Control method for smart home system
By updating the voice recognition model in the smart home system and adjusting the recognition module according to the noise type, the inaccurate voice recognition problem of household appliances under noise changes is solved, and the user experience is improved.
Patent Information
- Application Number
- CN202011450932.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-12-09
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-12-09
AI Technical Summary
In smart home systems, it is difficult for household appliances such as washing machines to effectively reduce noise and process voice recognition under changes in noise and environmental noise, resulting in inaccurate recognition and affecting the user experience.
By establishing a voice training mode between home appliances and servers, obtaining and updating the voice recognition model, selectively updating dedicated or general voice recognition modules according to the noise type, ensuring that the model matches the latest noise and improving the noise reduction processing effect.
Effectively avoid the impact of noise changes on voice recognition, improve user experience, and ensure accurate recognition and processing of voice commands.
Smart Images

Figure CN114625014B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of smart home, and specifically provides a control method for a smart home system. Background Art
[0002] In smart home life, taking washing machines as an example, washing machines with voice functions are becoming more and more popular. Users can free their hands through voice operation and achieve contactless control. However, the use environment of the washing machine has a direct impact on the voice recognition effect, and the washing machine itself also has a certain amount of noise. The superposition of these noises will affect the recognition algorithm of the voice module itself, resulting in inaccurate voice recognition, and ultimately unable to obtain the recognition results that users want, resulting in user complaints.
[0003] In order to solve the above problems, in the prior art, the noise of the washing machine itself and the common environmental noise of the washing machine are recorded in advance, and noise reduction processing is performed through algorithm training, which can alleviate the impact of some noise on speech recognition to a certain extent. However, the noise of the washing machine itself and the common environmental noise of the washing machine will change. When any of the noises changes, it will still lead to the inability to effectively reduce the noise of the voice information and accurately recognize the user's voice information, thereby failing to fundamentally solve the impact of noise on speech recognition.
[0004] Therefore, the art needs a new control method for a smart home system to solve the above problems. Summary of the invention
[0005] In order to solve the above-mentioned problems in the prior art, that is, to solve the problem that household appliances in the existing smart home system cannot effectively solve the influence of noise on voice recognition, the present invention provides a control method for a smart home system, the smart home system includes a household appliance and a server, the household appliance can communicate with the server, the household appliance has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on input first audio information; the control method comprises the following steps: when the household appliance is in the voice training mode, the household appliance obtains second audio information; the household appliance sends the second audio information to the server; the server updates the first voice recognition model according to the second audio information.
[0006] In the preferred technical solution of the above-mentioned control method, the first speech recognition model includes a first dedicated speech recognition module and a first general speech recognition module; the step of "the server updates the first speech recognition model according to the second audio information" specifically includes: obtaining the noise type of the second audio information; according to the noise type and the second audio information, the server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module.
[0007] In the preferred technical solution of the above control method, the noise includes a first type of noise and a second type of noise; the step of "according to the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module" specifically includes: if the noise type is the first type of noise, the server obtains the MAC address of the household appliance that sends the second audio information, and the server updates the first dedicated voice recognition module corresponding to the MAC address according to the second audio information; or if the noise type is the second type of noise, the server updates the first general voice recognition module according to the second audio information.
[0008] In the preferred technical scheme of the above-mentioned control method, the noise includes a first type of noise and a second type of noise; the step of "according to the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module" also includes: if the noise type includes the first type of noise and the second type of noise, the server obtains the MAC address of the household appliance that sends the second audio information, the server extracts the first noise information belonging to the first type of noise from the second audio information, the server updates the first dedicated voice recognition module corresponding to the MAC address according to the first noise information, and the server extracts the second noise information belonging to the second type of noise from the second audio information, and the server updates the first general voice recognition module according to the second noise information.
[0009] In the preferred technical solution of the above-mentioned control method, the household appliance has a second voice recognition model, the second voice recognition model includes a second dedicated voice recognition module and a second general voice recognition module, wherein the second voice recognition model is used to identify the control instruction information after performing noise reduction processing on the input third audio information; after "according to the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module", the control method also includes: the server sends the updated first dedicated voice recognition module and / or the first general voice recognition module to the household appliance; the household appliance updates the second dedicated voice recognition module and / or the second general voice recognition module according to the updated first dedicated voice recognition module and / or the first general voice recognition module.
[0010] In the preferred technical solution of the above-mentioned control method, the step of "the household appliance sends the second audio information to the server" specifically includes: the household appliance stores the second audio information; the household appliance marks the noise type on the second audio information according to the attributes of the second audio information; the household appliance sends the second audio information marked with the noise type to the server in the form of a digital signal; wherein the attributes include at least one of frequency, amplitude and signal strength.
[0011] In a preferred technical solution of the above control method, the household appliance is provided with an operation key, and the operation key is configured to make the household appliance enter the voice training mode when operated; before "when the household appliance is in the voice training mode, the household appliance obtains the second audio information", the control method also includes: judging whether the operation key is operated; if the operation key is operated, making the household appliance enter the voice training mode; or if the operation key is not operated, making the household appliance not enter the voice training mode.
[0012] In the preferred technical solution of the above control method, before "when the household appliance is in the voice training mode, the household appliance obtains second audio information", the control method also includes: the household appliance obtains fourth audio information of the user; determines whether valid control command information is obtained; if the household appliance does not obtain valid control command information, enables the household appliance to enter the voice training mode; or if the household appliance obtains valid control command information, enables the household appliance not to enter the voice training mode.
[0013] In addition, the present invention also provides a control method for a smart home system, which includes a household appliance and a server, the household appliance can communicate with the server, the household appliance has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on input first audio information; the control method includes the following steps: when the household appliance is in the voice training mode, the household appliance obtains second audio information; the household appliance sends the second audio information to the server, so that the server updates the first voice recognition model according to the second audio information.
[0014] In addition, the present invention also provides a control method for a smart home system, which includes a household appliance and a server, wherein the household appliance can communicate with the server, the household appliance has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on input first audio information; the control method includes the following steps: when the household appliance is in the voice training mode, the server receives second audio information from the household appliance; the server updates the first voice recognition model according to the second audio information.
[0015] It can be understood by those skilled in the art that in the preferred technical solution of the control method of the present invention, the smart home system includes a washing machine and a server, the washing machine can communicate with the server, the washing machine has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify the control instruction information after performing noise reduction processing on the input first audio information. The control method of the present invention includes the following steps: when the washing machine is in the voice training mode, the washing machine obtains the second audio information; the washing machine sends the second audio information to the server; and the server updates the first voice recognition model according to the second audio information.
[0016] Compared with the technical solution in the prior art of pre-recording the noise of the washing machine itself and the common environmental noise of the washing machine, and performing noise reduction processing through algorithm training, the server of the present invention updates the first speech recognition model according to the second audio information, so that the first speech recognition model can be updated in time, so that the updated first speech recognition model is more matched with the latest noise, and can more effectively perform noise reduction processing on the input first audio information, fundamentally avoiding the influence of noise changes on the recognition of the first audio information, and can accurately identify control command information, thereby improving user experience.
[0017] Furthermore, based on the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module, which not only updates the first dedicated voice recognition module dedicated to the washing machine so that the washing machine can effectively reduce the noise unique to the input first audio information, but also updates the first general voice recognition module applicable to all washing machines of the same model as the washing machine, so that all washing machines of the same model as the washing machine can more effectively reduce the noise common to the input first audio information, thereby further improving the user experience.
[0018] Furthermore, the server sends the updated first dedicated voice recognition module and / or the first general voice recognition module to the washing machine, and the washing machine updates the second dedicated voice recognition module and / or the second general voice recognition module according to the updated first dedicated voice recognition module and / or the first general voice recognition module, so that the second voice recognition model on the washing machine can also be updated in time, so that the updated second voice recognition model can also better match the latest noise, and can more effectively perform noise reduction processing on the input third audio information, fundamentally avoiding the influence of noise changes on the recognition of the third audio information, and can accurately identify the control command information, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] The smart home system and control method of the present invention will be described below with reference to the accompanying drawings and in combination with a washing machine, wherein:
[0020] Figure 1 is a structural diagram of the smart home system of the present invention;
[0021] Figure 2 is a flow chart of the control method of the present invention;
[0022] Figure 3 is a flow chart of a control method for causing a washing machine to enter a voice training mode according to a first embodiment of the present invention;
[0023] Figure 4 is a flow chart of a control method for causing a washing machine to enter a voice training mode according to a second embodiment of the present invention;
[0024] Figure 5 is a flow chart of a control method for a washing machine of the present invention to send second audio information to a server;
[0025] Figure 6 is a flow chart of a control method of the server of the present invention for updating the first speech recognition model according to the second audio information;
[0026] Figure 7is a flow chart of a control method for a server of the present invention to selectively update a first dedicated voice recognition module and / or a first general voice recognition module;
[0027] Figure 8 is a flow chart of a control method for updating a second speech recognition model of a washing machine of the present invention;
[0028] Fig. 9 It is a logic diagram of the control method of the present invention.
[0029] Reference numerals list
[0030] 1. Washing machine; 2. Server. DETAILED DESCRIPTION
[0031] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention. For example, although the present application is described in conjunction with a washing machine, the technical solution of the present invention is not limited thereto, and the smart home system and control method can obviously also be applied to other household appliances such as dryers, integrated washing and drying machines, shoe washing machines, nursing machines, air conditioners, televisions, and refrigerators, and such changes do not deviate from the principles and scope of the present invention.
[0032] It should be noted that, in the description of the present invention, the terms “first”, “second”, “third” and “fourth” are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.
[0033] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0034] Based on the technical problems raised in the background technology, the present invention provides a control method for a smart home system, which aims to enable a server to update a first voice recognition model according to a second audio information when a first washing machine is in a voice training mode, so that the first voice recognition model can be updated in time, so that the updated first voice recognition model is more matched with the latest noise, and the input first audio information can be more effectively subjected to noise reduction processing, thereby fundamentally avoiding the influence of noise changes on the recognition of the first audio information, and can accurately identify control instruction information, thereby improving user experience.
[0035] See first Figure 1 , the smart home system of the present invention is described. Among them, Figure 1 It is a structural diagram of the smart home system of the present invention.
[0036] like Figure 1 As shown, the smart home system of the present invention includes a washing machine 1 and a server 2. The washing machine 1 can communicate with the server 2. The washing machine 1 has a voice training mode. The server 2 has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on the input first audio information.
[0037] Preferably, the first speech recognition model includes a first dedicated speech recognition module and a first general speech recognition module, wherein the first dedicated speech recognition module is a dedicated speech recognition module suitable for the washing machine 1 that sends the second audio information to the server 2; the first general speech recognition module is a general speech recognition module suitable for all washing machines 1 of the same model as the washing machine 1 that sends the second audio information to the server 2.
[0038] Furthermore, the number of first dedicated voice recognition modules is multiple, for example, two, six, ten or more, and accordingly, the number of washing machines 1 is also multiple, for example, two, six, ten or more, and each washing machine 1 corresponds to a first dedicated voice recognition module associated with its MAC address.
[0039] Furthermore, the number of the first universal voice recognition modules is multiple, for example, two, three, four or more, and accordingly, the models of the washing machine 1 are also multiple, for example, two, three, four or more, and washing machines 1 of the same model all correspond to the first universal voice recognition module associated with the model.
[0040] Preferably, the first speech recognition model may be a Gaussian mixture model, a mixed Laplace model, a continuous hidden Markov model, a semi-continuous hidden Markov model, a discrete hidden Markov model or other speech recognition models.
[0041] Although not shown in the drawings, the washing machine 1 includes a recording module, a communication module, a second voice recognition model, a voice module and a controller.
[0042] The recording module is used to record the second audio information near the washing machine 1. Preferably, the recording module can be a Mic recording module, a WTV-SR recording module, a WT8603 recording module, a NR7100S recording module or other recording modules.
[0043] The washing machine 1 communicates with the server 2 via a communication module. Preferably, the communication module may be a WiFi module, a Bluetooth module, a ZIGBEE module, an NFC module, a GPRS module or other communication modules.
[0044] The second speech recognition model is used to identify the control instruction information after performing noise reduction processing on the input third audio information. Preferably, the second speech recognition model can be a Gaussian mixture model, a mixed Laplace model, a continuous hidden Markov model, a semi-continuous hidden Markov model, a discrete hidden Markov model or other speech recognition models. It should be noted that the first audio information and the third audio information can be the same or different.
[0045] Preferably, the second speech recognition model includes a second dedicated speech recognition module and a second general speech recognition module, wherein the second dedicated speech recognition module is a speech recognition module dedicated to the washing machine 1; the second general speech recognition module is a speech recognition module common to all washing machines 1 of the same model as the washing machine 1.
[0046] Among them, the voice module includes a voice interaction submodule, a storage submodule and a data processing submodule. The voice interaction submodule is used to obtain the user's first audio information, and the voice interaction submodule can send the first audio information to the server 2 through the communication module; the storage submodule is used to store the second audio information recorded by the recording module; the data processing submodule is provided with an audio processing chip, and the audio processing chip is used to mark the noise type on the second audio information according to the attribute of the second audio information, and convert the second audio information marked with the noise type into a digital signal, and send it to the server 2 in the form of a digital signal, wherein the attributes include frequency, amplitude and signal strength. Of course, the attributes can also include only frequency, amplitude or signal strength, or include any two of frequency, amplitude and signal strength, and those skilled in the art can flexibly adjust and set according to actual usage requirements.
[0047] Preferably, the voice interaction submodule is an SDK module, NV020S module, YQ5969 module, NRK10B or other voice modules.
[0048] Preferably, the storage submodule can store the audio information recorded by the recording module in real time, or can store the audio information recorded by the recording module in batches according to preset time intervals, for example, storing the second audio information once every 5 seconds, or storing the second audio information once every 10 seconds. Technical personnel in this field can flexibly adjust the settings according to the storage capacity of the storage submodule.
[0049] Among them, the controller is connected to the communication module, the recording module and the voice module respectively. The controller can control the washing machine 1 to enter the voice training mode. The controller can also control the communication module to send the first audio information to the server 2 and receive the control instruction information of the recognized first audio information from the server 2; the controller can also control the communication module to send the second audio information to the server 2 and receive the first dedicated voice recognition module and the first general voice recognition module updated according to the second audio information from the server 2, and control the washing machine 1 to update the second dedicated voice recognition module and the second general voice recognition module of the second voice recognition model according to the updated first dedicated voice recognition module and the first general voice recognition module.
[0050] Among them, the speech training mode is a mode used to train and optimize the first speech recognition model and the second speech recognition model, so that the optimized first speech recognition model and the second speech recognition model are more matched with the latest noise.
[0051] Furthermore, the washing machine 1 is also provided with an operation key, which is configured to cause the washing machine 1 to enter the voice training mode when operated. The operation key may be a key, an operation knob, a touch key or other types of operation keys.
[0052] Preferably, the server 2 can receive the first audio information from the washing machine 1 and send the control instruction information of the recognized first audio information to the washing machine 1; the server 2 can also receive the second audio information from the washing machine 1 and update the first voice recognition model according to the second audio information, for example, updating the first dedicated voice recognition module and the first general voice recognition module; the server 2 can also send the updated first dedicated voice recognition module and the first general voice recognition module to the washing machine 1, so that the washing machine 1 updates the second dedicated voice recognition module and the second general voice recognition module according to the updated first dedicated voice recognition module and the first general voice recognition module respectively.
[0053] The server 2 may be, but is not limited to, a cloud server and a backend server.
[0054] It should be noted that, although the above example shows that the washing machine 1 performs the work of "storing audio information and marking noise types on digital signals", this is not restrictive, but only exemplary, and the server 2 of the smart home system can also perform the work of "storing audio information and marking noise types on digital signals". Those skilled in the art can flexibly adjust and set the execution subject of the above work, and the present invention does not impose any limitation on this.
[0055] The following takes a washing machine as an example to further illustrate the control method for a smart home system of the present invention.
[0056] See also Figure 2, the control method for the smart home system of the present invention is described. Figure 2 is a flow chart of the control method of the present invention.
[0057] like Figure 2 As shown, the control method for the smart home system of the present invention comprises the following steps:
[0058] S100, when the washing machine is in a voice training mode, the washing machine obtains second audio information;
[0059] S200, the washing machine sends the second audio information to the server;
[0060] S300. The server updates the first speech recognition model according to the second audio information.
[0061] In step S100, the washing machine records the second audio information in its vicinity through the recording module.
[0062] Refer to the following Figure 3 and Figure 4 , two embodiments of the control method of the present invention for making a washing machine enter a voice training mode are described. Among them, Figure 3 is a flow chart of a control method for causing a washing machine to enter a voice training mode according to a first embodiment of the present invention; Figure 4 4 is a flow chart of a control method for causing a washing machine to enter a voice training mode according to a second embodiment of the present invention.
[0063] Embodiment 1
[0064] like Figure 3 As shown, before step S100 "when the washing machine is in the voice training mode, the washing machine obtains the second audio information", the control method further includes:
[0065] S011. The washing machine obtains fourth audio information of the user;
[0066] S012, determining whether valid control instruction information is obtained;
[0067] S013: If the washing machine does not obtain valid control instruction information, the washing machine enters a voice training mode;
[0068] S014: If the washing machine obtains valid control instruction information, the washing machine does not enter the voice training mode.
[0069] Among them, effective control instructions refer to control instructions that can reflect the user's clear control purpose, for example, start the washing program immediately, wash for 30 minutes, schedule to start washing at 8 o'clock, spin dry for 5 minutes, rinse 3 times, cancel the prompt sound, add a sterilization program, and other control instructions with clear control purposes.
[0070] The fourth audio information and the first audio information may be the same or different.
[0071] In step S011, when the washing machine is in a standby state, the fourth audio information of the user is obtained. Of course, the fourth audio information of the user can also be obtained when the washing machine is just started; or, the fourth audio information of the user can also be obtained when the washing machine ends. Those skilled in the art can flexibly adjust and set the timing of obtaining the fourth audio information of the user. No matter how the timing of obtaining the fourth audio information of the user is adjusted and set, it does not affect the normal operation of the washing machine and the user's experience.
[0072] In step S012, the first speech recognition model of the server can be used to perform noise reduction processing on the input fourth audio information and then identify the control instruction information; of course, the second speech recognition model of the washing machine can also be used to perform noise reduction processing on the input fourth audio information and then identify the control instruction information. Those skilled in the art can flexibly adjust and set the speech recognition model for performing noise reduction processing on the input fourth audio information and identifying the control instruction information, and the present invention does not impose any restrictions on this.
[0073] In step S013, if the washing machine does not obtain valid control command information from the fourth audio information, it means that the user has not performed voice interaction with the washing machine, or the valid control command information cannot be accurately identified from the fourth audio information due to noise elimination failure and other reasons. At this time, the audio information obtained may be noise, or it may be the sound of communication between users, or a sound mixed with noise and control commands. Then, if making the washing machine enter the voice training mode will not affect the user's use, the washing machine is made to enter the voice training mode, so that the first voice recognition model can be updated in time, thereby improving the user's use experience; if making the washing machine enter the voice training mode affects the user's use, the user can be prompted to manually enter the command or re-enter the fourth audio information, and then the fourth audio information just entered is stored and the background processing transmission service is used for learning.
[0074] In step S014, if the washing machine obtains valid control instruction information from the fourth audio information, it means that the user is interacting with the washing machine by voice. For example, the user is setting the washing program by voice, such as heating washing for 30 minutes, heating temperature of 40°C, etc. If the washing machine enters the voice training mode at this time, it will affect the user's use. In this case, the washing machine does not enter the voice training mode and continues to interact with the user by voice.
[0075] It should be further explained that, in the above process, step S013 and step S014 have no order of precedence and are parallel, and are only related to the judgment result of whether valid control instruction information is obtained. The corresponding steps can be executed according to different judgment results.
[0076] Embodiment 2
[0077] like Figure 4 As shown, before step S100 "when the washing machine is in the voice training mode, the washing machine obtains the second audio information", the control method further includes:
[0078] S021. Get the status of the operation key;
[0079] S022, determining whether the operation key is operated;
[0080] S023. If the operation key is operated, the washing machine enters a voice training mode;
[0081] S024: If the operation key is not operated, the washing machine does not enter the voice training mode.
[0082] The state of the operation key includes a state where the operation key is operated and a state where the operation key is not operated.
[0083] In steps S021 and S022, when the washing machine is in the standby state, the state of the operation key is obtained and it is determined whether the operation key is operated. Of course, the state of the operation key can also be obtained and it is determined whether the operation key is operated when the washing machine is just started; or, the state of the operation key can also be obtained and it is determined whether the operation key is operated after the washing machine has finished running. Those skilled in the art can flexibly adjust and set the timing of obtaining the state of the operation key and determining whether the operation key is operated. No matter how the timing of obtaining the state of the operation key and determining whether the operation key is operated is adjusted and set, it does not affect the normal operation of the washing machine and the user's experience.
[0084] In step S023, if the operation key is operated, it means that the user believes that the washing machine cannot accurately recognize its control instruction information, which affects the user's experience. The user hopes to update the first voice recognition model, so the washing machine enters the voice training mode.
[0085] In step S024, if the operation key is not operated, it means that the user believes that the washing machine can accurately recognize its control instruction information and the user does not want to update the first voice recognition model yet, so the washing machine does not enter the voice training mode.
[0086] It should be further explained that, in the above process, step S023 and step S024 have no order of precedence and are parallel, and are only related to the judgment result of whether the operation key is operated. The corresponding steps can be executed according to different judgment results.
[0087] Refer to the following Figure 5 , taking the attributes of the second audio information including frequency, amplitude and signal strength as an example, the control method of the washing machine of the present invention for sending the second audio information to the server is described. Among them, Figure 5 It is a flow chart of a control method for a washing machine of the present invention to send second audio information to a server.
[0088] like Figure 5 As shown, in step S200, the step of "the washing machine sends the second audio information to the server" specifically includes:
[0089] S211, the washing machine stores the second audio information;
[0090] S212: The washing machine marks the noise type on the second audio information according to the attribute of the second audio information;
[0091] S213: The washing machine sends the second audio information marked with the noise type to the server in the form of a digital signal.
[0092] Among them, attributes include frequency, amplitude, and signal strength.
[0093] Among them, the noise types include the first type of noise and the second type of noise. The first type of noise is mainly noise that is unique to the washing machine, that is, the noise of the environment in which the washing machine is used, such as household noise, dry cleaning noise, hospital laundry room noise and other environmental noises; the second type of noise is mainly noise common to all washing machines of the same model, that is, the noise of the washing machine itself, such as motor noise, dehydration noise, washing tub rotation noise and other noises of the washing machine itself. Of course, the classification method of noise types is not limited to the above-mentioned methods, and can also be classified according to the decibel value of the noise, and the noise with a decibel value greater than the first preset value is set as the first type of noise, the noise with a decibel value greater than the second preset value and less than or equal to the first preset value is set as the second type of noise, and the noise with a decibel value less than or equal to the second preset value is set as the third type of noise. No matter which noise classification method is adopted, as long as different noises can be distinguished, it is sufficient.
[0094] In step S211, the storage submodule of the washing machine can store the audio information recorded by the recording module in real time, or store the audio information recorded by the recording module in batches according to preset time intervals, for example, storing the second audio information once every 5 seconds, or storing the second audio information once every 10 seconds. Technical personnel in this field can flexibly adjust and set the preset time interval according to the storage capacity of the storage submodule.
[0095] In step S212, the audio processing chip of the washing machine determines the type of noise included in the second audio information according to the frequency, amplitude and signal strength of the second audio information, such as household environmental noise, motor noise, dehydration noise, washing tub rotation noise and other noises, and marks noise labels at corresponding positions of the second audio information according to different noise types. The specific method of determining the noise type according to the frequency, amplitude and signal strength, and marking the noise label is common knowledge in the art and will not be repeated here.
[0096] In step S213, the washing machine sends the second audio information with the noise label to the server in the form of a digital signal, so that the server can directly determine the noise type of the second audio information based on the noise label, and selectively update the first dedicated voice recognition module and / or the first general voice recognition module based on the noise type. The specific method is detailed in the corresponding description in step S312 below and will not be repeated here.
[0097] It should be noted that, in the above process, the execution order and method of step S211 and step S212 are not limited to the order listed above. Step S212 may be executed first and then step S211, or step S211 and step S212 may be executed at the same time. Those skilled in the art can flexibly adjust and set the execution order of step S211 and step S212, and the present invention does not impose any limitations on this.
[0098] In addition, it should be noted that although step S211 lists the washing machine as executing the task of "storing the second audio information" and step S212 lists the washing machine as executing the task of "marking the noise type on the second audio information according to the attributes of the second audio information", this is not restrictive but only exemplary. The tasks of "storing the second audio information" and "marking the noise type on the second audio information according to the attributes of the second audio information" can also be completed by the server. Those skilled in the art can flexibly adjust and set the execution subject of the task, and the present invention does not impose any restrictions on this.
[0099] Refer to the following Figure 6 and Figure 7 , a control method for the server of the present invention to update the first speech recognition model according to the second audio information is described. Among them, Figure 6 is a flow chart of a control method of the server of the present invention for updating the first speech recognition model according to the second audio information; Figure 7 It is a flow chart of a control method for a server of the present invention to selectively update a first dedicated voice recognition module and / or a first general voice recognition module. Figure 6 and Figure 7The example in which there is a voice training model and a first voice recognition model on the server, but there is no second voice recognition model on the washing machine is further explained.
[0100] like Figure 6 As shown, in step S300, the step of "the server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module according to the noise type and the second audio information" specifically includes:
[0101] S311. The server obtains a noise type of the second audio information;
[0102] S312: According to the noise type and the second audio information, the server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module.
[0103] The control method of the present invention is further described below by taking the first type of noise being mainly noise specific to a washing machine and the second type of noise being mainly noise common to all washing machines of the same model as the washing machine as an example.
[0104] like Figure 7 As shown, in step S312, the step of "the server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module according to the noise type and the second audio information" specifically includes:
[0105] S321: If the noise type is the first type of noise, the server obtains the MAC address of the washing machine that sends the second audio information, and the server updates the first dedicated voice recognition module corresponding to the MAC address according to the second audio information;
[0106] S322: If the noise type is the second type of noise, the server updates the first general speech recognition module according to the second audio information;
[0107] S323. If the noise type includes the first type of noise and the second type of noise, the server obtains the MAC address of the washing machine that sends the second audio information, extracts the first noise information belonging to the first type of noise from the second audio information, and updates the first dedicated voice recognition module corresponding to the MAC address according to the first noise information. The server extracts the second noise information belonging to the second type of noise from the second audio information, and updates the first general voice recognition module according to the second noise information.
[0108] In step S321, if the noise type is the first type of noise, it means that the noise of the second audio information is the noise specific to the washing machine that sends the second audio information, and the noise is only applicable to the washing machine, not to other washing machines. For example, the washing machine is placed in the user's home, and the noise of the second audio information is the noise specific to the user's home, and the noise in each home is different. Therefore, the first general voice recognition module corresponding to the washing machine that sends the second audio information cannot be updated according to the second audio information, and only the first dedicated voice recognition module corresponding to the washing machine that sends the second audio information can be updated. In order to accurately update the first dedicated voice recognition module corresponding to the washing machine that sends the second audio information, the server obtains the MAC address of the washing machine that sends the second audio information, and the server updates the first dedicated voice recognition module corresponding to the MAC address according to the second audio information, so that the updated first dedicated voice recognition module can effectively perform noise reduction processing on the noise specific to the first audio information input by the washing machine, thereby accurately identifying the control instruction information.
[0109] In step S322, if the noise type is the second type of noise, it means that the noise of the second audio information is the noise common to all washing machines of the same model as the washing machine that sends the second audio information, and the noise is applicable to all washing machines of the same model as the washing machine that sends the second audio information. For example, the noise of the second audio information is the noise of the motor of the washing machine, and the noise generated by the motors of washing machines of the same model should be the same, close or similar. Therefore, the first general voice recognition module corresponding to the washing machine that sends the second audio information can be updated according to the second audio information, and there is no need to update the first special voice recognition module corresponding to the washing machine that sends the second audio information, so that the updated first general voice recognition module can effectively perform noise reduction processing on the noise common to all washing machines of the same model as the washing machine and the input first audio information, so as to accurately identify the control instruction information.
[0110] In step S323, if the noise type includes the first type of noise and the second type of noise, it means that the noise of the second audio information includes both the noise unique to the washing machine that sends the second audio information and the noise common to all washing machines of the same model as the washing machine that sends the second audio information. For example, the washing machine is placed in a dry cleaning shop, and the noise of the second audio information includes the noise unique to the dry cleaning shop and the noise of the rotation of the washing tub. Both the first general voice recognition module corresponding to the washing machine that sends the second audio information and the first special voice recognition module corresponding to the washing machine that sends the second audio information can be updated. Therefore, the server obtains the MAC address of the washing machine that sends the second audio information, extracts the first noise information belonging to the first category of noise from the second audio information, and updates the first dedicated voice recognition module corresponding to the MAC address according to the first noise information, so that the updated first dedicated voice recognition module can effectively reduce the noise unique to the first audio information input by the washing machine; and the server extracts the second noise information belonging to the second category of noise from the second audio information, and updates the first general voice recognition module corresponding to the washing machine that sends the second audio information according to the second noise information, so that the updated first general voice recognition module can effectively reduce the noise common to the first audio information input by all washing machines of the same model as the washing machine, thereby further improving the user experience.
[0111] It should be noted that, in the above process, steps S321 to S323 have no order but are parallel and are only related to the noise type. Corresponding steps can be executed according to different noise types.
[0112] See next Figure 8 , the control method for the smart home system of the present invention is further described. Among them, Figure 8 It is a flow chart of a control method for updating a second speech recognition model of a washing machine according to the present invention. Figure 8 The example in which a second speech recognition model is provided on a washing machine, a speech training model is provided on a server, and a first speech recognition model is provided is further explained.
[0113] like Figure 8 As shown, after step S312 "according to the noise type and the second audio information, the server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module", the control method further includes:
[0114] S411, the server sends the updated first dedicated voice recognition module and / or the first general voice recognition module to the washing machine;
[0115] S412: The washing machine updates the second dedicated voice recognition module and / or the second general voice recognition module according to the updated first dedicated voice recognition module and / or the first general voice recognition module.
[0116] Specifically, if in step S321, the server updates the first dedicated voice recognition module corresponding to the MAC address of the washing machine according to the second audio information, the server will send the updated first dedicated voice recognition module to the washing machine corresponding to the MAC address according to the MAC address, so that the washing machine corresponding to the MAC address can update the second dedicated voice recognition module according to the updated first dedicated voice recognition module, so that the updated second voice recognition model can also better match the latest first-class noise, and can more effectively perform noise reduction processing on the input third audio information, fundamentally avoiding the influence of changes in the first-class noise on the recognition of the third audio information, and can accurately recognize the third audio information.
[0117] Furthermore, if in step S322, the server updates the first universal voice recognition module corresponding to the model of the washing machine that sends the second audio information according to the second audio information, the server will send the updated first universal voice recognition module to the washing machines corresponding to the MAC addresses according to the MAC addresses of all washing machines with the same model as the washing machine that sends the second audio information, so that all washing machines with the same model as the washing machine that sends the second audio information can update the second universal voice recognition module according to the updated first universal voice recognition module, so that the updated second voice recognition model can also better match the latest second-type noise, and can more effectively perform noise reduction processing on the input third audio information, fundamentally avoiding the influence of changes in the second-type noise on the recognition of the third audio information, and can accurately recognize the third audio information.
[0118] Further, if in step S323, the server updates the first dedicated voice recognition module corresponding to the MAC address of the washing machine and the first general voice recognition module corresponding to the model of the washing machine that sends the second audio information according to the second audio information, the server will send the updated first dedicated voice recognition module and the first general voice recognition module to the washing machine corresponding to the MAC address according to the MAC address, so that the washing machine corresponding to the MAC address can update the second dedicated voice recognition module according to the updated first dedicated voice recognition module, and update the second general voice recognition module according to the updated first general voice recognition module, so that the updated second voice recognition model can match the latest first and second type noise, and can more effectively perform noise reduction processing on the input third audio information, fundamentally avoiding the influence of noise changes on the recognition of the third audio information, and can accurately recognize the third audio information.
[0119] Furthermore, the server will also send the updated first universal voice recognition module to the washing machines corresponding to the MAC addresses according to the MAC addresses of other washing machines with the same model as the washing machine that sends the second audio information, so that other washing machines with the same model as the washing machine that sends the second audio information can also update the second universal voice recognition module according to the updated first universal voice recognition module, so that the updated second voice recognition model can also better match the latest second-type noise, and can more effectively perform noise reduction processing on the input third audio information, fundamentally avoiding the influence of changes in the second-type noise on the recognition of the third audio information, and can accurately recognize the third audio information.
[0120] It should be noted that the first audio information and the third audio information may be the same or different.
[0121] Refer to the following Fig. 9 , a possible control process of the present invention is introduced. Among them, Fig. 9 is a logic diagram of the control method of the present invention. Fig. 9 The example in which the fourth audio information is the same as the first audio information is further explained.
[0122] like Fig. 9 As shown, a possible complete process of the control method of the present invention is:
[0123] S501, when the washing machine is in a standby state, the washing machine obtains first audio information of a user;
[0124] S502, determine whether valid control instruction information is obtained; if so, execute step S503; if not, execute step S504;
[0125] S503, preventing the washing machine from entering the voice training mode;
[0126] S504, making the washing machine enter a voice training mode;
[0127] After step S504, execute step S505;
[0128] S505, the washing machine obtains second audio information;
[0129] S506, the washing machine stores the second audio information;
[0130] S507: The washing machine marks the noise type on the second audio information according to the attributes of the second audio information; wherein the attributes include frequency, amplitude and signal strength;
[0131] S508, the washing machine sends the second audio information marked with the noise type to the server in the form of a digital signal;
[0132] S509: The server obtains the noise type of the second audio information;
[0133] S510: If the noise type is the first type of noise, the server obtains the MAC address of the washing machine that sends the second audio information, and the server updates the first dedicated voice recognition module corresponding to the MAC address according to the second audio information;
[0134] S511: If the noise type is the second type of noise, the server updates the first universal voice recognition module corresponding to the model of the washing machine that sends the second audio information according to the second audio information;
[0135] S512: if the noise type includes the first type of noise and the second type of noise, the server obtains the MAC address of the washing machine that sends the second audio information, the server extracts first noise information belonging to the first type of noise from the second audio information, the server updates a first dedicated voice recognition module corresponding to the MAC address according to the first noise information, and the server extracts second noise information belonging to the second type of noise from the second audio information, and the server updates a first general voice recognition module corresponding to the model of the washing machine that sends the second audio information according to the second noise information;
[0136] After step S510, execute step S513 and step S514;
[0137] S513, the server sends the updated first dedicated voice recognition module to the washing machine according to the MAC address of the washing machine that sends the second audio information;
[0138] S514, the washing machine that sends the second audio information updates the second dedicated voice recognition module according to the updated first dedicated voice recognition module;
[0139] After step S511, execute step S515 and step S516;
[0140] S515, the server sends the updated first universal voice recognition module to the washing machines corresponding to the MAC addresses of all washing machines of the same model as the washing machine that sends the second audio information;
[0141] S516, all washing machines of the same model as the washing machine that sent the second audio information update their second general voice recognition modules according to the updated first general voice recognition module;
[0142] After step S512, execute steps S517 and S518;
[0143] S517, the server sends the updated first dedicated voice recognition module to the washing machine according to the MAC address of the washing machine that sends the second audio information; and the server sends the updated first general voice recognition module to the washing machines corresponding to the MAC addresses of all washing machines of the same model as the washing machine that sends the second audio information respectively;
[0144] S518. The washing machine that sends the second audio information updates its second dedicated voice recognition module and its second general voice recognition module respectively according to the updated first dedicated voice recognition module and the first general voice recognition module, and other washing machines of the same model as the washing machine that sends the second audio information update their second general voice recognition modules respectively according to the updated first general voice recognition module.
[0145] It should be pointed out that the above embodiment is only a preferred implementation of the present invention, which is only used to illustrate the principle of the method of the present invention, and is not intended to limit the scope of protection of the present invention. In practical applications, those skilled in the art can allocate the above functions to different steps as needed, that is, decompose or combine the steps in the embodiment of the present invention. For example, the steps of the above embodiment can be combined into one step, or further divided into multiple sub-steps to complete all or part of the functions described above. The names of the steps involved in the embodiments of the present invention are only for distinguishing the steps and are not regarded as limitations of the present invention.
[0146] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.
Claims
1. A control method for a smart home system, characterized in that: The smart home system includes a home appliance and a server, the home appliance can communicate with the server, the home appliance has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on input first audio information; The control method comprises the following steps: When the household appliance is in the voice training mode, the household appliance acquires second audio information; The household appliance sends the second audio information to the server; The server updates the first speech recognition model according to the second audio information; The first speech recognition model includes a first dedicated speech recognition module and a first general speech recognition module; The step of "the server updates the first speech recognition model according to the second audio information" specifically includes: Acquire the noise type of the second audio information; The server selectively updates the first dedicated speech recognition module and / or the first general speech recognition module according to the noise type and the second audio information; The noise includes the first type of noise and the second type of noise; The step of "according to the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module" specifically includes: if the noise type is the first type of noise, the server obtains the MAC address of the household appliance that sends the second audio information, and the server updates the first dedicated voice recognition module corresponding to the MAC address according to the second audio information; or if the noise type is the second type of noise, the server updates the first general voice recognition module according to the second audio information; Alternatively, the step of "the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module based on the noise type and the second audio information" also includes: if the noise type includes the first category of noise and the second category of noise, the server obtains the MAC address of the household appliance that sends the second audio information, the server extracts the first noise information belonging to the first category of noise from the second audio information, the server updates the first dedicated voice recognition module corresponding to the MAC address based on the first noise information, and the server extracts the second noise information belonging to the second category of noise from the second audio information, and the server updates the first general voice recognition module based on the second noise information.
2. The control method according to claim 1, characterized in that: The household appliance has a second voice recognition model, the second voice recognition model includes a second dedicated voice recognition module and a second general voice recognition module, wherein the second voice recognition model is used to identify control instruction information after performing noise reduction processing on the input third audio information; After "according to the noise type and the second audio information, the server selectively updates the first dedicated voice recognition module and / or the first general voice recognition module", the control method further includes: The server sends the updated first dedicated voice recognition module and / or the first general voice recognition module to the household appliance; The household appliance updates the second dedicated voice recognition module and / or the second general voice recognition module according to the updated first dedicated voice recognition module and / or the first general voice recognition module.
3. The control method according to claim 1 or 2, characterized in that: The step of "the household appliance sending the second audio information to the server" specifically includes: The household appliance stores the second audio information; The household appliance marks the noise type on the second audio information according to the attribute of the second audio information; The household appliance sends the second audio information marked with the noise type to the server in the form of a digital signal; The attribute includes at least one of frequency, amplitude and signal strength.
4. The control method according to claim 1 or 2, characterized in that: The household appliance is provided with an operation key, and the operation key is configured to cause the household appliance to enter the voice training mode when operated; Before “when the household appliance is in the voice training mode, the household appliance acquires second audio information”, the control method further includes: Determining whether the operation key is operated; If the operation key is operated, the household appliance enters the voice training mode; or If the operation key is not operated, the home appliance does not enter the voice training mode.
5. The control method according to claim 1 or 2, characterized in that: Before “when the household appliance is in the voice training mode, the household appliance acquires second audio information”, the control method further includes: The household appliance acquires fourth audio information of the user; Determine whether valid control instruction information is obtained; If the household appliance fails to obtain valid control instruction information, the household appliance is caused to enter the voice training mode; or If the household appliance obtains valid control instruction information, the household appliance is prevented from entering the voice training mode.
6. A control method for a smart home system, characterized in that: The smart home system includes a home appliance and a server, the home appliance can communicate with the server, the home appliance has a voice training mode, and the server has a first voice recognition model, wherein the first voice recognition model is used to identify control instruction information after performing noise reduction processing on input first audio information; The control method comprises the following steps: In a case where the home appliance is in the voice training mode, the server receives second audio information from the home appliance; According to the control method according to any one of claims 1 to 5, the server updates the first speech recognition model according to the second audio information.
