Whole-house smart home interaction and control system

Abstract

A whole-house smart home interaction and control system comprises a network transmission module, an equipment naming module, a sound receiver module, a control unit module, a data processing module, a sound recognition module and a control platform module. The smart home carries out data transmission through a wireless network, the control unit processes commands of users and controls the equipment, and the control unit distinguishes different users by analyzing voice characteristics of the users and provides different working styles of the smart equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of smart home, in particular to a whole-house smart home interaction and control system. BACKGROUND

[0002] In today's society, the development of smart home technology is increasingly mature, and more and more families begin to use intelligent devices and systems to improve the quality of life and convenience. However, the existing smart home system still has some limitations in interaction and control, such as single interaction mode between users and the system, and not enough intelligent control operation. In order to solve these problems, a new smart home interaction and control system is needed, which can better meet the needs of users and provide more intelligent and convenient home experience. It can realize multiple interaction modes such as voice control and gesture control, and has intelligent voice recognition and control capabilities, thereby improving the operation convenience and experience of users. The system should be able to adapt to the needs of different home environments, have good stability and reliability, and provide users with more intelligent and convenient home life experience. SUMMARY

[0003] The purpose of the present application is to provide a whole-house smart home interaction and control system, which aims to solve the problems mentioned in the background.

[0004] In order to achieve the above purpose, the present application proposes a whole-house smart home interaction and control system, which includes a network transmission module, a device naming module, a sound receiver module, a control unit module, a data processing module, a sound recognition module, and a control platform module. In the network transmission module, devices are connected and data are transmitted through wireless networks. In the device naming module, users name all devices individually and classify them. The sound receiver module is installed with an array of sound receivers in each room to collect user commands. The control unit module analyzes data and transmits commands through a central processor. The data processing module proposes a data dimension reduction algorithm to convert and reduce data. The sound recognition module classifies the collected sounds through a sound classification algorithm to distinguish different users. The control platform module controls and manages devices through a device operation panel.

[0005] Further, the network transmission module connects all devices in the whole-house smart home through wireless networks for data transmission.

[0006] Further, the device naming module names the devices in the whole-house smart home by users, and the naming method is set by users, and the naming result is unique.

[0007] Further, the sound receiver module is installed with an array of sound receivers in each room to collect user voice commands and transmit them through wireless networks.

[0008] Further, the control unit module is installed with a central server for data analysis and instruction transmission, the voice receiver collects the user's voice command and transmits it to the central server, the central server analyzes and obtains the user's instruction, and transmits the instruction to the corresponding device to control the device.

[0009] Further, the data dimension reduction processing algorithm, the central server processes the collected sound data, first, calculates the effective neighbor points of the high-dimensional data sample points, then, selects the effective neighbor points based on the maximum dimension parameter difference of the high-dimensional data sample points, represents the high-dimensional data sample points through the effective neighbor points, and then constructs an error function based on the weight matrix to process the low-dimensional data sample points.

[0010] Further, the data dimension reduction processing algorithm, the central server processes the collected sound data, first, calculates the effective neighbor points of the high-dimensional data sample points, then, selects the effective neighbor points based on the maximum dimension parameter difference of the high-dimensional data sample points, represents the high-dimensional data sample points through the effective neighbor points, and then constructs an error function based on the weight matrix to process the low-dimensional data sample points.

[0011] The collected sound data signal is first preprocessed, a signal length N is set, each point of the data signal is processed, and a processing function is defined As follows:

[0012]

[0013] Among them, Indicates the improvement parameter, Indicates the combination parameter, k indicates the value of each point in the signal length N, and then the preprocessed sound data signal is converted into a digital signal through analog-digital conversion. First, the data is reduced in dimension to improve the data processing efficiency. For a high-dimensional signal data set , it is expressed as , where , respectively, represent the 1st, 2nd, i-th, and n-th high-dimensional data sample points in the signal data set. The effective neighbor points of the high-dimensional data sample points , assuming the dimension is D, the high-dimensional data sample points are expressed as , where , respectively, represent the parameters of each dimension within the D dimensions, and the effective neighbor points of the high-dimensional data sample points are calculated as follows:

[0014]

[0015] Among them, , respectively, represent the parameters of each dimension within the D dimensions of the high-dimensional data sample points , , represents the maximum dimension parameter difference between the high-dimensional data sample points and the high-dimensional data sample points , and the high-dimensional data sample points The maximum dimension parameter difference of all the neighbor points is arranged from small to large, and the high-dimensional data sample points of the first k items are taken as the high-dimensional data sample points The effective neighbor points of the high-dimensional data sample points are represented by the effective neighbor points, and the high-dimensional data sample points The effective neighbor set of the high-dimensional data sample points is defined as , which is represented as , wherein The first, second and kth effective neighbor points of the high-dimensional data sample points Error function is constructed as follows:

[0016]

[0017] , wherein The weight of the kth effective neighbor point of the high-dimensional data sample points , wherein The 2-norm is calculated, the weight matrix is obtained when the error function is minimized, and the low-dimensional data sample point set The low-dimensional processing of the high-dimensional data sample points is performed, and the formula is as follows:

[0018]

[0019] , wherein The low-dimensional data sample point set is obtained by taking the weight matrix as the minimum error function, and the low-dimensional data sample point set The data dimension reduction processing algorithm provided by the application divides the neighbor points of the high-dimensional signal data into effective neighbor points, reduces the error of non-uniform sampling based on the selection of the first k items as effective neighbor points, constructs an error function, calculates the effective neighbor weight matrix through the error function, and performs low-dimensional representation of the high-dimensional data based on the effective neighbor weight matrix and the low-dimensional sample point set. The low-dimensional processing of the high-dimensional data, the data dimension reduction processing provided by the application reduces the complexity of high-dimensional data analysis and speeds up the data processing time.

[0020] Further, the sound classification algorithm identifies the user's voice and determines which user issued the command. First, the voice is feature extracted, and the processed voice data is converted, then a center value function is constructed, the center value is calculated, and then based on the center value, a set of optimal partition functions is used to divide the voice data by minimizing the error principle.

[0021] Further, the voice data division has the following detailed process:

[0022] First, the voice is feature extracted, and the processed voice data is converted, and the data discrimination degree is improved, and the formula is as follows:

[0023] ​

[0024] wherein, representing data of conversion data, representing data jth dimension value, representing conversion parameters, the center value is calculated based on the conversion data, and the formula is as follows:

[0025]

[0026] wherein, representing sampling frequency, m represents the number of filter groups, representing the center value function, using the center value as a feature, classifying the sound, based on the minimum error principle, dividing the data through a set of optimal partition functions, and the division formula is as follows:

[0027]

[0028] wherein respectively represent the bias of the partition interval, respectively represent the classification parameters of the first unknown quantity of the first optimal partition, the classification parameters of the second unknown quantity of the first optimal partition, the classification parameters of the first unknown quantity of the i-th optimal partition, the classification parameters of the second unknown quantity of the i-th optimal partition, the classification parameters of the first unknown quantity of the K-th optimal partition, and the classification parameters of the second unknown quantity of the K-th optimal partition. respectively represent the first unknown quantity of the first optimal partition, the second unknown quantity of the first optimal partition, the third unknown quantity of the first optimal partition, the first unknown quantity of the i-th optimal partition, the second unknown quantity of the i-th optimal partition, the third unknown quantity of the i-th optimal partition, the first unknown quantity of the K-th optimal partition, the second unknown quantity of the K-th optimal partition, and the third unknown quantity of the K-th optimal partition. respectively, the Euclidean distance between each feature point and each optimal partition is calculated through a set of optimal partitions, the Euclidean distance is used to judge the classification result, and the data is scored based on the partition structure. The highest score of the partition result is the category of the data. The sound classification algorithm proposed in the application first converts the data signal to improve the data signal differentiation, then calculates the center value, extracts the features through the center value, and divides the region categories through a set of optimal partition functions. The sound classification algorithm proposed in the application improves the accuracy of sound differentiation.

[0029] ​​​​​Further, the control platform module installs a device operation panel, and a user controls the device, checks device detailed information and manages intelligent devices through the operation panel.

[0030] Advantages

[0031] 1. The data dimension reduction processing algorithm provided by the application effectively divides the adjacent points of high-dimensional signal data, selects the first k items as effective adjacent points, reduces the error of non-uniform sampling, constructs an error function, calculates the effective adjacent point weight matrix through the error function, and then based on the effective adjacent point weight matrix and the low-dimensional sample point set, performs low-dimensional representation of the high-dimensional data, and performs low-dimensional processing on the high-dimensional data.

[0032] 2. The sound classification algorithm provided by the application first converts the data signal to improve the distinguishability of the data signal, then calculates the center value, extracts features through the center value, divides the region categories through a set of optimal division functions, and distinguishes the sound data. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0034] Figure 1 is a structural schematic diagram of the present application;

[0035] Figure 2 is a schematic diagram of the sound classification algorithm of the present application. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0037] In order to achieve the above object, the present application proposes a whole-house smart home interaction and control system, comprising a network transmission module, a device naming module, a sound receiver module, a control unit module, a data processing module, a sound recognition module, a control platform module; in the network transmission module, the devices are connected and data are transmitted through wireless network; in the device naming module, the user names all the devices individually and classifies the devices; the sound receiver module is installed with a sound receiver array in each room to collect the user's commands; the control unit module analyzes the data and transmits the commands through the central processor; the data processing module proposes a data dimension reduction processing algorithm to convert and reduce the dimension of the data; the sound recognition module classifies the collected sound through a sound classification algorithm to distinguish different users; the control platform module controls and manages the devices through the device operation panel.

[0038] Specifically, the network transmission module, all the devices in the whole-house smart home are connected and data are transmitted through wireless network.

[0039] Specifically, the device naming module, the user names the devices in the whole-house smart home, the naming method is set by the user, and the naming result is unique.

[0040] Specifically, the sound receiver module, a sound receiver array is installed in each room to collect the user's voice commands and transmit them through wireless network.

[0041] Specifically, the control unit module, a central server is installed to analyze the data and transmit the instructions, the user's voice commands collected by the sound receiver are transmitted to the central server, the central server analyzes and obtains the user's instructions, and transmits the instructions to the corresponding devices to control the devices.

[0042] Specifically, the data dimension reduction processing algorithm, the central server processes the collected sound data, first, calculates the effective neighbor points of the high-dimensional data sample points, then, selects the effective neighbor points based on the maximum dimension parameter difference of the high-dimensional data sample points, represents the high-dimensional data sample points through the effective neighbor points, and then constructs an error function based on the weight matrix to process the low-dimensional data sample points in detail as follows:

[0043] The collected sound data signal is first preprocessed, a signal length N is set, each point of the data signal is processed, and a processing function is defined as follows:

[0044]

[0045] Among them, Indicates the improved parameter, ​The combination parameter is represented by k, and k represents the value of each point in the signal length N. The preprocessed sound data signal is then subjected to analog-to-digital conversion to convert it into a digital signal. The data is first subjected to dimension reduction to improve data processing efficiency. For high-dimensional signal data sets , are represented as , where represent the 1st, 2nd, i-th, and n-th high-dimensional data sample points in the signal data set, respectively. The high-dimensional data sample point is represented as , where represent the parameters of each dimension within the D dimensions, and the effective neighbor points of the high-dimensional data sample point are calculated using the following formula:

[0046]

[0047] , where represent the parameters of each dimension within the D dimensions of the high-dimensional data sample point , and represents the maximum dimension parameter difference between the high-dimensional data sample point and the high-dimensional data sample point . The maximum dimension parameter differences of all neighbor points of the high-dimensional data sample point are sorted from small to large, and the first k high-dimensional data sample points are taken as the effective neighbor points of the high-dimensional data sample point . The high-dimensional data sample point is represented by its effective neighbor points, and the effective neighbor set of the high-dimensional data sample point is defined as , which is represented as , where represent the 1st, 2nd, and k-th effective neighbor points of the high-dimensional data sample point . The error function is constructed as follows:

[0048]

[0049] , where represents the weight of the k-th effective neighbor point of the high-dimensional data sample point , and represents the 2-norm. The weight matrix that minimizes the error function is obtained, and based on the weight matrix, the low-dimensional data sample set is constructed. The low-dimensional processing of the high-dimensional data sample is performed, and the formula is as follows:

[0050]

[0051] , where The weight matrix is expressed as minimizing the error function, and a low-dimensional data sample point set is obtained .

[0052] Specifically, the sound classification algorithm identifies the user's voice and determines which user issued the command. First, the sound is feature extracted, the processed sound data is converted, and then the center value function is constructed, the center value is calculated, and then based on the center value, a set of optimal partition functions is used to divide the sound data by minimizing the error principle. The detailed process is as follows:

[0053] First, the sound is feature extracted, the processed sound data is converted, and the data discrimination degree is improved. The formula is as follows:

[0054]

[0055] Wherein, represents the conversion data of the data , represents the jth dimension value of the data , represents the conversion parameter, and the center value is calculated based on the conversion data. The formula is as follows:

[0056]

[0057] Wherein, represents the sampling frequency, and m represents the number of filter groups, represents the center value function, and the center value is used as a feature for sound classification. Based on the minimum error principle, a set of optimal partition functions is used for data division. The division formula is as follows:

[0058]

[0059] Wherein respectively represent the bias of the partition interval, respectively represent the classification parameters of the first unknown quantity of the first optimal partition, the classification parameters of the second unknown quantity of the first optimal partition, the classification parameters of the first unknown quantity of the i-th optimal partition, the classification parameters of the second unknown quantity of the i-th optimal partition, the classification parameters of the first unknown quantity of the K-th optimal partition, and the classification parameters of the second unknown quantity of the K-th optimal partition. respectively represent the first unknown quantity of the first optimal partition, the second unknown quantity of the first optimal partition, the third unknown quantity of the first optimal partition, the first unknown quantity of the i-th optimal partition, the second unknown quantity of the i-th optimal partition, and the third unknown quantity of the i-th optimal partition. ​​​​The unknown quantity, the first unknown quantity of the Kth optimal partition, the second unknown quantity of the Kth optimal partition, and the unknown quantity of the Kth optimal partition. For each unknown variable, the Euclidean distance between each feature point and each optimal partition is calculated using a set of optimal partitions. The classification result is determined by the Euclidean distance. The data is scored based on the partition structure, and the partition with the highest score is the category of the data.

[0060] Specifically, the control platform module includes an equipment operation panel, through which users can control the equipment, view detailed equipment information, and manage smart devices.

[0061] In a specific embodiment, the user controls the device, views detailed device information, and manages the smart device through the device operation panel. MEMS microphones are deployed in each room, with a microphone spacing greater than 0.5 meters and less than 1.2 times the sound wavelength. The user operates the device through voice commands. The user can issue commands from any location in the home. The sound collector array transmits the received sound information to the central processor. The central processor processes the sound signal, distinguishes and analyzes the user's commands, interprets the commands, and then transmits the commands to the corresponding devices, which then perform the corresponding operations.

[0062] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0063] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A whole-house smart home interaction and control system, comprising a network transmission module, a device naming module, a sound receiver module, a control unit module, a data processing module, a sound recognition module, and a control platform module; in the network transmission module, devices connect and transmit data via a wireless network; in the device naming module, users personalize and categorize all devices; the sound receiver module has a sound receiver array installed in the room to collect user commands; the control unit module analyzes data and transmits commands through a central processor; the data processing module proposes a data dimensionality reduction algorithm to transform and reduce the dimensions of the data; The voice recognition module uses a voice classification algorithm to categorize the collected voices and distinguish between different users; The control platform module controls and manages the equipment through the equipment operation panel.

2. The whole-house smart home interaction and control system according to claim 1, characterized in that, The network transmission module enables all devices in the smart home to connect and transmit data via a wireless network.

3. The whole-house smart home interaction and control system according to claim 1, characterized in that, The device naming module allows users to name devices in their smart home system. Users can set their own naming methods, and the naming results are unique.

4. The whole-house smart home interaction and control system according to claim 1, characterized in that, The sound receiver module is installed in each room with a sound receiver array to collect users' voice commands and transmit them via a wireless network.

5. The whole-house smart home interaction and control system according to claim 1, characterized in that, The control unit module is equipped with a central server for data analysis and command transmission. The user's voice commands collected by the sound receiver are transmitted to the central server, which analyzes them and obtains the user's instructions. The central server then transmits the instructions to the corresponding devices to control them.

6. The whole-house smart home interaction and control system according to claim 1, characterized in that, The data dimensionality reduction algorithm performs data transformation and dimensionality reduction on the collected sound data. First, it calculates the effective nearest neighbors of the high-dimensional data sample points. Then, based on the maximum dimensionality parameter difference between the nearest neighbors of the high-dimensional data sample points, it selects the effective nearest neighbors and represents the high-dimensional data sample points through the effective neighbors. Next, it constructs an error function and performs low-dimensional processing on the high-dimensional data sample points based on the weight matrix.

7. A whole-house smart home interaction and control system according to claim 6, characterized in that, The detailed process of data transformation and dimensionality reduction is as follows: The audio signals collected by the sound controller and transmitted to the control unit are converted from analog to digital signals. First, a dimensionality reduction operation is performed on the data, especially for high-dimensional signal datasets. Among them, high-dimensional data sample points Assuming the dimension is D, then the high-dimensional data sample points are represented as follows: ,in Let each dimension of D represent a parameter, and calculate the high-dimensional data sample points. The effective nearest neighbor is calculated using the following formula: in Representing high-dimensional data sample points respectively The parameters of each dimension within dimension D. Represents high-dimensional data sample points With high-dimensional data sample points Find the maximum dimensional parameter difference to obtain the high-dimensional data sample points. Arrange all nearest neighbors by their maximum dimensionality difference in ascending order, and select the high-dimensional data sample points from the first k directions as the high-dimensional data sample points. Effective nearest neighbors are used to represent high-dimensional data sample points. Define its effective neighbor set as , represented as ,in, Represents high-dimensional data sample points Construct the error function for the 1st, 2nd, and kth effective neighbors. Define a weight matrix to represent high-dimensional data sample points, find the weight matrix that minimizes the error function, and construct a low-dimensional data sample point set. The formula for performing low-dimensional processing on high-dimensional data sample points is as follows: in, Describing the 2-norm, This represents the weight matrix that minimizes the error function, resulting in the low-dimensional data sample point set. .

8. A whole-house smart home interaction and control system according to claim 1, characterized in that, The aforementioned sound classification algorithm identifies the user's voice and determines which user issued the command. First, it extracts features from the sound and transforms the processed sound data. Then, it constructs a center value function and calculates the center value. Next, based on the center value, it uses a set of optimal partitioning functions to partition the sound data by minimizing the error.

9. A whole-house smart home interaction and control system according to claim 8, characterized in that, The detailed process of dividing the sound data is as follows: First, feature extraction is performed on the sound data. The processed sound data is then transformed to improve data discriminability. Based on the transformed data, the center value is calculated using the following formula: in, This represents the sampling frequency, and m represents the number of filter groups. This indicates a data transformation operation. The center value function is used to classify sounds using the center value as a feature. Based on the principle of minimizing error, the data is divided using a set of optimal partitioning functions. The partitioning formula is as follows: in These represent the bias values ​​for dividing the intervals. Let the classification parameters of the first unknown variable in the first optimal partition and the classification parameters of the first optimal partition be respectively represented. The classification parameters of the first unknown variable, the classification parameters of the first unknown variable in the i-th optimal partition, and the classification parameters of the i-th optimal partition. The classification parameters of the first unknown variable in the Kth optimal partition, and the classification parameters of the first unknown variable in the Kth optimal partition. Classification parameters for an unknown quantity Represent the first unknown of the first optimal partition, the second unknown of the first optimal partition, and the third unknown of the first optimal partition, respectively. The unknown quantity, the first unknown quantity of the i-th optimal partition, the second unknown quantity of the i-th optimal partition, and the first unknown quantity of the i-th optimal partition. The unknown quantity, the first unknown quantity of the Kth optimal partition, the second unknown quantity of the Kth optimal partition, and the unknown quantity of the Kth optimal partition. Given an unknown quantity, the Euclidean distance between each feature point and each optimal partition is calculated using a set of optimal partitions. The classification result is then determined using the Euclidean distance.

10. A whole-house smart home interaction and control system according to claim 1, characterized in that, The control platform module is equipped with a device operation panel, through which users can control the device, view detailed device information, and manage the smart device.

Images