Socket intelligent control system based on voice instruction recognition

Abstract

The present application relates to the technical field of intelligent device control, especially to a socket intelligent control system based on voice instruction recognition, which is provided with an instruction generation module, an authority analysis module, a preprocessing module and a voice instruction matching module, acquires a voiceprint signal, a socket state and a pre-set instruction label through the instruction generation module, marks a reference socket through the authority analysis module, determines the authority category of a user, determines the socket control preprocessing mode as marking a key socket or determines whether there is a recognition deviation risk through the preprocessing module, performs instruction compensation identification on the voiceprint signal of each socket, determines the voice instruction matching mode as cutting off a key socket or performs instruction re-verification through the voice instruction matching module, and thus, accurate authority division according to actual environmental conditions, adaptive adjustment of the processing mode, further matching verification in the interference environment and improved accuracy of socket intelligent control are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent device control, in particular to a socket intelligent control system based on voice instruction recognition. BACKGROUND

[0002] With the continuous progress of Internet of Things technology, network communication technology and the like, people have higher requirements for the convenience, comfort and intelligent degree of home life, and hope to be able to control various electrical appliances in the home in a more convenient way. As a very commonly used electrical device in the home, the socket provides a power supply interface for various household appliances, and its intelligentization is very important. The application of voice instruction recognition technology to the socket intelligent control system can effectively solve the problem of complicated operation of traditional intelligent sockets. Users can easily control the on-off of the socket through voice instructions. However, in actual application, the degree of environmental interference in different application scenarios is different, and the degree of influence of environmental interference on voice recognition is also different. At the same time, children in the imitating stage are in a rapid development period of language learning ability, and they are full of curiosity and keen on imitation of the behavior of adults around them. Since they lack sufficient cognitive and judgment ability for danger, once they imitate adults to issue voice instructions to control the socket, and the socket intelligent control system fails to effectively identify, it is very likely to cause the socket to execute the instructions incorrectly. Therefore, it is a technical problem to be solved to accurately divide the permissions according to the actual environmental conditions, adaptively adjust the processing mode, and improve the accuracy of socket intelligent control.

[0003] For example, Chinese patent No. CN113314113B discloses an intelligent socket control method, device, equipment and storage medium. The method comprises: establishing a communication connection between a terminal device and an audio receiver; collecting an audio signal through the terminal device or playing an audio signal through the terminal device; receiving the audio signal as a voice input signal through the audio receiver; performing feature recognition on the voice input information through the processor of the intelligent socket to extract the features of the audio signal; matching the control command of the socket according to the features of the audio signal, and controlling the operation action of the socket according to the matched control command; or performing semantic recognition on the features of the audio signal, and controlling the operation action of the socket according to the result of the semantic recognition.

[0004] The existing technology still has the following problems:

[0005] The existing technology does not consider that collecting voice instructions in a closed space may have reverberation interference, which affects the collection of voice instructions. The existing technology cannot accurately divide the permissions according to the actual environmental conditions, adaptively adjust the processing mode, and cannot further match and verify in the interference environment, which affects the accuracy of socket intelligent control. SUMMARY

[0006] To this end, the application provides a socket intelligent control system based on voice instruction recognition to overcome the problem that the prior art cannot accurately divide the authority according to the actual environment, adaptively adjust the processing mode, and further match and verify the interference environment, thereby affecting the accuracy of socket intelligent control.

[0007] To achieve the above-mentioned purpose, the application provides a socket intelligent control system based on voice instruction recognition, comprising:

[0008] An instruction generation module, which comprises a voiceprint acquisition unit for acquiring user voice and converting it into a voiceprint signal, a monitoring unit for acquiring the state of the socket, and a setting unit for pre-setting the instruction label of each socket;

[0009] An authority analysis module connected to the instruction generation module, for determining a voiceprint factor and a characteristic tendency coefficient based on the voiceprint signal, marking a reference socket based on the voiceprint factor, and obtaining the characteristic tendency coefficient of the reference socket to determine the authority category of the user;

[0010] A preprocessing module connected to the instruction generation module and the authority analysis module, respectively, for determining the socket control preprocessing mode according to the authority category, marking a key socket according to the comparison of a plurality of characteristic tendency coefficients, or determining whether there is a risk of recognition deviation according to the comparison of the reference socket and a plurality of characteristic sockets, and performing instruction compensation recognition on the voiceprint signal of the socket with the risk of recognition deviation;

[0011] The characteristic socket is determined according to the voiceprint signal of the reference socket.

[0012] A voice instruction matching module connected to the instruction generation module and the preprocessing module, respectively, for determining the voice instruction matching mode according to the socket control preprocessing mode, cutting off the key socket, or obtaining the socket state of the socket matched by the instruction compensation recognition for instruction re-verification, and controlling the socket according to the verification result.

[0013] Further, the authority analysis module is used to mark the reference socket and determine the characteristic tendency coefficient, wherein,

[0014] The authority analysis module obtains a plurality of amplitudes on the voiceprint signal, determines the average value of the amplitudes as the voiceprint factor of the voiceprint signal, and determines the socket corresponding to the voiceprint signal with the maximum value of the voiceprint factor as the reference socket.

[0015] The authority analysis module calculates the peak interval variance of the voiceprint signal according to a plurality of peak interval distances of the voiceprint signal, and determines the peak interval variance as the characteristic tendency coefficient.

[0016] Further, the permission analysis module is configured to determine the permission category of the user, wherein

[0017] If the feature tendency coefficient meets the first permission determination condition, the permission analysis module determines that the permission category of the user is a first permission category.

[0018] If the feature tendency coefficient does not meet the first permission determination condition, the permission analysis module determines that the permission category of the user is a second permission category.

[0019] The first permission determination condition is that the feature tendency coefficient obtained in a preset sampling period exceeds a preset voice feature tendency difference reference value.

[0020] Further, the preprocessing module is configured to determine a socket control preprocessing mode, wherein

[0021] If the permission category is the first permission category, the preprocessing module determines that the socket control preprocessing mode is to mark a key socket according to a comparison situation of a plurality of feature tendency coefficients.

[0022] If the permission category is the second permission category, the preprocessing module determines that the socket control preprocessing mode is to determine whether there is a recognition deviation risk according to a comparison situation of the flag socket and a plurality of feature sockets, and to perform instruction compensation recognition on the voiceprint signal of the socket with the recognition deviation risk.

[0023] Further, the preprocessing module is configured to mark the key socket, wherein

[0024] If the feature tendency coefficient of the socket meets a marking condition, the preprocessing module marks the socket as a key socket.

[0025] The marking condition is that a difference between the feature tendency coefficient in a preset first monitoring time period and a preset second monitoring time period exceeds a preset feature tendency difference reference value.

[0026] Further, the preprocessing module is configured to determine the flag socket and the feature socket, wherein

[0027] The preprocessing module sorts initial time points of the voiceprint signals collected by each socket in time sequence, and determines the socket corresponding to the initial time point ranked first as the flag socket.

[0028] The preprocessing module calculates the similarity between the voiceprint signal of the reference socket and the voiceprint signals of the remaining sockets, respectively, and determines the socket with a similarity greater than a preset similarity threshold as a feature socket.

[0029] Further, the preprocessing module is configured to determine whether there is a recognition deviation risk, wherein

[0030] If the mark socket is not any one of the feature sockets, the preprocessing module determines that there is a risk of identification deviation;

[0031] If the mark socket is any one of the feature sockets, the preprocessing module determines that there is no risk of identification deviation.

[0032] Further, the preprocessing module is also used to perform instruction compensation identification on the voiceprint signals of each socket, wherein,

[0033] The preprocessing module determines the amplitudes according to adjacent peaks and valleys on the voiceprint signals, determines a voiceprint signal region with amplitudes meeting an energy concentration condition as an energy concentration region, and identifies a user instruction according to the energy concentration region.

[0034] The energy concentration condition is that the difference between the amplitudes does not exceed a preset difference threshold value, and the amplitudes exceed a preset amplitude threshold value.

[0035] Further, the voice instruction matching module is used to determine a voice instruction matching mode, wherein,

[0036] If the socket control preprocessing manner is to mark a key socket, the voice instruction matching module determines that the voice instruction matching mode is to cut off the key socket.

[0037] If the socket control preprocessing manner is to perform instruction compensation identification on the voiceprint signals of each socket, the voice instruction matching module determines that the voice instruction matching mode is to obtain the socket state of the socket matched by the instruction compensation identification to perform instruction re-verification, and performs socket control according to the verification result.

[0038] Further, the voice instruction matching module is used to perform instruction re-verification, and performs socket control according to the verification result, wherein,

[0039] The voice instruction matching module performs pre-matching according to the instruction compensation identification, and determines the socket state of the socket meeting the pre-matching result.

[0040] If the socket state meets a verification condition, the voice instruction matching module determines that the pre-matching result is correct, and performs socket control according to the pre-matching result.

[0041] The verification condition is that the pre-matched socket accesses an electrical equipment.

[0042] Compared with the prior art, the present application has the beneficial effects that the present application is provided with an instruction generation module, an authority analysis module, a preprocessing module and a voice instruction matching module, the voiceprint signal, the socket state and the pre-set instruction label are obtained through the instruction generation module, the reference socket is marked through the authority analysis module to determine the authority category of the user, the socket control preprocessing mode is determined as marking the key socket through the preprocessing module, or it is determined whether there is a recognition deviation risk, the voiceprint signal of each socket is subjected to instruction compensation recognition, the voice instruction matching mode is determined as cutting off the key socket through the voice instruction matching module, or instruction re-verification is performed, and the socket control is performed according to the verification result, thereby realizing accurate authority division according to the actual environmental conditions, adaptively adjusting the processing mode, further matching and verifying in the interference environment, and improving the accuracy of socket intelligent control.

[0043] Especially, the characteristic tendency coefficient of the reference socket is obtained through the authority analysis module to determine the authority category of the user, that is, the voiceprint factor is obtained by using the amplitudes on the voiceprint signal, the socket corresponding to the maximum value of the voiceprint factor is taken as the reference socket, and the characteristic tendency coefficient is obtained on the basis to determine the authority category, the clearest voiceprint of the user can be obtained for characteristic analysis, the authorities of different users can be accurately distinguished, the authority misjudgment is avoided, the accuracy of system authority management is improved, the voiceprint signal collected based on the reference socket is deeply analyzed, the user does not need to input complex identity verification information, and accurate authority determination can be automatically performed, and intelligent authority management is realized, the characteristic tendency coefficient of the reference socket is obtained through the authority analysis module to determine the authority category of the user, thereby realizing accurate authority division according to the actual environmental conditions, and improving the accuracy of socket intelligent control.

[0044] Especially, the key socket is marked according to the comparison of the characteristic tendency coefficients of the first authority category, that is, the key socket is marked when the user is identified as a child, the socket that may be touched by the child can be effectively identified, especially the socket at the level of the child, the risk of accidental electric shock of the child is reduced, the behavior characteristics and the activity range of the child are represented according to the comparison of the characteristic tendency coefficients, the socket that needs to be focused on is screened out, and more targeted protection is provided, the key socket is marked according to the comparison of the characteristic tendency coefficients of the first authority category, thereby realizing accurate authority division according to the actual environmental conditions, adaptively adjusting the processing mode, and improving the accuracy of socket intelligent control.

[0045] Especially, the application determines whether there is a recognition deviation risk according to the comparison of the landmark socket and the plurality of feature sockets in the second permission category, and compensates and recognizes the voiceprint signal of the socket with the recognition deviation risk, it can be understood that the second permission category is that the user is an adult, in the actual environment, reverberation can easily cause distortion of the voiceprint signal, and further affect the accuracy of recognition, by analyzing the comparison of the landmark socket and the plurality of feature sockets, whether there is reverberation interference can be effectively detected, once it is determined that there is reverberation, the voiceprint signal of each socket is compensated and recognized in time, the signal deviation caused by reverberation can be corrected, the accuracy of adult voice command recognition is significantly improved, the accuracy of socket control is ensured, adults have high requirements for the accuracy and stability of control when using the socket intelligent control system, through this detailed analysis and compensation mechanism specially for the adult permission category, the individualized needs of adult users can be better met, there is no need to worry about the recognition problem caused by environmental factors, the satisfaction of user experience is improved, and thus, accurate permission division is realized according to the actual environmental conditions, the processing mode is adaptively adjusted, and the accuracy of socket intelligent control is improved.

[0046] Especially, the application obtains the socket state of the socket matched in the instruction compensation recognition after compensating and recognizing the voiceprint signal of each socket, and performs instruction verification again, and controls the socket according to the verification result, it can be understood that people will first insert the plug of the electrical appliance into the socket, and then control the socket to start the electrical appliance, so that it works normally, which is a way conforming to the daily operation habit and the electrical appliance use process, if the socket is started first and then the electrical appliance plug is inserted, an electric spark may be generated, there is a certain safety hazard, and it also does not conform to the operation habit of most people who connect the device first and then power on, reverberation can cause distortion, delay and superposition of the voiceprint signal, and increase the difficulty of accurate recognition of the voiceprint signal and judgment of the instruction of the system, although the processing mode is adjusted for instruction compensation recognition, there may be misjudgment, and a safety hazard may be generated, the socket state is verified again, which can provide additional judgment basis for the socket control system, help to correct the wrong judgment caused by reverberation interference, and thus, accurate permission division is realized according to the actual environmental conditions, the processing mode is adaptively adjusted, further matching verification is performed on the interference environment, and the accuracy of socket intelligent control is improved. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 It is a function block diagram of the socket intelligent control system based on voice command recognition of the embodiment of the application;

[0048] Figure 2 It is a logic flow chart for determining the permission category of the user by the permission analysis module of the embodiment of the application;

[0049] Figure 3The logic flow chart for determining the pre-processing mode of the socket control pre-processing module for the embodiment of the present application is shown in the figure;

[0050] Figure 4 The logic flow chart for marking the key socket for the pre-processing module of the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0051] In order to make the objects and advantages of the present application more clear, the present application is further described below in conjunction with the embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0052] The preferred embodiments of the present application 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 application and are not intended to limit the protection scope of the present application.

[0053] It should be noted that in the description of the present application, the terms "upper", "lower", "inner", "outer" and the like indicating the direction or positional relationship are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description and does not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0054] In addition, it should also be noted that in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through an intermediate medium, or internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0055] Please refer to Figure 1 The figure shows the functional block diagram of the socket intelligent control system based on voice instruction recognition of the embodiment of the present application, the socket intelligent control system based on voice instruction recognition of the present application comprises:

[0056] The instruction generation module comprises a voiceprint acquisition unit for acquiring user voice and converting it into a voiceprint signal, a monitoring unit for acquiring the state of the socket, and a setting unit for pre-setting the instruction label of each socket;

[0057] Specifically, the present application does not limit the specific structure of the voiceprint acquisition unit, preferably, it can be an integrated voice sensor, which performs real-time processing and analysis on the collected sound signal and extracts the voiceprint feature, which will not be described here.

[0058] Specifically, the specific structure of the monitoring unit is not limited in the present application, preferably, it can be a capacitive sensor, an induction electrode is arranged near the socket, when an electrical appliance plug is inserted into the socket, the plug and the induction electrode will form a capacitive coupling, resulting in a change in the capacitance value of the induction electrode, by detecting the change in the capacitance value, it is determined whether the socket constitutes a closed loop, which will not be repeated here.

[0059] Specifically, the specific structure of the setting unit is not limited in the present application, it can be a smart terminal device with a graphical user interface, preferably, it can be a smart phone application or a smart control panel, used to pre-set each socket instruction label, which will not be repeated here.

[0060] The permission analysis module is connected with the instruction generation module, and is used to determine a voiceprint factor and a characteristic tendency coefficient according to the voiceprint signal, mark a reference socket based on the voiceprint factor, and determine the permission category of the user by acquiring the characteristic tendency coefficient of the reference socket;

[0061] Specifically, the specific structure of the permission analysis module is not limited in the present application, preferably, it can be a processor used by a computer, used to determine the voiceprint factor and the characteristic tendency coefficient, mark the reference socket, and determine the permission category of the user, which will not be repeated here.

[0062] The preprocessing module is connected with the instruction generation module and the permission analysis module respectively, and is used to determine a socket control preprocessing mode according to the permission category, mark a key socket according to the comparison of a plurality of characteristic tendency coefficients, or determine whether there is a recognition deviation risk according to the comparison of the reference socket and a plurality of characteristic sockets, and perform instruction compensation identification on the voiceprint signal of the socket with the recognition deviation risk;

[0063] The characteristic socket is determined according to the voiceprint signal of the reference socket.

[0064] Specifically, the specific structure of the preprocessing module is not limited in the present application, preferably, it can be a microprocessor, used to determine the socket control preprocessing mode, mark the key socket, determine whether there is a recognition deviation risk, and perform instruction compensation identification on the voiceprint signal of each socket, which will not be repeated here.

[0065] The voice instruction matching module is connected with the instruction generation module and the preprocessing module respectively, and is used to determine a voice instruction matching mode according to the socket control preprocessing mode, cut off the key socket, or acquire the socket state of the socket matched by the instruction compensation identification to perform instruction re-verification, and perform socket control according to the verification result.

[0066] Specifically, the present application does not limit the specific structure of the voice instruction matching module, and preferably, it can be an intelligent terminal device used to determine the voice instruction matching mode and control the opening and closing of the socket, which will not be described here.

[0067] Specifically, the permission analysis module is used to mark the reference socket and determine the characteristic tendency coefficient, wherein,

[0068] The permission analysis module obtains a plurality of amplitudes on the voiceprint signal, determines the average amplitude as the voiceprint factor of the voiceprint signal, and determines the socket corresponding to the voiceprint signal with the maximum voiceprint factor as the reference socket.

[0069] The permission analysis module calculates the wave peak interval variance according to a plurality of wave peak interval spacings of the voiceprint signal, and determines the wave peak interval variance as the characteristic tendency coefficient.

[0070] Please refer to Figure 2 The permission analysis module is used to determine the permission category of the user, wherein,

[0071] If the characteristic tendency coefficient meets the first permission determination condition, the permission analysis module determines the permission category of the user as the first permission category.

[0072] If the characteristic tendency coefficient does not meet the first permission determination condition, the permission analysis module determines the permission category of the user as the second permission category.

[0073] The first permission determination condition is that the characteristic tendency coefficient obtained within a preset sampling period exceeds a preset voice feature tendency difference reference value.

[0074] Specifically, the preset sampling period and the preset voice feature tendency difference reference value can be set by a person skilled in the art according to the accuracy requirement of the socket intelligent control. The higher the accuracy requirement, the longer the preset sampling period, and the smaller the preset voice feature tendency difference reference value. The value range of the preset sampling period can be [1, 5] with the interval unit being milliseconds, and the value range of the preset voice feature tendency difference reference value can be [0.2, 0.8]. Preferably, the preset sampling period can be 3 milliseconds, and the preset voice feature tendency difference reference value can be 0.6.

[0075] Specifically, the application determines the user's permission category by obtaining the characteristic tendency coefficient of the reference socket through the permission analysis module. It can be understood that the voiceprint factor is obtained by using several amplitudes on the voiceprint signal, and the socket corresponding to the maximum value of the voiceprint factor is taken as the reference socket. On this basis, the characteristic tendency coefficient is obtained to determine the permission category. The clearest voiceprint of the user can be obtained for feature analysis, and the permissions of different users can be accurately distinguished to avoid permission misjudgment and improve the accuracy of system permission management. Based on the in-depth analysis of the voiceprint signal collected based on the reference socket, the user does not need to input complex identity verification information, and the accurate permission determination can be automatically performed to realize intelligent permission management. The application determines the user's permission category by obtaining the characteristic tendency coefficient of the reference socket through the permission analysis module. Furthermore, accurate permission division is realized according to the actual environmental conditions to improve the accuracy of socket intelligent control.

[0076] Specifically, it can be understood that in a closed environment, sound propagation will be reflected by walls, furniture and other objects, thereby producing a reverberation phenomenon. The reverberation will distort the voiceprint signal, resulting in that the collected voiceprint signal contains more reflected echoes and interference components, which will interfere with the feature extraction and analysis of the voiceprint signal. According to several amplitudes on the voiceprint signal, the average value of the amplitudes is taken as a voiceprint factor to represent the intensity of the voiceprint signal. The larger the voiceprint factor, the stronger the voice energy collected by the socket, and the greater the interference of the reflected sound. The voiceprint signal collected at the position with the largest voiceprint factor is relatively less affected by reverberation, and can more truly reflect the voice features of the user. The characteristic tendency coefficient is determined based on the peak distance of the voiceprint signal, which reflects the stability and feature consistency of the voiceprint signal. The larger the characteristic tendency coefficient, the worse the stability. The voiceprint signal emitted by a child has poor stability. According to the characteristic tendency coefficient, the permission category is divided to ensure the reliability of the system. Furthermore, accurate permission division is realized according to the actual environmental conditions to improve the accuracy of socket intelligent control.

[0077] Please refer to Figure 3 The preprocessing module is used to determine the socket control preprocessing mode, and the preprocessing module comprises a first determining unit, a second determining unit and a third determining unit.

[0078] If the permission category is the first permission category, the preprocessing module determines that the socket control preprocessing mode is to mark the key socket according to the comparison of the several characteristic tendency coefficients.

[0079] Specifically, the application marks the key socket according to the comparison of the first permission category and the several characteristic tendency coefficients, which can be understood as that the first permission category identifies that the user is a child, and the key socket is marked, which can effectively identify the socket that the child may touch, especially the socket at the level of the child, reducing the risk of accidental electric shock of the child, and screening the socket that needs to be focused on according to the behavior characteristics and the activity range of the child, providing more targeted protection, the application marks the key socket according to the comparison of the first permission category and the several characteristic tendency coefficients, and then, the accurate permission division is realized according to the actual environment, the processing mode is adaptively adjusted, and the accuracy of the socket intelligent control is improved.

[0080] If the permission category is the second permission category, the preprocessing module determines that the socket control preprocessing mode is to determine whether there is an identification deviation risk according to the comparison of the landmark socket and the several characteristic sockets, and to perform instruction compensation identification on the voiceprint signal of the socket with the identification deviation risk.

[0081] Specifically, the application determines whether there is an identification deviation risk according to the comparison of the landmark socket and the several characteristic sockets for the second permission category, and performs instruction compensation identification on the voiceprint signal of the socket with the identification deviation risk, which can be understood as that the second permission category identifies that the user is an adult, and in the actual environment, reverberation can easily cause distortion of the voiceprint signal, and then affect the accuracy of identification, by analyzing the comparison of the landmark socket and the several characteristic sockets, whether there is reverberation interference can be effectively detected, once it is determined that there is reverberation, the voiceprint signal of each socket is compensated and identified in time, which can correct the signal deviation caused by reverberation, significantly improve the accuracy of adult voice instruction identification, ensure the accuracy of socket control, and the adult has a higher requirement for the accuracy and stability of the control when using the socket intelligent control system, through this detailed analysis and compensation mechanism specially for the adult permission category, the individualized needs of the adult user can be better met, without worrying about the identification problem caused by environmental factors, the satisfaction of user experience is improved, and then, the accurate permission division is realized according to the actual environment, the processing mode is adaptively adjusted, and the accuracy of the socket intelligent control is improved.

[0082] Please refer to Figure 4 As shown in the figure, it is a logic flow chart of the preprocessing module of the embodiment of the application for marking the key socket, the preprocessing module is used to mark the key socket, wherein,

[0083] If the characteristic tendency coefficient of the socket meets the marking condition, the preprocessing module marks the socket as a key socket;

[0084] If the feature tendency coefficient of the socket does not meet the marking condition, the preprocessing module does not mark the socket;

[0085] The marking condition is that a difference between the feature tendency coefficient in a preset first monitoring time period and a preset second monitoring time period exceeds a preset feature tendency difference reference value.

[0086] Specifically, the timing relationship between the preset first monitoring time period and the preset second monitoring time period is that the first monitoring time period is before the second monitoring time period, and the duration of the first monitoring time period and the duration of the second monitoring time period can be set by a person skilled in the art according to the accuracy requirement of socket intelligent control. The higher the accuracy requirement, the longer the duration of the first monitoring time period and the duration of the second monitoring time period are set. The duration of the first monitoring time period and the duration of the second monitoring time period are the same, and the value range can be [1, 2], the interval unit is s, preferably, the duration of the first monitoring time period can be 1.5 s, and the duration of the second monitoring time period can be 1.5 s.

[0087] Specifically, the feature tendency difference reference value is the product of the average value of the feature tendency coefficient obtained by a plurality of experimental data and a tendency factor. The average value of the feature tendency coefficient is the average value of the feature tendency coefficient obtained by a plurality of experimental data. The tendency factor can be set by a user according to the actual age of the infant. The smaller the actual age of the infant, the smaller the tendency factor. The value range of the tendency factor can be [0.6, 0.8]. Preferably, when the actual age of the infant is 4 years old, the tendency factor can be 0.7.

[0088] Specifically, it can be understood that the activity range of the infant is limited and usually in a lower position, mainly concentrated in the lower part of the space. The infant is more likely to contact the socket in a relative horizontal position, while the socket above the space is usually not within the activity range. The difference between the feature tendency coefficient in the preset first monitoring time period and the preset second monitoring time period can represent the decay rate of high frequency. The larger the difference, the greater the change of the feature tendency coefficient. When the collection position and the sound source are in the same horizontal position, the sound wave propagation path is relatively direct, and the attenuation of high frequency component is relatively small, so the difference is small. When the collection position is above the sound source, the sound wave needs to propagate upward and will encounter more air resistance and reflection, resulting in faster attenuation of high frequency component, so the difference is large. Therefore, the sockets in the relative horizontal position and the space above can be distinguished by the feature tendency coefficient, and the key socket can be marked. The present application marks the key socket by comparing a plurality of feature tendency coefficients, and then realizes accurate permission division according to the actual environment, adaptively adjusts the processing mode, and improves the accuracy of socket intelligent control.

[0089] Specifically, the preprocessing module is configured to determine the mark socket and the plurality of feature sockets, wherein,

[0090] The preprocessing module sorts the initial time of the voiceprint signal collected by each socket in time sequence, and determines the socket corresponding to the initial time ranked first as the mark socket.

[0091] The preprocessing module calculates the similarity of the voiceprint signal of the reference socket and the voiceprint signal of the remaining sockets respectively, and determines the socket with a similarity greater than a preset similarity threshold as a feature socket.

[0092] Specifically, the preset similarity threshold can be set by a person skilled in the art according to the accuracy requirement of socket intelligent control. The higher the accuracy requirement, the greater the preset similarity threshold. The value range of the similarity threshold can be [0.6, 0.9], and preferably, the similarity threshold can be 0.7.

[0093] Specifically, the preprocessing module is configured to determine whether there is a risk of recognition deviation, wherein,

[0094] If the mark socket is not any of the feature sockets, the preprocessing module determines that there is a risk of recognition deviation.

[0095] If the mark socket is any of the feature sockets, the preprocessing module determines that there is no risk of recognition deviation.

[0096] Specifically, the preprocessing module is further configured to perform instruction compensation recognition on the voiceprint signal of each socket, wherein,

[0097] The preprocessing module determines the amplitude according to adjacent peaks and valleys on the voiceprint signal, determines a voiceprint signal region with an amplitude meeting an energy aggregation condition as an energy aggregation region, and recognizes a user instruction according to the energy aggregation region.

[0098] The energy aggregation condition is that the difference between the amplitudes does not exceed a preset difference threshold, and the amplitude exceeds a preset amplitude threshold.

[0099] Specifically, the preset difference threshold and the preset amplitude threshold can be set by a person skilled in the art according to the accuracy requirement of socket intelligent control. The higher the accuracy requirement, the smaller the preset difference threshold, and the greater the preset amplitude threshold. The value range of the preset difference threshold can be [0.01, 0.1], and the value range of the preset amplitude threshold can be [0.2, 0.5]. Preferably, the preset difference threshold can be 0.05, and the preset amplitude threshold can be 0.3.

[0100] Specifically, it can be understood that by acquiring the signal initial time of the voiceprint signal collected by each socket and determining the marker socket in chronological order, the time characteristics of sound propagation are utilized. In the ideal case without reverberation, the socket closer to the sound source will receive the sound signal earlier, and the socket with higher similarity to the reference socket voiceprint signal will have a closer distance to the sound source and receive the sound signal earlier. Reverberation causes the sound to reflect multiple times in space, causing other sockets to also receive reflected sound earlier, so the marker socket is not in the several feature sockets, and there is a risk of recognition deviation in the current environment. The voiceprint signal of each socket needs to be compensated and recognized according to the instruction. The difference between the amplitudes and the amplitude size can represent the energy aggregation area. Identifying the user's instruction in the energy aggregation area can effectively overcome the interference of reverberation, and then, accurate authority division according to the actual environmental conditions is realized, the processing mode is adaptively adjusted, and the accuracy of socket intelligent control is improved.

[0101] Specifically, the voice instruction matching module is used to determine the voice instruction matching mode, wherein,

[0102] If the socket control preprocessing mode is to mark the key socket, the voice instruction matching module determines that the voice instruction matching mode is to cut off the key socket.

[0103] If the socket control preprocessing mode is to compensate and recognize the voiceprint signal of each socket according to the instruction, the voice instruction matching module determines that the voice instruction matching mode is to acquire the socket state of the socket matched by the instruction compensation recognition for instruction re-verification, and to control the socket according to the verification result.

[0104] Specifically, the voice instruction matching module is also used for instruction re-verification, and controls the socket according to the verification result, wherein,

[0105] The voice instruction matching module pre-matches according to the instruction compensation recognition, and determines the socket state of the socket that meets the pre-matching result.

[0106] If the socket state meets the verification condition, the voice instruction matching module determines that the pre-matching result is correct, and controls the socket according to the pre-matching result.

[0107] If the socket state does not meet the verification condition, the voice instruction matching module determines that the pre-matching result is incorrect, and does not control the socket according to the pre-matching result.

[0108] The verification condition is that the pre-matched socket accesses the electrical equipment.

[0109] Specifically, the voiceprint signal of each socket is subjected to command compensation identification, the socket state of the socket subjected to command compensation identification matching is acquired, and command re-verification is performed, and socket control is performed according to the verification result. It can be understood that people will first insert the plug of the electrical appliance into the socket, and then control the socket to start the electrical appliance, so that it works normally. This is a way that conforms to the daily operation habit and the electrical appliance use process. If the socket is started first, and then the electrical appliance plug is inserted, an electric spark may be generated, which has certain safety hazards, and also does not conform to the operation habit of most people who first connect the device and then power on. Reverberation will cause distortion, delay and superposition of the voiceprint signal, increasing the difficulty of accurate identification of the voiceprint signal and judgment of the command. Although the processing mode is adjusted for command compensation identification, there may be misjudgment, which may cause safety hazards. The socket state re-verification can provide additional judgment basis for the socket control system, help to correct the false judgment caused by reverberation interference, and then realize accurate authority division according to the actual environment, adaptive adjustment of the processing mode, further matching verification for the interference environment, and improvement of the accuracy of the socket intelligent control.

[0110] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will fall within the protection scope of the present application.

[0111] The above description is only the preferred embodiments of the present application and is not used to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A smart socket control system based on voice command recognition, characterized in that, include: The instruction generation module includes a voiceprint acquisition unit for acquiring user voice and converting it into a voiceprint signal, a monitoring unit for acquiring socket status, and a setting unit for presetting instruction tags for each socket. The permission analysis module, which is connected to the instruction generation module, is used to mark the reference socket according to the voiceprint factor determined by the voiceprint signal, and to obtain the characteristic tendency coefficient of the reference socket to determine the user's permission category. The voiceprint factor is the average value of several amplitudes on the voiceprint signal obtained by the permission analysis module, and the characteristic tendency coefficient is the variance value of several peak spacings in the voiceprint signal obtained by the permission analysis module. The preprocessing module, which is connected to the instruction generation module and the permission analysis module respectively, is used to determine the socket control preprocessing method according to the permission category, which is to mark key sockets based on the comparison of several feature tendency coefficients, or to determine whether there is a risk of identification deviation based on the comparison of the marked sockets and several feature sockets. The preprocessing module is also used to perform instruction compensation recognition on the voiceprint signals of sockets with identification deviation risk. The preprocessing module calculates the similarity between the voiceprint signal of the reference socket and the voiceprint signals of the other sockets respectively, and determines the sockets with the similarity greater than the preset similarity threshold as feature sockets. The preprocessing module sorts the initial times of the voiceprint signals collected from each socket according to time sequence, and determines the socket corresponding to the first initial time as the marked socket. The feature socket is determined based on the voiceprint signal of the reference socket; The voice command matching module is connected to the command generation module and the preprocessing module respectively. It is used to determine the voice command matching mode as cutting off the key socket according to the socket control preprocessing method, or to obtain the socket status of the socket matched by the command compensation identification and verify the command again. The voice command matching module is also used to control the socket according to the verification result.

2. The smart socket control system based on voice command recognition according to claim 1, characterized in that, The reference socket is the socket corresponding to the voiceprint signal where the maximum value of the voiceprint factor is located.

3. The smart socket control system based on voice command recognition according to claim 2, characterized in that, The permission analysis module is used to determine the user's permission category, wherein... If the feature tendency coefficient meets the first permission determination condition, then the permission analysis module determines that the user's permission category is the first permission category; If the characteristic tendency coefficient does not meet the first permission determination condition, the permission analysis module determines that the user's permission category is the second permission category; The first permission determination condition is that the feature tendency coefficient obtained within the preset sampling period exceeds the preset speech feature tendency difference reference value.

4. The smart socket control system based on voice command recognition according to claim 3, characterized in that, The preprocessing module is used to determine the socket control preprocessing method, wherein... If the permission category is the first permission category, then the preprocessing module determines that the socket control preprocessing method is to mark key sockets based on the comparison of several feature tendency coefficients. If the permission category is the second permission category, the preprocessing module determines that the socket control preprocessing method is to determine whether there is a risk of identification deviation based on the comparison between the marked socket and several feature sockets, and to perform instruction compensation identification on the voiceprint signal of the socket with identification deviation risk.

5. The smart socket control system based on voice command recognition according to claim 4, characterized in that, The preprocessing module is used to mark key sockets, wherein, If the characteristic tendency coefficient of the socket meets the marking conditions, the preprocessing module marks the socket as a key socket; The marking condition is that the difference between the characteristic tendency coefficients in the preset first monitoring time period and the preset second monitoring time period exceeds the preset characteristic tendency difference reference value.

6. The smart socket control system based on voice command recognition according to claim 5, characterized in that, The preprocessing module is used to determine whether there is a risk of identification bias. If the marked socket is not any of the featured sockets, the preprocessing module determines that there is a risk of identification deviation. If the sign socket is any one of the feature sockets, the preprocessing module determines that there is no risk of identification deviation.

7. The smart socket control system based on voice command recognition according to claim 6, characterized in that, The preprocessing module is also used to perform instruction compensation recognition on the voiceprint signals of each socket, wherein... The preprocessing module determines the amplitude based on adjacent peaks and troughs on the voiceprint signal, identifies the voiceprint signal region whose amplitude meets the energy accumulation condition as the energy accumulation region, and identifies user commands based on the energy accumulation region. The energy accumulation condition is that the difference between the amplitudes does not exceed a preset difference threshold, and the amplitude exceeds a preset amplitude threshold.

8. The smart socket control system based on voice command recognition according to any one of claims 5 or 7, characterized in that, The voice command matching module is used to determine the voice command matching pattern, wherein... If the socket control preprocessing method is to mark key sockets, then the voice command matching module determines the voice command matching mode as cutting off the key sockets; If the preprocessing method for socket control is to perform command compensation recognition on the voiceprint signal of each socket, then the voice command matching module determines the voice command matching mode to obtain the socket status of the socket matched by command compensation recognition and verify the command again, and then controls the socket according to the verification result.

9. The smart socket control system based on voice command recognition according to claim 8, characterized in that, The voice command matching module is used to re-verify the command and control the socket based on the verification result. The voice command matching module performs pre-matching based on the command compensation recognition to determine the socket status of sockets that match the pre-matching results; If the socket status meets the verification conditions, the voice command matching module determines that the pre-matching result is correct and controls the socket according to the pre-matching result; The verification condition is that the pre-matched socket is connected to the electrical equipment.

Images