influence of different types of attribute information changes on the voice wake-up signal. By optimizing the matching degree of the voice wake-up signal with the user's expected wake-up signal, the influence of attribute information on the propagation of the voice wake-up signal is eliminated. The optimized matching degree then determines whether the home appliance is awake or not, which helps improve the accuracy of voice wake-up and avoids the problem of home appliances being mistakenly woken up or unable to be woken up due to the impact of different types of attribute information changes on the propagation effect of the voice wake-up signal. This ensures that the woken-up home appliances better match the user's actual wishes and improves the user experience.

[0044] In this embodiment, the initial value of the wake-up static weight coefficient can be set according to actual needs, and can be adjusted later according to attribute information.

[0045] In some embodiments, the attribute information includes environmental state information of the home appliance, performance information of the home appliance, and user behavior information. A wake-up static weight coefficient is determined based on the attribute information. Specifically, the control module is configured to determine the wake-up static weight coefficient based on one or more of the environmental state information, performance information, and user behavior information. For example, the static weight coefficient may be determined based on environmental state information and performance information; or, based on environmental state information and user behavior information; or, based on performance information and user behavior information; or, based on environmental state information, performance information, and user behavior information.

[0046] The environmental status information of home appliances can be understood as environmental information related to the location of the home appliance that affects the propagation of the voice wake-up signal. Environmental status information can include environmental parameters and device parameters. Environmental parameters include the user's geographical location, season, weather, air quality, indoor temperature, and humidity. Device parameters can include the type and model of the home appliance, its installation location, and its current status. The installation location can be the living room, kitchen, or bedroom, and the current status can be standby or not powered on, without restrictions. It should be noted that a home appliance cannot be a wake-up device when it is not powered on; therefore, the weighting coefficient is adjusted accordingly. The performance information of the home appliance can be understood as performance information related to the home appliance's ability to receive voice wake-up signals. User behavior information can be the user's operation information on the home appliance and other home appliances after the home appliance is woken up as the user's wake-up target device. User behavior information can include manually switching device information and voice correction commands.

[0047] Specifically, due to differences in the performance information of different home appliances, their reception speed for voice wake-up signals varies. Furthermore, differences in the environmental conditions surrounding the appliances affect the propagation speed of the voice wake-up signal through the air. Additionally, changes in user behavior information for appliances using the same wake-up word result in different voice wake-up signals; the woken appliance may not meet the user's current needs, prompting the user to manually or verbally switch to another appliance with the same wake-up word. Therefore, a static wake-up weight is determined based on one or more of the environmental conditions, performance information, and user behavior information. The wake-up static weight coefficient is a dynamically updated coefficient that combines one or more of the following: environmental status information, performance information, and user behavior information. This allows the wake-up static weight coefficient to reflect information affecting the propagation of the voice wake-up signal and changes in the user. Therefore, when determining the wake-up state of home appliances through the wake-up static weight coefficient and the voice wake-up signal, it can adapt to the usage needs of different users for home appliances with the same wake word. It can also avoid the problem of home appliances being mistakenly woken up due to changes in performance information and environmental status information. Moreover, the wake-up static weight coefficient is optimized based on user behavior information, thereby avoiding the problem of rigid wake-up of home appliances. This makes the woken-up home appliances more in line with the user's actual wishes and improves the user experience.

[0048] In some embodiments, the attribute information includes environmental state information of the home appliance, performance information of the home appliance, and user behavior information. A wake-up dynamic weighting coefficient is determined based on the attribute information. Specifically, the control module is configured to determine the dynamic weighting coefficient based on one or more of the environmental state information, performance information, and user behavior information. For example, the dynamic weighting coefficient may be determined based on both environmental state information and performance information; or, it may be determined based on both environmental state information and user behavior information; or, it may be determined based on both performance information and user behavior information; or, it may be determined based on a combination of environmental state information, performance information, and user behavior information.

[0049] Specifically, due to differences in the performance information of different home appliances, their reception speed for voice wake-up signals varies. Furthermore, differences in the environmental conditions surrounding the appliances affect the propagation speed of the voice wake-up signal through the air. Additionally, changes in user behavior (i.e., different users using the same wake-up word) result in different voice wake-up signals, and the woken-up appliance may not meet the user's current needs. In such cases, the user may manually or verbally switch to another appliance with the same wake-up word. Therefore, a dynamic wake-up weighting system is determined based on one or more of the environmental condition information, performance information, and user behavior information. The wake-up dynamic weight coefficient is dynamically updated by combining one or more of the following: environmental status information, performance information, and user behavior information. This ensures that the wake-up dynamic weight coefficient takes into account factors affecting the propagation of the voice wake-up signal and changes in user behavior. Therefore, when determining the wake-up status of home appliances through the wake-up dynamic weight coefficient and the voice wake-up signal, it can adapt to the usage needs of different users for home appliances with the same wake word. It can also avoid the problem of home appliances being mistakenly woken up due to changes in performance information and environmental status information. Moreover, the wake-up dynamic weight coefficient is optimized based on user behavior information, thereby avoiding the problem of rigid wake-up of home appliances. This makes the woken-up home appliances more in line with the user's actual wishes and improves the user experience.

[0050] In addition, it should be noted that different home appliances have different performance information, and therefore different wake-up weight coefficients.

[0051] In this embodiment, the wake-up static weight coefficient corresponding to different attribute information can be called through a mapping table between environmental state information, performance information, and user behavior information and wake-up static weight coefficient; and the wake-up dynamic weight coefficient corresponding to different attribute information can be called through a mapping table between environmental state information, performance information, and user behavior information and wake-up dynamic weight coefficient.

[0052] In some embodiments, the wake-up weight coefficient is determined based on attribute information, with reference to... Figure 3 As shown, the control module is specifically configured to execute steps S3-S4, and the specific steps are as follows.

[0053] Step S3: Obtain an adaptive model based on training multiple attribute information.

[0054] Step S4: Determine the wake-up weight coefficients based on the attribute information using an adaptive model.

[0055] The adaptive model dynamically adjusts the output wake-up weight coefficients based on different attribute information. The adaptive model can be a Transformer model or an optimization learning adaptive model. It is a lightweight, large-scale model. The fine-tuning weights and thresholds of the adaptive model can be updated through an online learning algorithm, such as the FTRL-Proximal algorithm. Specifically, the online learning algorithm outputs the optimal values ​​for both the dynamic and static wake-up weight coefficients for each home appliance.

[0056] Specifically, to improve the accuracy of the wake-up weight coefficients output by the adaptive model, the adaptive model needs to continuously learn changes in environmental state information, performance information, and user behavior information. It is trained using multiple attribute information to update the weights and thresholds of the adaptive model. The attribute information determines the wake-up weight coefficients through the adaptive model; that is, the attribute information is input into the adaptive model to output different wake-up weight coefficients. In other words, feature extraction is performed on the attribute information, and then the extracted information is used to train the adaptive model, updating the fine-tuned weights and thresholds. This allows the adaptive model to dynamically output wake-up weight coefficients based on real-time changes in attribute information. The wake-up weight coefficients include static wake-up weight coefficients and / or dynamic wake-up weight coefficients. Therefore, the adaptive model in this application can continuously learn real-time changing attribute information, adaptively upgrade and optimize the wake-up weight coefficients, and automate the entire process of perceiving and optimizing the wake-up weight coefficients. This greatly improves the accuracy of the wake-up weight coefficients output by the adaptive model, increases the efficiency of iterative optimization of the wake-up weight coefficients, and enables continuous optimization of wake-up device recognition through the adaptive model. It can more flexibly and accurately identify the unique home appliance to be woken up from multiple home appliances, achieving a more personalized wake-up strategy that better suits the user's actual needs, and is more flexible and accurate.

[0057] In this embodiment, some or all of the stored user behavior information is periodically input into the adaptive model at preset intervals, so that the adaptive model can fine-tune the weights and thresholds in a timely manner based on the user behavior information. The preset intervals can be set according to actual needs.

[0058] In this embodiment, after the bedroom air conditioner is woken up, the user manually turns off the bedroom air conditioner A and turns on the living room air conditioner B, or the user turns off the bedroom air conditioner A and turns on the living room air conditioner through a voice correction command. This user behavior information is then used as a negative sample input to the adaptive model.

[0059] In some embodiments, such as Figure 4 As shown, the home appliance 10 also includes a communication module 7.

[0060] The communication module 7 is connected to the control module 3 and the user terminal device 20. The communication module 7 is used to obtain static wake-up characteristic information about home appliances provided by the user terminal device. The communication module can be a Bluetooth module.

[0061] For determining the wake-up status of home appliances based on voice wake-up signals and wake-up weighting coefficients, refer to... Figure 5 As shown, the control module is specifically configured to execute steps S5-S7, and the specific steps are as follows.

[0062] Step S5: Extract dynamic wake-up feature information based on the voice wake-up signal.

[0063] Dynamic wake-up feature information can be understood as feature information extracted from the voice wake-up signal that reflects the user's wake-up intention. Dynamic wake-up feature information is used to determine whether a home appliance is a wake-up device.

[0064] Specifically, after receiving a voice wake-up signal, the voice module of the home appliance actively extracts the dynamic wake-up feature information from the voice wake-up signal.

[0065] Step S6: Determine the wake-up score of the home appliance based on the dynamic wake-up feature information, the static wake-up feature information, and the wake-up weight coefficient.

[0066] Static wake-up feature information can be understood as information pre-set by the user that reflects the user's wake-up intention. The wake-up score is used to assess the likelihood that home appliances are likely to be woken up by the user's intention; the wake-up score is a quantitative indicator.

[0067] Specifically, the wake-up score of a home appliance is determined based on dynamic wake-up feature information, static wake-up feature information, and a wake-up weight coefficient. In other words, the likelihood of the home appliance being woken up by the user's intention is determined using dynamic and static wake-up feature information. The wake-up weight coefficient optimizes the attribute information's representation of the likelihood of the home appliance being woken up by the user's intention, thus eliminating the influence of attribute information on the wake-up score. Therefore, this application avoids the problem of home appliances being falsely woken up or unable to be woken up due to changes in attribute information affecting the propagation of the voice wake-up signal, thereby ensuring that the woken-up home appliance better matches the user's actual intention and improving the user experience.

[0068] Step S7: Determine the wake-up status of the home appliance based on the static wake-up feature information and the wake-up score.

[0069] Specifically, the wake-up score determines the likelihood that a home appliance is the device the user intends to wake up. Combined with the wake-up priority and wake-up sensitivity of the home appliance, it is further determined whether the home appliance is the user's desired wake-up target. Thus, by determining the wake-up status of home appliances through static wake-up feature information and wake-up score, the accuracy of the wake-up status judgment of home appliances can be improved, and the problem of the currently wake-up device being falsely woken up can be avoided.

[0070] In an embodiment, such as Figure 6 As shown, the device parameters, environmental parameters, and user behavior information stored in the memory module 4 are input into the adaptive model. The adaptive model dynamically updates the wake-up weight coefficient. The intelligent decision module 31 of the control module 3 receives the dynamically updated wake-up weight coefficient, static wake-up feature information, and dynamic wake-up feature information to determine the wake-up status of the home appliance based on the wake-up weight coefficient, static wake-up feature information, and dynamic wake-up feature information.

[0071] The memory module stores long-term or short-term user behavior feedback information regarding wake-up results. This feedback reflects user habits and personal preferences. The memory module can also periodically clean up older data stored for longer periods by setting a preset time, while simultaneously storing new data. This optimizes the wake-up weight coefficients output by the adaptive model, enabling dynamic updates to the wake-up weight coefficients. The preset time can be set according to actual needs.

[0072] In some embodiments, dynamic wake-up feature information includes one or more of the following: noise intensity, signal strength, signal arrival time, and direct current mixing ratio of the voice wake-up signal; or, the dynamic wake-up feature information includes the noise intensity and signal strength of the voice wake-up signal; or, the dynamic wake-up feature information includes the noise intensity, signal strength, signal arrival time, and direct current mixing ratio of the voice wake-up signal. Static wake-up feature information includes one or more of the following: wake-up sensitivity and wake-up priority level of the home appliance. For example, static wake-up feature information may include wake-up sensitivity and wake-up priority level.

[0073] The voice wake-up signal includes not only the original voice containing specific keywords or sounds emitted by the user, but also noise. The noise intensity of the voice wake-up signal refers to the energy level of the noise in the audio. The signal strength of the voice wake-up signal refers to the energy level of the sound. The direct-to-indirect mixing ratio is the proportion of the original sound in the entire voice wake-up signal. The signal arrival time can be understood as the time it takes for the home appliance to receive the voice wake-up signal provided by the user. Generally, the closer the home appliance is to the user, the shorter the signal arrival time. The signal arrival time can determine the actual distance between the home appliance and the user, thus enabling more accurate voice-based wake-up based on proximity when determining the wake-up device, improving the accuracy of voice wake-up for home appliances.

[0074] Among these, wake-up sensitivity refers to the degree to which a home appliance responds to a voice wake-up signal. Wake-up priority refers to the order in which a particular home appliance responds to a voice wake-up signal among all home appliances with the same wake-up word. Users can choose and set wake-up sensitivity and wake-up priority based on their preferences and usage habits. For example, users in the same household can choose and set the wake-up sensitivity and wake-up priority of their home appliances through an app on their user terminal device.

[0075] For example, the wake-up sensitivity of home appliances can be set to five levels: lowest, low, medium, high, and highest. If the user has not yet set the wake-up sensitivity of the home appliances, the default wake-up sensitivity is medium. When the user sets the wake-up sensitivity of the home appliances to the highest level, it means that the home appliances are most easily woken up. Regarding wake-up priority, if home appliances A, B, and a refrigerator have the same wake-up word, and the wake-up priority of air conditioner A > the wake-up priority of air conditioner B > the wake-up priority of the refrigerator, then air conditioner A will be woken up first. Furthermore, if the user has not set the wake-up priority level for home appliances with the same wake-up word, the default device sensitivity of all home appliances is medium wake-up priority.

[0076] In this embodiment, after a user sets the wake-up sensitivity and wake-up priority level of a certain home appliance through a user terminal device, the settings are sent to the home appliance via the communication module of the user terminal device. Upon receiving the wake-up sensitivity and wake-up priority level provided by the user, the home appliance device can also proactively request the user terminal device to obtain the wake-up sensitivity and wake-up priority level. When the communication module or voice module receives the new wake-up sensitivity and wake-up priority level settings, it needs to apply and memorize the wake-up sensitivity and wake-up priority level. After the wake-up sensitivity and wake-up priority level settings are successfully set, the results are synchronized to the user terminal device and displayed to the user.

[0077] In this embodiment, static wake-up feature information can also be automatically generated based on user usage through a generative large model. Different home appliances possess different static wake-up feature information depending on user usage. Static wake-up feature information includes, but is not limited to, wake-up sensitivity and wake-up priority level, as well as the mode the home appliance is in, which can be a nighttime focus mode. Dynamic wake-up feature information may also include user behavior information. For example, when a home appliance receives a voice wake-up signal, it can simultaneously determine the distance between the device and the user by measuring the signal strength received from the user's mobile phone / watch. Generally speaking, the higher the signal strength, the closer the device is to the user.

[0078] In some embodiments, the wake-up weighting coefficient includes a static wake-up weighting coefficient and a dynamic wake-up weighting coefficient. For determining the wake-up score of a home appliance based on dynamic wake-up feature information, static wake-up feature information, and the wake-up weighting coefficient, refer to... Figure 7 As shown, the control module is specifically configured to execute steps S8-S10, and the specific steps are as follows.

[0079] Step S8: Determine the first product value of the static wake-up feature information and the wake-up static weight coefficient.

[0080] Specifically, taking the static wake-up feature information, including the wake-up sensitivity and wake-up priority level of home appliances, as an example, the wake-up sensitivity and wake-up priority level are multiplied by the wake-up static weight coefficient and then added together to obtain the first product value. Alternatively, corresponding values ​​can be set for different wake-up sensitivities and different wake-up priority levels, with higher values ​​corresponding to higher wake-up sensitivity and wake-up priority levels. When the wake-up sensitivity and wake-up priority level are involved in the calculation of the first product value, the corresponding values ​​of the wake-up sensitivity and wake-up priority level can be multiplied together to obtain the first product value.

[0081] Step S9: Determine the second product value of the dynamic wake-up feature information and the wake-up dynamic weight coefficient.

[0082] The second product value refers to the dynamic wake-up score derived from dynamic wake-up feature information and wake-up dynamic weight coefficient.

[0083] Specifically, taking the dynamic wake-up feature information, including the noise intensity, signal strength, signal arrival time, and direct mixing ratio of the voice wake-up signal, as an example, the noise intensity, signal strength, signal arrival time, and direct mixing ratio are multiplied by the wake-up dynamic weighting coefficient and then added together to obtain the second product value.

[0084] Step S10: Determine the wake-up score based on the first product value and the second product value.

[0085] Specifically, the wake-up score is obtained by adding the first product value and the second product value. Thus, the wake-up score is calculated using static wake-up feature information, static wake-up weight coefficient, dynamic wake-up feature information, and dynamic wake-up weight coefficient. This allows the static wake-up weight coefficient to optimize the static wake-up feature information, eliminating the influence of attribute information of the corresponding type on the propagation of the voice wake-up signal. Simultaneously, the dynamic wake-up weight coefficient can optimize the dynamic wake-up feature information, eliminating the influence of attribute information of the corresponding type on the propagation of the voice wake-up signal. Therefore, when determining the wake-up score by optimizing the static and dynamic wake-up feature information, the wake-up score can eliminate the adverse effects of attribute information on the propagation of the voice wake-up signal, enabling a more flexible and accurate identification of the unique appliance to be woken up from multiple appliances.

[0086] In some embodiments, the communication module is further configured to receive reference wake-up information from other home appliances, for determining the wake-up state of home appliances based on static wake-up characteristic information and wake-up score, the reference... Figure 8 As shown, the control module is specifically configured to execute steps S11-S12, and the specific steps are as follows.

[0087] Step S11: Obtain reference wake-up information from other home appliances.

[0088] The reference wake-up information for other home appliances can be understood as information used to determine whether other home appliances are the devices woken up by the user. This reference wake-up information is used to determine the current wake-up status of the home appliance. There can be multiple other home appliances. The wake-up word for each other home appliance is the same as that for the current home appliance.

[0089] Specifically, currently, home appliances send reference wake-up information via a Wi-Fi (Wi-Fi network) local area network. This means that multiple smart home appliances within the same Wi-Fi network compare reference wake-up information to determine which appliance should be woken up locally. However, this method requires all appliances to be connected to the same Wi-Fi network. This suffers from uneven network coverage and signal interference, making it difficult to quickly identify the unique appliance to be woken up from among multiple devices. Other appliances can send their own reference wake-up information to the first appliance via their communication modules. The first appliance then receives this information, avoiding uneven network coverage and signal interference, and can quickly identify the unique appliance to be woken up from among multiple devices, making the wake-up process more flexible and accurate.

[0090] Step S12: Determine the wake-up status of the home appliance based on the static wake-up feature information, wake-up score, and reference wake-up information.

[0091] Specifically, since the wake-up words of other home appliances are the same as those of the current home appliance, in order to improve the accuracy of the wake-up status judgment of the current home appliance and avoid the problem of the current wake-up device being falsely woken up, this application determines the probability that the current home appliance is the user's desired wake-up target by using the wake-up score of the current home appliance and the reference wake-up information of other home appliances. Furthermore, it further determines whether the current device is the user's desired wake-up target by combining the wake-up priority level and wake-up sensitivity of the current home appliance. Thus, by determining the wake-up status of the home appliance through static wake-up feature information, wake-up score, and reference wake-up information, the accuracy of the wake-up status judgment of the current home appliance can be further improved, and the problem of the current wake-up device being falsely woken up can be avoided.

[0092] In this embodiment, when a home appliance receives a voice wake-up signal, it can also determine the distance between the home appliance and the user by measuring the signal strength received from the user's mobile phone and / or watch. Generally, the higher the signal strength received from the user's mobile phone and / or watch, the closer the home appliance is to the user.

[0093] In some embodiments, static wake-up feature information includes the wake-up priority level of the home appliance, and reference wake-up information includes the reference wake-up score and reference wake-up priority level of other home appliances. The wake-up state of the home appliance is determined based on the static wake-up feature information, the wake-up score, and the reference wake-up information. The control module is specifically configured to perform the following steps.

[0094] Step S22: If the wake-up score is determined to be lower than the reference wake-up score, then the wake-up state of the home appliance is determined to be not woken up.

[0095] Step S23: If it is determined that the wake-up score is higher than the reference wake-up score and the wake-up priority level of the home appliance is higher than the reference wake-up priority level, then the wake-up state of the home appliance is determined to be "wake-up".

[0096] Among them, the reference wake-up score and reference wake-up priority level are the wake-up scores and wake-up priority levels of other home appliances.

[0097] Specifically, if the wake-up score of the current home appliance is determined to be lower than the reference wake-up score of other home appliances, it indicates a high probability that the current home appliance is the user's intended wake-up target, and the current home appliance's wake-up state is determined to be "not woken up." However, since the wake-up score being higher than the reference wake-up score includes situations where the home appliance's wake-up score is equal to the wake-up score of other home appliances, if the home appliance's wake-up priority is determined to be higher than the reference wake-up priority, it indicates that the current home appliance is the user's desired wake-up target, and the current home appliance's wake-up state is determined to be "wake-up." In other words, if multiple home appliances have the same and highest wake-up score, further filtering is performed based on the reference wake-up information, and the device with the higher user preference level will be selected as the responding device. Therefore, this application, by utilizing the reference wake-up information of other home appliances when determining whether a wake-up device is the user's wake-up target, can effectively improve the accuracy of the wake-up status judgment of the wake-up device, thereby avoiding false wake-ups and improving the user experience.

[0098] In addition, if the wake-up score is greater than the reference wake-up score, the current wake-up state of the home appliance is determined to be "wake-up".

[0099] The following is for reference. Figure 10 The control process of the control module in an embodiment of the present invention is illustrated by example, and the specific content is as follows.

[0100] Step S13: Obtain static wake-up feature information provided by the user.

[0101] Step S14: Obtain dynamic wake-up feature information from the voice wake-up signal.

[0102] Step S15: Calculate the wake-up score of the home appliance based on the static wake-up feature information, dynamic wake-up feature information, and wake-up weight coefficient.

[0103] Step S16: The home appliance obtains the reference wake-up score of other home appliances through the Bluetooth network.

[0104] Step S17: Determine whether the wake-up score is higher than the reference wake-up score. If yes, proceed to step S18; otherwise, proceed to step S19.

[0105] Step S18: Determine whether the number of home appliances with the highest wake-up score is greater than 1. If yes, proceed to step S20; otherwise, proceed to step S21.

[0106] Step S19: Determine that the wake-up state of the home appliance is not woken up, and control the home appliance to standby mode.

[0107] Step S20: Determine whether the wake-up priority level of the home appliance is higher than the reference wake-up priority level of other home appliances. If yes, proceed to step S21; otherwise, proceed to step S19.

[0108] Step S21: Determine that the wake-up state of the home appliance is "wake-up" and control the home appliance to respond.

[0109] In this embodiment, if there is only one home appliance in the same home scenario or only one home appliance's wake-up score is obtained, then the reference wake-up scores of other home appliances are considered to be zero. At this time, if the wake-up score of the current home appliance is higher than the reference wake-up scores of other home appliances, then the wake-up state of the home appliance is confirmed to be awake.

[0110] In this embodiment, once the wake-up status of a home appliance is determined, a notification is broadcast to other home appliances via the communication module to ensure that other home appliances receive the wake-up decision result. Simultaneously, the home appliances can broadcast a response to confirm that the user has successfully woken them up, and the home appliances then perform corresponding operations according to the user's commands. Other home appliances then return to standby mode, waiting for the next voice wake-up signal from the user.

[0111] The above technical solution has the following advantages or beneficial effects: by comparing the wake-up score of the current home appliance with the reference wake-up score of other home appliances, and by comparing the wake-up priority level of the current home appliance with the reference wake-up priority level of other home appliances to determine the wake-up status of the home appliance, the accuracy of the wake-up result can be improved.

[0112] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0113] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. An electric home appliance characterized by comprising: include: The voice module is used to receive voice wake-up signals provided by the user; The acquisition module is used to acquire attribute information that affects the propagation of the voice wake-up signal to the home appliance; A control module, connected to the voice module and the acquisition module, is specifically configured to: determine a wake-up weight coefficient based on the attribute information, wherein the attribute information includes environmental state information of the home appliance, performance information of the home appliance, and user behavior information; and determine the wake-up state of the home appliance based on the voice wake-up signal and the wake-up weight coefficient. Specifically, the control module is configured to determine the wake-up weight coefficient based on the attribute information as follows: The wake-up weight coefficient is determined by an adaptive model based on the attribute information, and the adaptive model is trained on multiple attribute information. The home appliance also includes a communication module, which is connected to the control module and the user terminal device, and is used to obtain static wake-up feature information about the home appliance provided by the user terminal device. The control module is specifically configured to determine the wake-up state of the home appliance based on the voice wake-up signal and the wake-up weight coefficient as follows: Dynamic wake-up feature information is extracted based on the voice wake-up signal; The wake-up score of the home appliance is determined based on the dynamic wake-up feature information, the static wake-up feature information, and the wake-up weight coefficient; The wake-up status of the home appliance is determined based on the static wake-up feature information and the wake-up score; The dynamic wake-up feature information includes one or more of the following: noise intensity, signal strength, signal arrival time, and direct mixing ratio of the voice wake-up signal. The static wake-up feature information includes one or more of the following: wake-up sensitivity and wake-up priority level of the home appliance.

2. The home appliance according to claim 1, characterized in that, in that, The wake-up weight coefficient includes a static wake-up weight coefficient and a dynamic wake-up weight coefficient. Specifically, for determining the static wake-up weight coefficient based on the attribute information, the control module is configured as follows: The wake-up static weight coefficient and / or the wake-up dynamic weight coefficient are determined based on the attribute information.

3. The home appliance of claim 2, wherein The attribute information includes environmental status information of the home appliance, performance information of the home appliance, and user behavior information. A wake-up static weight coefficient is determined based on the attribute information. The control module is specifically configured as follows: The wake-up static weight coefficient is determined based on one or more of the environmental state information, the performance information, and the user behavior information.

4. The household appliance according to claim 2, characterized in that, The attribute information includes environmental status information of the home appliance, performance information of the home appliance, and user behavior information. The wake-up dynamic weight coefficient is determined based on the attribute information. The control module is specifically configured as follows: The wake-up dynamic weight coefficient is determined based on one or more of the environmental state information, the performance information, and the user behavior information.

5. The household appliance according to claim 1, characterized in that, The wake-up weighting coefficient includes a static wake-up weighting coefficient and a dynamic wake-up weighting coefficient. To determine the wake-up score of the home appliance based on the dynamic wake-up feature information, the static wake-up feature information, and the wake-up weighting coefficient, the control module is specifically configured as follows: Determine the first product value of the static wake-up feature information and the static wake-up weight coefficient; Determine the second product value of the dynamic wake-up feature information and the wake-up dynamic weight coefficient; The wake-up score is determined based on the first product value and the second product value.

6. The household appliance according to claim 1, characterized in that, The communication module is also used to receive reference wake-up information from other home appliances. Specifically, the control module is configured to: determine the wake-up state of the home appliance based on the static wake-up feature information and the wake-up score. Obtain reference wake-up information from the other home appliances; The wake-up status of the home appliance is determined based on the static wake-up feature information, the wake-up score, and the reference wake-up information.

7. The household appliance according to claim 6, characterized in that, The static wake-up feature information includes the wake-up priority level of the home appliance, and the reference wake-up information includes the reference wake-up score and reference wake-up priority level of other home appliances. The wake-up state of the home appliance is determined based on the static wake-up feature information, the wake-up score, and the reference wake-up information. The control module is specifically configured as follows: If the wake-up score is determined to be lower than the reference wake-up score, then the wake-up state of the home appliance is determined to be not woken up; If it is determined that the wake-up score is higher than the reference wake-up score and the wake-up priority level of the home appliance is higher than the reference wake-up priority level, then the wake-up state of the home appliance is determined to be "wake-up".

Images