Electric blanket control method and system based on AI voice interaction

Abstract

The invention discloses an electric blanket control method and system based on AI voice interaction, and the method comprises the steps: collecting a noisy voice signal through a microphone array, carrying out the filtering of an analog front-end circuit, the echo elimination of an AEC chip, and the noise reduction of a DSP, and obtaining a high-signal-to-noise-ratio pure voice. The AI or voice recognition chip extracts features and matches the features with a pre-stored template, and the master control MCU drives the voltage regulation circuit to regulate the heating power according to an instruction through PWM (Pulse Width Modulation). And dynamically adjusting the power of each region according to the temperature difference by using a closed-loop control algorithm according to ergonomic partitioning. The intelligent electric blanket solves the problems that a traditional electric blanket is inconvenient to operate, inaccurate in temperature adjustment and lack of effective safety protection, and an existing intelligent electric blanket is insufficient in voice interaction function, convenient voice control, accurate partition temperature control and comprehensive safety protection are achieved, and the user experience is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric blankets, and particularly relates to a control method and system for an electric blanket based on AI voice interaction. Background Art

[0002] As a common heating device, the traditional electric blanket provides warmth for people in cold seasons. Its working principle is mainly to generate heat by energizing the heating wire, converting electrical energy into heat energy, so as to increase the temperature of the quilt. With the rapid development of artificial intelligence technology, voice interaction technology has gradually become one of the important ways of human-computer interaction. Voice interaction technology allows users to interact with devices through voice commands without manual operation, greatly improving the convenience and efficiency of operation. In the field of smart home, voice interaction technology has been widely applied. Devices such as smart speakers, smart air conditioners, and smart lights all support voice control functions. Users can easily control the on / off, temperature adjustment, brightness adjustment and other parameters of these devices through voice commands, realizing an intelligent home life.

[0003] However, realizing the control of an electric blanket based on AI voice interaction also faces some challenges. First of all, the acquisition and processing of voice signals is a key issue. The usage environment of the electric blanket is usually relatively complex, and there may be various noise interferences, such as the sound of a fan, the sound of a TV, etc., which will affect the quality of the voice signal and lead to a decrease in the accuracy of voice recognition. Secondly, how to achieve accurate voice command parsing and control is also a difficult problem. There are differences in the voice habits and accents of different users, and it is necessary to develop highly adaptable and accurate voice recognition algorithms and command parsing models. In addition, the safety performance of the electric blanket needs to be considered to ensure that there are no safety problems such as overheating and electric leakage during the voice control process.

[0004] Although there have already been some electric blankets with intelligent functions on the market, the intelligence level of most products is relatively low and still cannot meet the needs of users for convenience, comfort and safety. Some so-called intelligent electric blankets only simply add the function of controlling through a mobile phone APP and do not really achieve voice interaction. For the few electric blanket products that support voice control, there are still obvious deficiencies in aspects such as voice recognition accuracy, command parsing ability and safety protection mechanisms. For example, in a complex noise environment, the accuracy of voice recognition is relatively low, and misrecognition or non-recognition is likely to occur; in terms of command parsing, only a limited number of simple commands can be recognized, which cannot meet the diverse needs of users. Summary of the Invention

[0005] To this end, the present invention provides an electric blanket control method and system based on AI voice interaction, which solves the problems of inconvenient operation, inaccurate temperature regulation, lack of effective safety protection of traditional electric blankets, and insufficient voice interaction functions of existing intelligent electric blankets, realizes convenient voice control, accurate zoned temperature control, and all-round safety protection, and improves the user experience.

[0006] To achieve the above object, the present invention provides the following technical solutions: An electric blanket control method based on AI voice interaction, comprising the following steps:

[0007] Use a microphone array to collect voice signals containing ambient noise, filter the voice signals containing ambient noise through an analog front-end circuit to remove interference signals; then, an AEC chip eliminates acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi-band noise suppression to output a pure voice signal with a signal-to-noise ratio ≥ 70dB.

[0008] Extract features from the preprocessed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre-stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized after the match, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi-way thyristor voltage regulating circuit according to the received digital control signal.

[0009] According to the ergonomic zoning model, the electric blanket is divided into multiple heating zones to adapt to different temperature requirements of different parts of the human body; during the process of adjusting the heating power of the electric blanket, a closed-loop control algorithm is adopted to collect the temperature data of each heating zone of the electric blanket in real time, and according to the difference between the set temperature and the actual temperature, dynamically adjust the heating power of each heating zone of the electric blanket.

[0010] As a preferred solution of the electric blanket control method based on AI voice interaction, the microphone array is a linear array or a circular array composed of at least three microphones, and the spacing between each microphone is optimized according to the frequency range of the voice signal and the expected noise reduction effect;

[0011] The low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal;

[0012] The anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal;

[0013] When the AEC chip eliminates acoustic echo based on the adaptive filter algorithm, it dynamically adjusts the step size according to the magnitude of the error signal, and adaptively adjusts the filtering parameters according to different acoustic environments through the acoustic model built in the AEC chip. The formula for eliminating acoustic echo based on the adaptive filter algorithm is:

[0014]

[0015] In the formula, w k is the filter coefficient vector, μ is the step size factor, e k is the error signal, and x k is the reference input signal.

[0016] As a preferred solution of the electric blanket control method based on AI voice interaction, when the DSP performs multi-band noise suppression, it divides the voice signal into multiple bands, and each band adopts an independent noise estimation and suppression strategy; for the low-frequency band, spectral subtraction combined with Wiener filtering is used for noise suppression; for the high-frequency band, a method based on sub-band spectral estimation is used to improve the clarity of the high-frequency voice signal; the DSP monitors the signal-to-noise ratio of each band in real time. When the signal-to-noise ratio of a band is lower than the set threshold, the noise suppression intensity of the band is enhanced, and finally a pure voice signal with a signal-to-noise ratio ≥ 70 dB is output.

[0017] As a preferred solution of the electric blanket control method based on AI voice interaction, the closed-loop control algorithm uses a PID control model:

[0018]

[0019] In the formula, ΔD is the PWM duty cycle adjustment amount, ΔT is the difference between the set temperature and the actual temperature, and K p , K i , K d are the proportional, integral, and differential coefficients respectively;

[0020] The heating power adjustment formula is:

[0021] P output = P max ·D

[0022] In the formula, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for low gear, 30% for medium gear, 60% for high gear, and 90% for ultra-high temperature gear.

[0023] As a preferred solution of the electric blanket control method based on AI voice interaction, during the process of adjusting the heating power of the electric blanket, it also includes using a dual safety protection measure combining hardware-level overcurrent protection and software over-temperature detection to protect the electric blanket;

[0024] Hardware-level overcurrent protection process: A fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse or an overcurrent situation occurs, the fuse melts to cut off the circuit, and the self-resetting fuse restores the circuit connection after the overcurrent situation disappears.

[0025] Software over-temperature detection process: The main control MCU reads the data of the temperature sensor in the set heating area of the electric blanket in real time. When it detects that the temperature in the set heating area of the electric blanket exceeds the set safety threshold, it cuts off the power supply of the heating circuit corresponding to the set heating area.

[0026] As a preferred solution of the electric blanket control method based on AI voice interaction, in the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is:

[0027]

[0028] In the formula, P max is the maximum heating power, and R load is the load resistance;

[0029] Software over-temperature detection process, the software over-temperature detection threshold T alert The formula is:

[0030] T alert = T set + ΔT`

[0031] In the formula, T set is the set temperature; ΔT` is the reserved safety margin.

[0032] The present invention also provides an electric blanket control system based on AI voice interaction, including:

[0033] A voice signal preprocessing module, which is used to collect voice signals containing environmental noise by using a microphone array, filter the voice signals containing environmental noise through an analog front-end circuit to remove interference signals; then, an AEC chip eliminates the acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi-band noise suppression to output a pure voice signal with a signal-to-noise ratio ≥ 70dB;

[0034] An instruction parsing and control module, which is used to extract features from the preprocessed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre-stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized after the match, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi-way thyristor voltage regulating circuit according to the received digital control signal;

[0035] The partition temperature control strategy adjustment module is used to divide the electric blanket into multiple heating zones according to the ergonomic partition model to meet the different temperature requirements of different parts of the human body; during the process of adjusting the heating power of the electric blanket, a closed-loop control algorithm is adopted to collect the temperature data of each heating zone of the electric blanket in real time, and the heating power of each heating zone of the electric blanket is dynamically adjusted according to the difference between the set temperature and the actual temperature.

[0036] As an optimal solution for the electric blanket control system based on AI voice interaction, in the voice signal preprocessing module: the microphone array is a linear array or a circular array composed of at least three microphones, and the spacing between each microphone is optimized according to the frequency range of the voice signal and the expected noise reduction effect;

[0037] The low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal;

[0038] The anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal;

[0039] When the AEC chip eliminates acoustic echo based on the adaptive filter algorithm, it dynamically adjusts the step size according to the magnitude of the error signal, and adaptively adjusts the filtering parameters according to different acoustic environments through the acoustic model built in the AEC chip. The formula for eliminating acoustic echo based on the adaptive filter algorithm is:

[0040]

[0041] In the formula, w k is the filter coefficient vector, μ is the step size factor, e k is the error signal, x k is the reference input signal.

[0042] As an optimal solution for the electric blanket control system based on AI voice interaction, when the DSP performs multi-band noise suppression, it divides the voice signal into multiple frequency bands, and each frequency band adopts an independent noise estimation and suppression strategy; for the low-frequency band, spectral subtraction combined with Wiener filtering is used for noise suppression; for the high-frequency band, a method based on sub-band spectral estimation is used to improve the clarity of the high-frequency voice signal; the DSP monitors the signal-to-noise ratio of each frequency band in real time. When the signal-to-noise ratio of a frequency band is lower than the set threshold, the noise suppression intensity of the frequency band is enhanced, and finally a pure voice signal with a signal-to-noise ratio ≥ 70dB is output.

[0043] As an optimal solution for the electric blanket control system based on AI voice interaction, in the partition temperature control strategy adjustment module, the closed-loop control algorithm adopts a PID control model:

[0044]

[0045] In the formula, ΔD is the PWM duty cycle adjustment amount, ΔT is the difference between the set temperature and the actual temperature, and K p , K i , K d are the proportional, integral, and differential coefficients respectively;

[0046] In the partition temperature control strategy adjustment module, the heating power adjustment formula is:

[0047] P output = P max ·D

[0048] In the formula, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for low gear, 30% for medium gear, 60% for high gear, and 90% for ultra-high temperature gear.

[0049] As an optimal solution for the electric blanket control system based on AI voice interaction, it further includes:

[0050] A safety protection module, which is used to adopt dual safety protection measures combining hardware-level overcurrent protection and software over-temperature detection to protect the electric blanket;

[0051] In the hardware-level overcurrent protection process, a fuse-type fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse-type fuse or an overcurrent situation occurs, the fuse-type fuse melts to cut off the circuit, and the self-resetting fuse restores the circuit connection after the overcurrent situation disappears;

[0052] In the software over-temperature detection process, the main control MCU reads the data of the temperature sensor in the set heating area of the electric blanket in real time. When it detects that the temperature in the set heating area of the electric blanket exceeds the set safety threshold, it cuts off the power supply of the heating circuit corresponding to the set heating area;

[0053] In the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is:

[0054]

[0055] In the formula, P max is the maximum heating power, and R load is the load resistance;

[0056] In the software over-temperature detection process, the software over-temperature detection threshold T alertThe formula is as follows:

[0057] T alert = T set + ΔT`

[0058] Wherein, T set is the set temperature; ΔT` is the reserved safety margin.

[0059] The beneficial effects of the present invention are as follows:

[0060] First, by using a microphone array, an analog front-end circuit, an AEC chip, and a DSP for preprocessing voice signals, it can effectively collect and process noisy voice signals and output pure voice signals with high signal-to-noise ratio. Whether it is an online AI chip or an offline voice recognition chip, it can accurately parse instructions, enabling the main control MCU to drive a multi-way thyristor voltage regulation circuit to adjust the heating power. Users can easily control the electric blanket just by speaking without manual operation, and can accurately recognize instructions even in complex environments, greatly improving the operation convenience.

[0061] Second, according to the ergonomic zoning model, the heating area is divided, and a closed-loop control algorithm is adopted to collect the temperature data of each area in real time and dynamically adjust the heating power. Accurate zoned temperature control is achieved, meeting the different temperature requirements of different parts of the human body. Compared with the single temperature setting of traditional electric blankets, the user comfort is greatly improved, and the accurate temperature control can avoid local overheating or overcooling, enhancing the use experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can be obtained based on the provided drawings.

[0063] The structures, proportions, sizes, etc. shown in this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the limited conditions for the implementation of the present invention. Therefore, they do not have technical essence. Any modification of the structure, change of the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed in the present invention.

[0064] Figure 1 is a schematic flow chart of the electric blanket control method based on AI voice interaction provided by the embodiment of the present invention;

[0065] Figure 2 is a temperature control schematic diagram in the electric blanket control method based on AI voice interaction provided by the embodiment of the present invention;

[0066] Figure 3 Schematic diagram of the control system architecture of the electric blanket based on AI voice interaction provided by the embodiments of the present invention. Specific implementation manners

[0067] The following specific embodiments illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0068] Embodiment 1

[0069] Refer to Figure 1 , Embodiment 1 of the present invention provides an electric blanket control method based on AI voice interaction, including the following steps:

[0070] S1. Use a microphone array to collect voice signals containing ambient noise, filter the voice signals containing ambient noise through an analog front-end circuit to remove interference signals; then, an AEC chip eliminates the acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi-band noise suppression to output a pure voice signal with a signal-to-noise ratio ≥ 70 dB.

[0071] In the actual use environment, there are various noises around. The microphone array can collect voice signals from multiple directions, increasing the comprehensiveness and accuracy of collection. The filter in the analog front-end circuit can specifically remove high-frequency or low-frequency interference components in the voice signal according to its frequency characteristics. The adaptive filter algorithm of the AEC chip continuously adjusts its own parameters to simulate the acoustic path from the speaker to the microphone, thereby canceling the echo. The noise characteristics of different frequency bands are different. The DSP divides the voice signal into multiple frequency bands, adopts different strategies to suppress noise according to the characteristics of each frequency band, and finally outputs a pure voice signal with a high signal-to-noise ratio, laying a foundation for accurately identifying voice commands subsequently.

[0072] S2. Extract features from the preprocessed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre-stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi-way thyristor voltage regulation circuit according to the received digital control signal.

[0073] Specifically, the AI chip or the voice recognition chip has powerful signal processing capabilities and can extract key features such as frequency and amplitude from the pure voice signal. The pre-stored instruction templates store the features of various preset voice instructions. By comparing and matching the extracted features with them, the control intention of the user can be recognized. For example, when the instruction "increase the temperature" is recognized, the chip generates a corresponding digital control signal. The main control MCU uses PWM technology to control the conduction time of the thyristor by changing the duty cycle of the pulse, thereby adjusting the voltage applied to the heating wire of the electric blanket and realizing the adjustment of the heating power to meet the user's temperature control requirements.

[0074] S3. According to the ergonomic zoning model, the electric blanket is divided into multiple heating zones to adapt to the different temperature requirements of different parts of the human body; in the process of adjusting the heating power of the electric blanket, a closed-loop control algorithm is adopted to collect the temperature data of each heating zone of the electric blanket in real time, and according to the difference between the set temperature and the actual temperature, dynamically adjust the heating power of each heating zone of the electric blanket.

[0075] Specifically, there are differences in the sensitivity and requirements of different parts of the human body for temperature. For example, the feet are usually colder than the chest. Dividing the heating zones based on the ergonomic zoning model can provide appropriate temperatures for different parts. The closed-loop control algorithm feeds back the actual temperature output to the input end for comparison with the set temperature. When the actual temperature is lower than the set temperature, the heating power is increased; when it is higher than the set temperature, the heating power is decreased. By continuously collecting temperature data in real time and adjusting the heating power, precise control of the temperature of each zone is achieved, improving the comfort of the user.

[0076] In this embodiment, in step S1, the microphone array is a linear array or a circular array composed of at least three microphones, and the spacing between the microphones is optimized according to the frequency range of the voice signal and the expected noise reduction effect.

[0077] Specifically, multiple microphones form an array, and the time difference and intensity difference of the signals received by different microphones can be used to enhance the useful signal and suppress noise. The linear array or the circular array can collect voice signals in different directions, expanding the collection range. The spacing between the microphones is crucial. If the spacing is too large, the signal collection may be discontinuous; if the spacing is too small, the signals collected by each microphone are similar and the signal differences cannot be effectively used for noise reduction. Optimizing the spacing according to the frequency range of the voice signal and the expected noise reduction effect can make the microphone array work better in different usage environments and improve the quality of voice signal collection.

[0078] In this embodiment, in step S1, the low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal; the anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal.

[0079] Specifically, the high gain-bandwidth product of the low-noise operational amplifier can ensure the effective amplification of weak voice signals within a wide frequency range, and the low input noise density ensures less additional noise introduced during the amplification process. Automatically adjusting the gain can adaptively amplify according to the strength of the input voice signal, avoiding distortion caused by overly strong or weak signals. The anti-aliasing filter uses a second-order or higher-order Butterworth filter because it has good frequency characteristics and can achieve steep attenuation at the cut-off frequency. Setting the cut-off frequency according to the highest frequency of the voice signal can effectively prevent noise above this frequency from mixing into the voice signal and ensure the quality of the subsequent processed voice signal.

[0080] Among them, when the AEC chip eliminates acoustic echo based on the adaptive filter algorithm, it dynamically adjusts the step size according to the magnitude of the error signal, and through the acoustic model built in the AEC chip, it adaptively adjusts the filtering parameters according to different acoustic environments. The formula for eliminating acoustic echo based on the adaptive filter algorithm is:

[0081]

[0082] In the formula, w k is the filter coefficient vector, μ is the step size factor, e k is the error signal, and x k is the reference input signal.

[0083] Specifically, the adaptive filter algorithm continuously adjusts the filter coefficient vector w k to match the acoustic path from the speaker to the microphone. The step size factor μ determines the adjustment speed of the filter coefficients. When the error signal e k is relatively large, increasing μ can speed up the adjustment speed and make the filter adapt to changes in the acoustic environment faster; when e k is relatively small, decreasing μ can avoid over-adjustment and make the filter more stable. The AEC chip's built-in acoustic model stores the parameter characteristics of different acoustic environments, and the chip can automatically adjust the filtering parameters according to the characteristics of the current environment, such as the filter order, gain, etc., to more effectively eliminate acoustic echo.

[0084] In this embodiment, in step S1, when the DSP performs multi-band noise suppression, the speech signal is divided into multiple frequency bands, and each frequency band adopts an independent noise estimation and suppression strategy; for the low-frequency band, spectral subtraction combined with Wiener filtering is used for noise suppression; for the high-frequency band, a method based on sub-band spectral estimation is adopted to improve the clarity of the high-frequency speech signal; the DSP monitors the signal-to-noise ratio of each frequency band in real time. When the signal-to-noise ratio of a frequency band is lower than the set threshold, the noise suppression intensity of the frequency band is enhanced, and finally a pure speech signal with a signal-to-noise ratio ≥ 70 dB is output.

[0085] Specifically, the noise sources and characteristics in different frequency bands are different. Dividing the frequency bands and processing them separately can suppress noise more precisely. The noise in the low-frequency band is usually relatively stable. Spectral subtraction estimates the noise spectrum and subtracts it from the noisy speech spectrum for preliminary noise reduction. Wiener filtering further optimizes the filtering effect based on the statistical characteristics of the noise and the signal, retaining the main components of the low-frequency speech signal. The high-frequency speech signal has weak energy and many details. The method based on sub-band spectral estimation can analyze the spectral characteristics of the high-frequency signal in detail, more effectively suppress high-frequency noise, and improve the speech clarity. The DSP monitors the signal-to-noise ratio of each frequency band in real time. When it is lower than the set threshold, the noise suppression intensity is enhanced by increasing the filtering coefficient or using a more complex filtering algorithm to ensure that the signal-to-noise ratio of the output speech signal meets the standard.

[0086] In a possible embodiment, in step S3, the closed-loop control algorithm uses a PID control model:

[0087]

[0088] In the formula, ΔD is the adjustment amount of the PWM duty cycle, ΔT is the difference between the set temperature and the actual temperature, and K p , K i , K d are the proportional, integral, and differential coefficients respectively.

[0089] Specifically, in the PID control model, the proportional coefficient K p adjusts the PWM duty cycle proportionally according to the magnitude of the current temperature error ΔT. The larger the error, the larger the adjustment amount, and the faster the heating power changes. The integral coefficient K i integrates the temperature error to eliminate the steady-state error of the system. For example, if the electric blanket is slightly lower than the set temperature for a long time, the integral term accumulates to increase the PWM duty cycle until the set temperature is reached. The differential coefficient K d adjusts the PWM duty cycle according to the rate of change of the temperature error, can predict the temperature change trend, and respond promptly to rapidly changing temperatures. For example, when the temperature rises too fast, the PWM duty cycle is quickly reduced to prevent the temperature from being too high. The three coefficients work together to achieve precise control of the heating power of the electric blanket, making the temperature of each area stable near the set value.

[0090] In a possible embodiment, the heating power adjustment formula is:

[0091] P output = P max ·D

[0092] In the formula, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for low gear, 30% for medium gear, 60% for high gear, and 90% for ultra-high temperature gear.

[0093] Specifically, the PWM duty cycle D determines the proportion of the thyristor conduction time in a cycle, which also determines the average voltage applied to the heating wire of the electric blanket. The larger the duty cycle, the higher the average voltage across the heating wire. According to the power calculation formula P = U×I (when the resistance of the electric blanket is constant, the higher U, the larger P), the heating power is greater. By setting different PWM duty cycles corresponding to different heating gears, users can select as needed and conveniently adjust the heating power of the electric blanket.

[0094] In a possible embodiment, during the process of adjusting the heating power of the electric blanket, it further includes adopting a dual safety protection measure combining hardware-level overcurrent protection and software over-temperature detection to protect the electric blanket. The hardware-level overcurrent protection and software over-temperature detection cooperate with each other to comprehensively ensure the safe use of the electric blanket. The hardware-level overcurrent protection is based on circuit principles. When the current is abnormal, the circuit is cut off to prevent component damage and accidents. The software over-temperature detection uses the main control MCU to read the data of the temperature sensor in real time. When the threshold is exceeded, the heating circuit is cut off to protect the safety of users and equipment.

[0095] Among them, during the hardware-level overcurrent protection process, a fuse-type fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse-type fuse or an overcurrent situation occurs, the fuse-type fuse melts and cuts off the circuit. The self-resetting fuse restores the circuit connection after the overcurrent situation disappears.

[0096] Specifically, the metal wire in the fuse-type fuse generates a large amount of heat according to Joule's law Q = I 2 Rt when the current is too large. After reaching the melting point, it melts and cuts off the circuit. The self-resetting fuse is a special thermistor. When there is an overcurrent, the resistance increases sharply to limit the current. After the current returns to normal, the temperature decreases, the resistance recovers, and the circuit is re-conducted. It can be used multiple times, improving safety and reliability.

[0097] Among them, during the software over-temperature detection process, the main control MCU reads the data of the temperature sensor in the set heating area of the electric blanket in real time. When it detects that the temperature in the set heating area of the electric blanket exceeds the set safety threshold, it cuts off the power supply of the heating circuit corresponding to the set heating area.

[0098] Specifically, the temperature sensor converts the temperature into an electrical signal, and the main control MCU converts the analog signal into a digital signal through ADC for processing. The main control MCU continuously reads the temperature data and compares it with the preset safety threshold. When the threshold is exceeded, an instruction is issued through the control circuit to cut off the power supply of the corresponding heating circuit and stop heating, avoiding danger caused by excessive temperature and realizing software-level safety protection.

[0099] In a possible embodiment, during the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is:

[0100]

[0101] In the formula, P max is the maximum heating power, and R load is the load resistance.

[0102] Specifically, according to the power calculation formula P = I 2 R is deformed to obtain In the electric blanket, to ensure that the fuse does not blow during normal operation and can cut off the circuit in time during overcurrent, the fuse current threshold needs to be determined according to the maximum heating power and the load resistance. When the current exceeds this threshold, the fuse blows to protect the circuit components.

[0103] Among them, in the software over-temperature detection process, the software over-temperature detection threshold T alert The formula is:

[0104] T alert = T set +ΔT`

[0105] In the formula, T set is the set temperature; ΔT` is the reserved safety margin.

[0106] Specifically, the set temperature is the temperature expected by the user. However, in practice, due to measurement errors of the temperature sensor and the thermal inertia of the electric blanket, the temperature may still fluctuate after reaching the set temperature. The reserved safety margin can avoid misjudging over-temperature caused by a small increase in temperature. When the actual temperature exceeds the set temperature plus the safety margin, it is determined that over-temperature occurs and the heating circuit is cut off, which not only ensures that normal operation does not frequently trigger protection but also can protect in time when there is real overheating.

[0107] In a possible embodiment, after the cloud voice command data passes through the AI semantic parsing engine, the natural language is converted into a control command recognizable by the device. For example, when the user says "Turn up the temperature a little", the semantic parsing engine will recognize the command "Increase the temperature" and convert it into the corresponding temperature adjustment value or gear switching command, and then transmit it back to the main control MCU of the electric blanket through the network for execution. Online mode: The user sends the "Preheat before going to bed" command through the mobile terminal App → Cloud AI parsing → The main control MCU starts zone heating.

[0108] See Figure 2 , an application process of an embodiment of the present invention is as follows:

[0109] First, the instruction input stage

[0110] Voice input: The user issues a voice command, such as "Set the electric blanket to high gear". If in the online state, the voice signal is first collected by the microphone array, preprocessed through filtering by the analog front-end circuit, echo cancellation by the AEC chip, and multi-band noise suppression by the DSP, and then transmitted to the online voice module. The online voice module uses the powerful computing resources and rich voice models in the cloud to extract features from the voice command and match and parse it with the pre-stored command template. If in the offline state, the preprocessed voice signal is processed by the offline voice recognition chip, and the pre-stored voice algorithm instructions in the chip extract and match the signal features to recognize the user's control intention.

[0111] App input: The user opens the mobile App and selects the corresponding temperature gear on the interface, such as clicking the "medium gear" button. At this time, the App establishes a connection with the main control MCU of the electric blanket through the WiFi or Bluetooth communication protocol, encodes and packs the instruction information selected by the user, and then transmits it to the MCU.

[0112] Operation keypad input: The user directly presses the buttons representing different gears on the operation keypad equipped with the electric blanket, such as pressing the button marked "low gear". The electrical signal generated by the button operation is transmitted to the main control MCU through the internal circuit to realize the input of the instruction.

[0113] Second, the instruction processing and power distribution stage

[0114] Instruction parsing: After receiving the instructions from different input methods, the main control MCU decodes and analyzes them. For voice instructions, according to the results parsed by the online or offline voice module, the temperature gear required by the user is determined; for App instructions, the selected gear information is parsed according to the communication protocol; for operation keypad instructions, the gear is identified according to the encoding corresponding to the button.

[0115] PWM duty cycle calculation and power distribution: The main control MCU matches the preset PWM duty cycle according to the parsed gear information, 15% for low gear, 30% for medium gear, 60% for high gear, and 90% for ultra-high temperature gear. For example, if it is recognized that the user selects "high gear", the corresponding duty cycle is 60%. At the same time, considering that the electric blanket may have multiple zones and the limitation of the overall power, the MCU dynamically balances the power of each zone to ensure that each zone can be heated at an appropriate power, meeting the user's demand for the overall temperature and ensuring the relative uniformity of the temperature of each zone.

[0116] Third, the heating execution stage

[0117] Signal drive: The main control MCU outputs a PWM signal with a corresponding duty cycle through its general-purpose input / output (GPIO) interface. This signal is transmitted to the multi-way thyristor voltage regulation circuit, and the high and low level changes of the PWM signal control the on and off times of the thyristor.

[0118] Zone control: Since the electric blanket is divided into multiple heating zones according to the ergonomic zoning model, the multi-way thyristor voltage regulation circuit independently controls the heating circuit of each zone according to the instructions of the main control MCU. For example, when it is detected that the user only wants a higher temperature in the foot area, the conduction time of the thyristor corresponding to the foot area increases, the heating power increases, while other areas maintain the original power or are fine-tuned according to the overall power distribution strategy.

[0119] Fourth, the temperature feedback and regulation stage

[0120] Temperature monitoring: An NTC (negative temperature coefficient) thermistor is installed in each heating zone, which can sense the temperature change in the zone in real time. The resistance value of the NTC thermistor decreases as the temperature increases. By measuring the change in its resistance value and using a specific conversion formula, the resistance value can be converted into the actual temperature value and fed back to the main control MCU.

[0121] Deviation processing: The main control MCU compares the actual temperature values of each zone received with the set temperature value. If the temperature deviation exceeds ±1°C, then according to the positive or negative and magnitude of the deviation, the value of the PWM duty cycle to be adjusted is calculated. For example, when the actual temperature of a certain zone is 1.5°C lower than the set temperature, the PWM duty cycle corresponding to this zone is appropriately increased to increase the heating power; conversely, if the actual temperature is too high, the duty cycle is decreased. At the same time, the latest actual temperature value is updated and displayed on the LED digital tube for the user to intuitively understand.

[0122] Fifth, the safety guarantee stage

[0123] Over-temperature protection: The main control MCU continuously monitors the temperature of each zone. When the temperature of a certain zone exceeds the set value +2°C, first try to reduce the heating level of this zone to reduce the heating power; if the temperature still does not drop or continues to rise, directly cut off the power supply of the heating circuit of this zone. If the temperature reaches 115°C, the hardware fuse (fuse-type fuse) will melt due to overheating, forcibly cutting off the entire circuit at the hardware level to prevent serious accidents such as fires.

[0124] Timeout protection: For the ultra-high temperature gear, the system sets a countdown. When the countdown ends, the heating function of the ultra-high temperature gear is automatically turned off to avoid safety hazards and unnecessary energy consumption caused by long-term high-temperature operation.

[0125] Overcurrent protection: A self - resetting fuse is installed in the circuit to monitor the current magnitude in real - time. When the current is abnormal, such as when a short - circuit causes an instantaneous increase in current, the resistance of the self - resetting fuse will rapidly increase, restricting the current flow and triggering the protection mechanism. Meanwhile, the main control MCU records the fault code E2 to facilitate users or maintenance personnel to troubleshoot the cause of the fault later.

[0126] Sixth, the status feedback stage

[0127] LED digital tube display: The LED digital tube displays the current gear information of the electric blanket and the actual temperature of each zone in real - time. For example, words such as "High - gear, 30°C" are clearly displayed on the digital tube, enabling users to directly obtain the operating status information from the electric blanket itself without the aid of other devices.

[0128] App push notification: The mobile App connected to the electric blanket will, when the status of the electric blanket changes, such as gear switching, temperature reaching the set value, or a fault occurring, promptly push status notifications to the user. Users can understand the operating conditions of the electric blanket at any time on their mobile phones and keep track of its dynamics even when not beside the electric blanket.

[0129] Offline voice broadcast: In the offline mode without a network connection, when important changes occur in the status of the electric blanket, such as power - on, power - off, or a fault, the built - in voice broadcast module will broadcast the current status to the user in voice form, such as "The electric blanket has been powered on, currently at medium - gear", providing a convenient way for users to obtain information.

[0130] Embodiment 2

[0131] See Figure 3 , Embodiment 2 of the present invention also provides an electric blanket control system based on AI voice interaction, including:

[0132] A voice signal pre - processing module 001, which is used to collect voice signals containing environmental noise using a microphone array, filter the voice signals containing environmental noise through an analog front - end circuit to remove interference signals; then, an AEC chip eliminates the acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi - band noise suppression to output a pure voice signal with a signal - to - noise ratio ≥ 70dB;

[0133] An instruction parsing and control module 002, which is used to extract features from the pre - processed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre - stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized after the match, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi - way thyristor voltage - regulating circuit according to the received digital control signal.

[0134] The partition temperature control strategy adjustment module 003 is used to divide the electric blanket into multiple heating zones according to the ergonomic partition model to adapt to the different temperature requirements of different parts of the human body; in the process of adjusting the heating power of the electric blanket, a closed-loop control algorithm is adopted to collect the temperature data of each heating zone of the electric blanket in real time, and the heating power of each heating zone of the electric blanket is dynamically adjusted according to the difference between the set temperature and the actual temperature.

[0135] In this embodiment, in the voice signal preprocessing module 001: the microphone array is a linear array or a circular array composed of at least three microphones, and the distance between each microphone is optimized according to the frequency range of the voice signal and the expected noise reduction effect;

[0136] The low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal;

[0137] The anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal;

[0138] When the AEC chip eliminates acoustic echo based on the adaptive filter algorithm, it dynamically adjusts the step size according to the magnitude of the error signal, and adaptively adjusts the filter parameters according to different acoustic environments through the acoustic model built in the AEC chip. The formula for eliminating acoustic echo based on the adaptive filter algorithm is:

[0139]

[0140] In the formula, w k is the filter coefficient vector, μ is the step size factor, e k is the error signal, x k is the reference input signal.

[0141] In this embodiment, when the DSP performs multi-band noise suppression, the voice signal is divided into multiple frequency bands, and each frequency band adopts an independent noise estimation and suppression strategy; for the low-frequency band, spectral subtraction combined with Wiener filtering is used for noise suppression; for the high-frequency band, a method based on sub-band spectral estimation is used to improve the clarity of the high-frequency voice signal; the DSP monitors the signal-to-noise ratio of each frequency band in real time. When the signal-to-noise ratio of a frequency band is lower than the set threshold, the noise suppression intensity of the frequency band is enhanced, and finally a pure voice signal with a signal-to-noise ratio ≥ 70 dB is output.

[0142] As a preferred solution of the electric blanket control system based on AI voice interaction, in the partition temperature control strategy adjustment module 003, the closed-loop control algorithm adopts a PID control model:

[0143]

[0144] Wherein, ΔD is the PWM duty cycle adjustment amount, ΔT is the difference between the set temperature and the actual temperature, and K p , K i , K d are the proportional, integral, and differential coefficients respectively;

[0145] In the partition temperature control strategy adjustment module 003, the heating power adjustment formula is:

[0146] P output = P max ·D

[0147] Wherein, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for low gear, 30% for medium gear, 60% for high gear, and 90% for ultra-high temperature gear.

[0148] In this embodiment, it further includes:

[0149] A safety protection module 004, which is used to protect the electric blanket by adopting a dual safety protection measure combining hardware-level overcurrent protection and software over-temperature detection;

[0150] During the hardware-level overcurrent protection process, a fuse-type fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse-type fuse or an overcurrent situation occurs, the fuse-type fuse melts to cut off the circuit, and the self-resetting fuse restores the circuit connection after the overcurrent situation disappears;

[0151] During the software over-temperature detection process, the main control MCU reads the data of the temperature sensor in the set heating area of the electric blanket in real time. When it detects that the temperature in the set heating area of the electric blanket exceeds the set safety threshold, it cuts off the power supply of the heating circuit corresponding to the set heating area;

[0152] During the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is:

[0153]

[0154] Wherein, P max is the maximum heating power, R load is the load resistance;

[0155] During the software over-temperature detection process, the software over-temperature detection threshold T alert The formula is:

[0156] T alert = T set +ΔT`

[0157] Wherein, T set is the set temperature; ΔT` is the reserved safety margin.

[0158] It should be noted that for the information interaction, execution process, etc. among the above-mentioned system modules, since they are based on the same concept as the method embodiment in Embodiment 1 of this application, the technical effects brought by them are the same as those of the method embodiment of this application. For specific content, reference can be made to the description in the method embodiment shown above in this application, and details will not be elaborated here.

[0159] Embodiment 3

[0160] Embodiment 3 of the present invention provides a non-transitory computer-readable storage medium, in which program codes of an electric blanket control method based on AI voice interaction are stored, and the program codes include instructions for executing the electric blanket control method based on AI voice interaction in Embodiment 1 or any possible implementation manner thereof.

[0161] The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center integrating one or more available media. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid-state disk (SSD)), etc.

[0162] Embodiment 4

[0163] Embodiment 4 of the present invention provides an electronic device, including: a memory and a processor;

[0164] The processor and the memory complete communication with each other through a bus; the memory stores program instructions executable by the processor, and the processor can execute the electric blanket control method based on AI voice interaction in Embodiment 1 or any possible implementation manner thereof by calling the program instructions.

[0165] Specifically, the processor can be implemented by hardware or by software. When implemented by hardware, the processor can be a logic circuit, an integrated circuit, etc.; when implemented by software, the processor can be a general-purpose processor, which is implemented by reading software codes stored in the memory. The memory can be integrated in the processor or can exist independently outside the processor.

[0166] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.).

[0167] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by the computing device. Thus, they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order than here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module to implement. In this way, the present invention is not limited to any specific combination of hardware and software.

[0168] Although the present invention has been described in detail above with general descriptions and specific embodiments, on the basis of the present invention, some modifications or improvements can be made, which are obvious to those skilled in the art. Therefore, these modifications or improvements made without departing from the spirit of the present invention all fall within the scope of protection required by the present invention.

Claims

1. An electric blanket control method based on AI voice interaction, characterized in that It includes the following steps: Collect voice signals containing ambient noise using a microphone array, filter the voice signals containing ambient noise through an analog front-end circuit to remove interference signals; then, an AEC chip eliminates acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi-band noise suppression to output a pure voice signal with a signal-to-noise ratio ≥ 70dB; Extract features from the preprocessed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre-stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized after the match, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi-way thyristor voltage regulating circuit according to the received digital control signal; According to the ergonomic zoning model, divide the electric blanket into multiple heating zones to adapt to different temperature requirements of different parts of the human body; during the process of adjusting the heating power of the electric blanket, use a closed-loop control algorithm to collect the temperature data of each heating zone of the electric blanket in real time, and dynamically adjust the heating power of each heating zone of the electric blanket according to the difference between the set temperature and the actual temperature.

2. The electric blanket control method based on AI voice interaction according to claim 1, wherein The microphone array is a linear array or a circular array composed of at least three microphones, and the spacing between each microphone is optimized according to the frequency range of the voice signal and the expected noise reduction effect; The low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal; The anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal; When the AEC chip eliminates acoustic echo based on the adaptive filter algorithm, dynamically adjusts the step size according to the magnitude of the error signal, and adaptively adjusts the filtering parameters according to different acoustic environments through the acoustic model built in the AEC chip. The formula for eliminating acoustic echo based on the adaptive filter algorithm is: where w k is the filter coefficient vector, μ is the step factor, and e k is the error signal, and x k is the reference input signal.

3. The electric blanket control method based on AI voice interaction according to claim 2, wherein When the DSP performs multi-band noise suppression, divides the voice signal into multiple frequency bands, and each frequency band adopts an independent noise estimation and suppression strategy; for the low-frequency band, uses spectral subtraction combined with Wiener filtering for noise suppression; for the high-frequency band, uses a method based on sub-band spectral estimation to improve the clarity of high-frequency voice signals; the DSP monitors the signal-to-noise ratio of each frequency band in real time, and when the signal-to-noise ratio of a frequency band is lower than the set threshold, enhances the noise suppression intensity of the frequency band, and finally outputs a pure voice signal with a signal-to-noise ratio ≥ 70dB.

4. The electric blanket control method based on AI voice interaction according to claim 1, characterized in that, The closed-loop control algorithm uses a PID control model: Where ΔD is the PWM duty cycle adjustment amount, ΔT is the difference between the set temperature and the actual temperature, and K p , K i , K d are the proportional, integral, and derivative coefficients respectively; The heating power adjustment formula is: P output = P max · D Wherein, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for the low gear, 30% for the medium gear, 60% for the high gear, and 90% for the ultra-high temperature gear.

5. The electric blanket control method based on AI voice interaction according to claim 1, wherein During the process of adjusting the heating power of the electric blanket, it also includes using a dual safety protection measure combining hardware-level overcurrent protection and software over-temperature detection to protect the electric blanket; Hardware-level overcurrent protection process, where a fuse-type fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse-type fuse or an overcurrent situation occurs, the fuse-type fuse melts to cut off the circuit, and the self-resetting fuse restores the circuit connection after the overcurrent situation disappears; Software over-temperature detection process, where the main control MCU reads the temperature sensor data of the set heating area of the electric blanket in real time. When it detects that the temperature of the set heating area of the electric blanket exceeds the set safety threshold, it cuts off the power supply of the heating circuit corresponding to the set heating area.

6. The electric blanket control method based on AI voice interaction according to claim 5, characterized in that, During the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is: where P max is the maximum heating power, and R load is the load resistance; Software over-temperature detection process, software over-temperature detection threshold T alert The formula is: T alert = T set + ΔT` where T set is the set temperature; ΔT` is the reserved safety margin.

7. The electric blanket control system based on AI voice interaction is characterized in that, Including: A voice signal preprocessing module, which is used to collect voice signals containing ambient noise by using a microphone array, filter the voice signals containing ambient noise through an analog front-end circuit to remove interference signals; then, an AEC chip eliminates the acoustic echo generated by speaker feedback based on an adaptive filter algorithm, and a DSP performs multi-band noise suppression to output a pure voice signal with a signal-to-noise ratio ≥ 70dB; An instruction parsing and control module, which is used to extract features from the preprocessed pure voice signal through an online AI chip or an offline voice recognition chip, match the extracted features with a pre-stored instruction template, and when the corresponding control instruction of the pure voice signal is successfully recognized after the match, generate a corresponding digital control signal and transmit it to the main control MCU. The main control MCU adjusts the heating power of the electric blanket through a PWM signal to drive a multi-way thyristor voltage regulation circuit according to the received digital control signal; A partition temperature control strategy adjustment module, which is used to divide the electric blanket into multiple heating areas according to the ergonomic partition model to adapt to the different temperature requirements of different parts of the human body; during the process of adjusting the heating power of the electric blanket, a closed-loop control algorithm is adopted to collect the temperature data of each heating area of the electric blanket in real time, and dynamically adjust the heating power of each heating area of the electric blanket according to the difference between the set temperature and the actual temperature.

8. The electric blanket control system based on AI voice interaction according to claim 7, characterized in that, In the voice signal preprocessing module: the microphone array is a linear array or a circular array composed of at least three microphones, and the spacing between each microphone is optimized according to the frequency range of the voice signal and the expected noise reduction effect; The low-noise operational amplifier in the analog front-end circuit uses a model with a high gain-bandwidth product and a low input noise density, and the gain of the low-noise operational amplifier in the analog front-end circuit is set to be automatically adjusted according to the intensity of the input voice signal; The anti-aliasing filter in the analog front-end circuit is a second-order or higher-order Butterworth filter, and the cut-off frequency is set according to the highest frequency of the voice signal; When the AEC chip eliminates the acoustic echo based on the adaptive filter algorithm, it dynamically adjusts the step size according to the magnitude of the error signal, and adaptively adjusts the filtering parameters according to different acoustic environments through the acoustic model built in the AEC chip. The formula for eliminating the acoustic echo based on the adaptive filter algorithm is: where \(w\) k is the filter coefficient vector, \(\mu\) is the step factor, and \(e\) k is the error signal, and \(x\) k is the reference input signal; When the DSP performs multi-band noise suppression, the speech signal is divided into multiple frequency bands, and each frequency band adopts an independent noise estimation and suppression strategy; for the low-frequency band, spectral subtraction combined with Wiener filtering is used for noise suppression; for the high-frequency band, a method based on sub-band spectral estimation is adopted to improve the clarity of high-frequency speech signals; the DSP monitors the signal-to-noise ratio of each frequency band in real time. When the signal-to-noise ratio of a frequency band is lower than the set threshold, the noise suppression intensity of the frequency band is enhanced, and finally a pure speech signal with a signal-to-noise ratio ≥ 70 dB is output.

9. The electric blanket control system based on AI voice interaction according to claim 7, characterized in that, In the partition temperature control strategy adjustment module, the closed-loop control algorithm adopts a PID control model: Where ΔD is the PWM duty cycle adjustment amount, ΔT is the difference between the set temperature and the actual temperature, and K p 、K i 、K d are the proportional, integral, and derivative coefficients respectively; In the partition temperature control strategy adjustment module, the heating power adjustment formula is: P output = P max ·D Wherein, P output is the output power, P max is the maximum power, D is the PWM duty cycle, 15% for the low gear, 30% for the medium gear, 60% for the high gear, and 90% for the ultra-high temperature gear.

10. The electric blanket control system based on AI voice interaction according to claim 7, characterized in that, It also includes: A safety protection module for protecting the electric blanket by adopting a dual safety protection measure combining hardware-level overcurrent protection and software over-temperature detection; In the process of hardware-level overcurrent protection, a fuse-type fuse and a self-resetting fuse are set. When the current in the circuit exceeds the rated current of the fuse-type fuse or an overcurrent situation occurs, the fuse-type fuse melts to cut off the circuit, and the self-resetting fuse restores the circuit connection after the overcurrent situation disappears; In the process of software over-temperature detection, the main control MCU reads the data of the temperature sensor in the set heating area of the electric blanket in real time. When it is detected that the temperature in the set heating area of the electric blanket exceeds the set safety threshold, the power supply of the heating circuit corresponding to the set heating area is cut off; During the hardware-level overcurrent protection process, the hardware fuse current threshold I fuse The calculation formula is as follows: Wherein, P max is the maximum heating power, and R load is the load resistance; Software over-temperature detection process, software over-temperature detection threshold T alert The formula is: T alert = T set + ΔT` where T set is the set temperature; ΔT` is the reserved safety margin.

Images