CN115171317A - Internet of things smart home method and system and electronic equipment - Google Patents

Internet of things smart home method and system and electronic equipment Download PDF

Info

Publication number
CN115171317A
CN115171317A CN202210222195.9A CN202210222195A CN115171317A CN 115171317 A CN115171317 A CN 115171317A CN 202210222195 A CN202210222195 A CN 202210222195A CN 115171317 A CN115171317 A CN 115171317A
Authority
CN
China
Prior art keywords
density map
gaussian density
vector
gaussian
matrix
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210222195.9A
Other languages
Chinese (zh)
Inventor
张文平
白维朝
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Isurpass Technology Co ltd
Original Assignee
Shenzhen Isurpass Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Isurpass Technology Co ltd filed Critical Shenzhen Isurpass Technology Co ltd
Priority to CN202210222195.9A priority Critical patent/CN115171317A/en
Publication of CN115171317A publication Critical patent/CN115171317A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • G08B17/06—Electric actuation of the alarm, e.g. using a thermally-operated switch
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
    • G08B17/117—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means by using a detection device for specific gases, e.g. combustion products, produced by the fire
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B25/00—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems
    • G08B25/01—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium
    • G08B25/08—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium using communication transmission lines

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Emergency Management (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Alarm Systems (AREA)

Abstract

The application discloses thing networking smart home method, system and electronic equipment, it is to each through time sequence encoder the sensor is in the measured value of a plurality of predetermined time points carries out high dimension correlation feature extraction to when the high dimension characteristic aspect fuses, because data distribution itself accords with the gaussian distribution, just each eigenvector has huge difference in the aspect of the characteristic field, consequently can be used as the characteristic of convolutional neural network's study target to the gaussian density map, carries out the feature fusion through the mean value and the oblique square difference iterative mode of gaussian density map. In this way, again through an iterative approach based on likelihood maximization, the target density map can be dynamically updated to modify the feature domain bias to help the convolutional neural network learn to a consistent feature representation and is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region. Therefore, the fire disaster at home can be accurately and timely pre-warned.

Description

Internet of things smart home method and system and electronic equipment
Technical Field
The application relates to the field of smart home, and more particularly, to an internet of things smart home method, system and electronic device.
Background
With the aggravation of the rhythm of life in modern cities and the increase of working pressure, more and more people desire a safe, comfortable and intelligent wind shelter. The desire is made possible by the rapid development of technologies such as the internet of things technology, the wireless communication technology, the embedded technology and the data fusion technology, so that intelligent and safe home life is no longer far out of reach.
In order to meet the requirements of users for intelligent home security and to take the disadvantages of the current intelligent home security into consideration, for example, most detection modules in the intelligent home security system only judge whether the decision is abnormal according to the data detected by a single sensor, especially the detection of fire, so that the obtained result has low precision and false alarm and missed alarm conditions exist. Therefore, in order to accurately and timely early warn the fire in the house by using the multi-sensor data fusion method, an internet of things intelligent home scheme is expected.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent home method, system and electronic equipment of thing networking, it is to each through time sequence encoder the sensor is in the measured value of a plurality of predetermined time points carries out high dimension correlation feature extraction to when high dimension feature level fuses, because data distribution itself accords with the gaussian distribution, just there is huge difference in the aspect of the characteristic field in each eigenvector, consequently can be used as the characteristic of convolutional neural network's learning target to the gaussian density map, carry out the feature fusion through the mean value and the oblique square difference iterative mode of gaussian density map. In this way, the target density map can be dynamically updated to modify the feature domain shifts to help the convolutional neural network learn to a consistent feature representation, and which is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region, again in an iterative manner based on likelihood maximization. Therefore, the fire disaster at home can be accurately and timely pre-warned.
According to one aspect of the application, an internet of things smart home method is provided, which includes:
obtaining measurements at a plurality of predetermined points in time from a plurality of sensors deployed in a home, wherein the plurality of sensors includes a temperature sensor, a smoke detector, and a carbon monoxide detector;
respectively arranging the measured values of the sensors at the preset time points into input vectors, and then respectively passing through a time sequence encoder comprising a one-dimensional convolutional layer and a full connection layer to obtain a first characteristic vector, a second characteristic vector and a third characteristic vector;
constructing first to third gaussian density maps of the first to third eigenvectors, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector;
fusing the first to third Gaussian density maps based on an iterative mode of likelihood maximization to obtain a final Gaussian density map, wherein the iterative mode of likelihood maximization represents that a mean vector of the fused Gaussian density map is a point product between mean vectors of two Gaussian density maps to be fused, and a covariance matrix of the fused Gaussian density map is a transpose of a product of a covariance matrix of the first Gaussian density map to be fused and a tensor product of a covariance matrix of the second Gaussian density map to be fused and a mean vector of the second Gaussian density map to be fused and a tensor product of the covariance matrix of the second Gaussian density map to be fused and the mean vector of the second Gaussian density map to be fused;
performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to obtain a classification matrix; and
and passing the classification matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire early warning is generated or not.
In the above method for internet of things smart home, after arranging the measured values of the sensors at the predetermined time points into input vectors respectively, obtaining a first feature vector, a second feature vector and a third feature vector by a time sequence encoder including a one-dimensional convolutional layer and a full link layer, respectively, the method includes: respectively encoding the input vector by using a full connection layer of the time sequence encoder according to the following formula to extract high-dimensional implicit characteristics of characteristic values of all positions in the input vector, wherein the formula is
Figure DEST_PATH_IMAGE001
Wherein
Figure DEST_PATH_IMAGE003
Is a function of the input vector or vectors,
Figure 925285DEST_PATH_IMAGE004
is the output vector of the digital video signal,
Figure DEST_PATH_IMAGE005
is a matrix of the weights that is,
Figure DEST_PATH_IMAGE007
is a vector of the offset to be used,
Figure 869101DEST_PATH_IMAGE008
represents a matrix multiplication; and one-dimensional convolution encoding the input vector using the one-dimensional convolution layer of the time-series encoder to extract the input vectorHigh-dimensional implicit relevance features of the relevance between the feature values of the positions in the input vector, wherein the formula is:
Figure 539117DEST_PATH_IMAGE010
wherein the content of the first and second substances,ais a convolution kernelxA width in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wthe size of the convolution kernel.
In the internet of things smart home method, fusing the first to third gaussian density maps based on an iterative manner of likelihood maximization to obtain a final gaussian density map, including: fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization to obtain a final Gaussian density map; wherein the formula is:
Figure 653703DEST_PATH_IMAGE012
Figure 434577DEST_PATH_IMAGE014
wherein
Figure DEST_PATH_IMAGE015
Which represents a dot product of the light beam,
Figure 731435DEST_PATH_IMAGE008
express tensor multiplication, and
Figure 572353DEST_PATH_IMAGE016
，
Figure DEST_PATH_IMAGE017
。
in the internet of things smart home method, the gaussian discretization of the gaussian distribution of each position in the final gaussian density map to obtain a classification matrix includes: performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to convert the Gaussian distribution of each position into a one-dimensional vector; and two-dimensionally arranging the one-dimensional vectors corresponding to the Gaussian distribution of each position in the final Gaussian density map to obtain the classification matrix.
In the method for intelligent home furnishing through internet of things, the classification matrix is used for obtaining a classification result through a classifier, and the classification result is used for indicating whether fire early warning is generated or not, and the method comprises the following steps: processing the classification matrix using the classifier to generate the classification result with the formula:
Figure 439814DEST_PATH_IMAGE018
wherein
Figure DEST_PATH_IMAGE019
Representing the projection of the classification matrix as a vector,
Figure 775112DEST_PATH_IMAGE020
to
Figure DEST_PATH_IMAGE021
Is a weight matrix of the fully connected layers of each layer,
Figure 411630DEST_PATH_IMAGE022
to is that
Figure DEST_PATH_IMAGE023
A bias matrix representing the fully connected layers of each layer.
In the internet of things smart home method, the internet of things smart home method further includes responding to the classification result to generate a fire early warning, and sending a fire early warning prompt.
According to another aspect of the application, an internet of things smart home system is provided, which includes:
a measurement value acquisition unit for acquiring measurement values of a plurality of sensors disposed in a home at a plurality of predetermined time points, wherein the plurality of sensors includes a temperature sensor, a smoke detector, and a carbon monoxide detector;
the encoding unit is used for respectively arranging the measured values of the sensors obtained by the measured value obtaining unit at the plurality of preset time points into input vectors and then respectively passing through a time sequence encoder comprising a one-dimensional convolution layer and a full connection layer to obtain a first characteristic vector, a second characteristic vector and a third characteristic vector;
a gaussian density map constructing unit, configured to construct first to third gaussian density maps of the first to third eigenvectors obtained by the encoding unit, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector;
a likelihood maximization fusion unit, configured to fuse the first to third gaussian density maps obtained by the gaussian density map construction unit based on an iterative manner of likelihood maximization to obtain a final gaussian density map, where the iterative manner of likelihood maximization indicates that a mean vector of the fused gaussian density map is a dot product between mean vectors of two gaussian density maps to be fused, and a covariance matrix of the fused gaussian density map is a transpose of a product between a covariance matrix of the first gaussian density map to be fused and a tensor product of a covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused and a tensor product of the covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused;
a gaussian discretization unit, configured to perform gaussian discretization on the gaussian distribution at each position in the final gaussian density map obtained by the likelihood maximization fusion unit to obtain a classification matrix; and
and the classification unit is used for enabling the classification matrix obtained by the Gaussian discretization unit to pass through a classifier so as to obtain a classification result, and the classification result is used for indicating whether fire early warning is generated or not.
In the above-mentioned thing networking smart home systems, the coding unit includes: a full-link layer feature extraction unit, configured to use a full-link layer of the time sequence encoder to encode the input vector by using a formula to extract high-dimensional implicit features of feature values of each position in the input vector, where the formula is
Figure 689027DEST_PATH_IMAGE001
In which
Figure 761894DEST_PATH_IMAGE003
Is a function of the input vector or vectors,
Figure 884571DEST_PATH_IMAGE004
is the output vector of the output vector,
Figure 110016DEST_PATH_IMAGE005
is a matrix of weights that is a function of,
Figure 27156DEST_PATH_IMAGE007
is a vector of the offset to be used,
Figure 603631DEST_PATH_IMAGE008
represents a matrix multiplication; and a one-dimensional convolutional layer feature extraction unit, configured to perform one-dimensional convolutional coding on the input vector by using a one-dimensional convolutional layer of the time-series encoder according to the following formula to extract a high-dimensional implicit correlation feature of a correlation between feature values of each position in the input vector, where the formula is:
Figure 15152DEST_PATH_IMAGE010
wherein the content of the first and second substances,ais a convolution kernel inxA width in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wthe size of the convolution kernel.
In the foregoing internet of things smart home system, the likelihood maximization fusion unit is further configured to: fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization according to the following formula to obtain a final Gaussian density map;
wherein the formula is:
Figure 829524DEST_PATH_IMAGE012
Figure 448725DEST_PATH_IMAGE014
wherein
Figure 981337DEST_PATH_IMAGE015
It is shown that the dot-product,
Figure 445817DEST_PATH_IMAGE008
express tensor multiplication, and
Figure 380274DEST_PATH_IMAGE016
，
Figure 154064DEST_PATH_IMAGE017
。
in the above-mentioned thing networking smart home systems, the gaussian discretization unit is further configured to: performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map so as to convert the Gaussian distribution of each position into a one-dimensional vector; and two-dimensionally arranging the one-dimensional vectors corresponding to the Gaussian distribution of each position in the final Gaussian density map to obtain the classification matrix.
In the above-mentioned intelligent home systems of internet of things, instituteThe classification unit is further configured to: processing the classification matrix using the classifier to generate the classification result with the formula:
Figure 173973DEST_PATH_IMAGE018
in which
Figure 442143DEST_PATH_IMAGE019
Representing the projection of the classification matrix as a vector,
Figure 231108DEST_PATH_IMAGE020
to is that
Figure 926531DEST_PATH_IMAGE021
Is a weight matrix of the fully connected layers of each layer,
Figure 168157DEST_PATH_IMAGE022
to
Figure 990750DEST_PATH_IMAGE023
A bias matrix representing the fully connected layers of each layer.
In the internet of things intelligent home system, the internet of things intelligent home method further comprises responding to the classification result to generate fire early warning and sending out a fire early warning prompt.
According to still another aspect of the present application, there is provided an electronic apparatus including: a processor; and a memory having stored therein computer program instructions that, when executed by the processor, cause the processor to perform the internet of things smart home method as described above.
According to yet another aspect of the present application, there is provided a computer readable medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the internet of things smart home method as described above.
Compared with the prior art, the application provides an intelligent home method, system and electronic equipment of thing networking, it is through time sequence encoder to each the sensor is in the measured value of a plurality of predetermined time points carries out high dimension correlation feature extraction to when high dimension feature level fuses, because data distribution itself accords with gaussian distribution, just each eigenvector has huge difference in the aspect of the characteristic field, consequently can be used as the characteristic of the learning target of convolution neural network to the gaussian density picture, carries out the feature fusion through the mean value and the oblique square difference iterative mode of gaussian density picture. In this way, the target density map can be dynamically updated to modify the feature domain shifts to help the convolutional neural network learn to a consistent feature representation, and which is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region, again in an iterative manner based on likelihood maximization. Therefore, the fire disaster at home can be accurately and timely pre-warned.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally indicate like parts or steps.
Fig. 1 is an application scene diagram of an internet of things smart home method according to an embodiment of the application.
Fig. 2 is a flowchart of an internet of things smart home method according to an embodiment of the application.
Fig. 3 is a schematic architecture diagram of an internet of things smart home method according to an embodiment of the application.
Fig. 4 is a block diagram of an internet of things smart home system according to an embodiment of the application.
Fig. 5 is a block diagram of an electronic device according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Overview of a scene
As mentioned above, as the pace of life in modern cities increases and the working pressure increases, more and more people desire a safe, comfortable and intelligent wind shelter. The desire is made possible by the rapid development of technologies such as internet of things, wireless communication technology, embedded technology and data fusion, so that intelligent and safe home life is no longer far out of reach.
In order to meet the requirements of users for intelligent home security and to take the disadvantages of the current intelligent home security into consideration, for example, most detection modules in the intelligent home security system only judge whether the decision is abnormal according to the data detected by a single sensor, especially the detection of fire, so that the obtained result has low precision and false alarm and missed alarm conditions exist. Therefore, in order to accurately and timely early warn the fire in the house by using the multi-sensor data fusion method, an internet of things intelligent home scheme is expected.
And it is considered that when multi-sensor information fusion is performed, extracted features need to be effectively fused in a high-dimensional feature space according to the data properties of each sensor so as to obtain features for classification, which express accuracy.
Based on this, in the technical scheme of this application, at first through a plurality of sensors that dispose in the family obtain the measured value of each sensor in a plurality of predetermined time points, here, a plurality of sensors include temperature sensor, smoke detector and carbon monoxide detector, then arrange the data of each sensor according to time as input vector after, obtain first through N eigenvector through the time sequence encoder.
Constructing a Gaussian density map for each feature vector
Figure 899800DEST_PATH_IMAGE024
Wherein
Figure DEST_PATH_IMAGE025
Is the desired vector, i.e. the feature vector itself, and
Figure DEST_PATH_IMAGE027
is an autocovariance matrix, is the variance between the eigenvalues of each two positions of the eigenvector.
Then, the above gaussian density maps are fused in an iterative manner based on likelihood maximization, which can be expressed as:
Figure 297284DEST_PATH_IMAGE012
Figure 557364DEST_PATH_IMAGE014
wherein
Figure 416604DEST_PATH_IMAGE015
Which represents a dot product of the light beam,
Figure 180161DEST_PATH_IMAGE008
express tensor multiplication, and
Figure 951808DEST_PATH_IMAGE016
，
Figure 433605DEST_PATH_IMAGE017
。
fusing N Gaussian density maps in a sequential manner to finally obtain a final Gaussian density map
Figure 847269DEST_PATH_IMAGE028
。
And finally, carrying out Gaussian discretization on the final Gaussian density graph to obtain a classification matrix, and obtaining a classification result through a classifier.
It should be understood that, when the high-dimensional features of the convolutional neural network are extracted and fused at the high-dimensional feature level, based on the fact that the data distribution conforms to the gaussian distribution itself and each feature vector has great difference in the feature domain, the feature fusion is performed in an iterative manner of mean and oblique square difference of the gaussian density map with respect to the characteristic that the gaussian density map can be used as the learning target of the convolutional neural network.
In this way, by an iterative approach based on likelihood maximization, the target density map can be dynamically updated to modify the feature domain shifts to help the convolutional neural network learn to a consistent feature representation, and is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region.
Based on this, the application provides an internet of things smart home method, which includes: obtaining measurements at a plurality of predetermined points in time from a plurality of sensors deployed in a home, wherein the plurality of sensors comprises a temperature sensor, a smoke detector, and a carbon monoxide detector; respectively arranging the measured values of the sensors at the preset time points into input vectors, and then respectively passing through a time sequence encoder comprising a one-dimensional convolutional layer and a full connection layer to obtain a first characteristic vector, a second characteristic vector and a third characteristic vector; constructing first to third gaussian density maps of the first to third eigenvectors, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector; fusing the first to third Gaussian density maps based on an iterative mode of likelihood maximization to obtain a final Gaussian density map, wherein the iterative mode of likelihood maximization represents that a mean vector of the fused Gaussian density map is a point product between mean vectors of two Gaussian density maps to be fused, and a covariance matrix of the fused Gaussian density map is a transpose of a product of a covariance matrix of the first Gaussian density map to be fused and a tensor product of a covariance matrix of the second Gaussian density map to be fused and a mean vector of the second Gaussian density map to be fused and a tensor product of the covariance matrix of the second Gaussian density map to be fused and the mean vector of the second Gaussian density map to be fused; performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to obtain a classification matrix; and passing the classification matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire early warning is generated or not.
Fig. 1 illustrates an application scene diagram of an internet of things smart home method according to an embodiment of the application. As shown in fig. 1, in this application scenario, first, measurement values of a plurality of sensors (e.g., T as illustrated in fig. 1) disposed in a home (e.g., H as illustrated in fig. 1) are acquired at a plurality of predetermined time points, wherein the plurality of sensors include a temperature sensor, a smoke detector, and a carbon monoxide detector. Then, the measured values of the sensors are input into a server (e.g., S as illustrated in fig. 1) deployed with an internet of things smart home algorithm, where the server can process the measured values of the sensors by the internet of things smart home algorithm to generate a classification result indicating whether a fire alarm is generated.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary method
Fig. 2 illustrates a flowchart of an internet of things smart home method according to an embodiment of the present application. As shown in fig. 2, the method for internet of things smart home according to the embodiment of the application includes the steps: s110, obtaining measurement values of a plurality of sensors deployed in a home at a plurality of preset time points, wherein the plurality of sensors comprise a temperature sensor, a smoke detector and a carbon monoxide detector; s120, after the measured values of the sensors at the preset time points are respectively arranged into input vectors, the input vectors respectively pass through a time sequence encoder comprising a one-dimensional convolution layer and a full connection layer to obtain a first characteristic vector, a second characteristic vector and a third characteristic vector; s130, constructing first to third gaussian density maps of the first to third eigenvectors, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector; s140, fusing the first to third gaussian density maps based on an iterative manner of likelihood maximization to obtain a final gaussian density map, where the iterative manner of likelihood maximization represents that a mean vector of the fused gaussian density map is a point product between mean vectors of two gaussian density maps to be fused, and a covariance matrix of the fused gaussian density map is a transpose of a product between a covariance matrix of the first gaussian density map to be fused and a tensor product of a covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused, multiplied by a tensor product of the covariance matrix of the second gaussian density map to be fused and the mean vector of the second gaussian density map to be fused; s150, carrying out Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to obtain a classification matrix; and S160, passing the classification matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire early warning is generated or not.
Fig. 3 illustrates an architecture diagram of an internet of things smart home method according to an embodiment of the present application. As shown in fig. 3, in the network architecture, firstly, after arranging the obtained measured values (e.g., P1, P2, P3 as illustrated in fig. 3) of each of the sensors at the plurality of predetermined time points respectively as input vectors (e.g., V1, V2, V3 as illustrated in fig. 3), respectively, pass through a time-sequence encoder (e.g., E as illustrated in fig. 3) including a one-dimensional convolutional layer and a fully-connected layer to obtain a first eigenvector (e.g., VF1 as illustrated in fig. 3), a second eigenvector (e.g., VF2 as illustrated in fig. 3), and a third eigenvector (e.g., VF3 as illustrated in fig. 3); s130, constructing first to third gaussian density maps (e.g., GD1, GD2, GD3 as illustrated in fig. 3) for the first to third eigenvectors; s140, fusing the first to third gaussian density maps based on an iterative manner of likelihood maximization to obtain a final gaussian density map (e.g., GD as illustrated in fig. 3); s150, performing gaussian discretization on the gaussian distribution of each position in the final gaussian density map to obtain a classification matrix (e.g., M as illustrated in fig. 3); and, finally, passing the classification matrix through a classifier (e.g., a classifier as illustrated in fig. 3) to obtain a classification result, which is used to indicate whether a fire pre-warning is generated.
In steps S110 and S120, measurement values of a plurality of sensors deployed in a home at a plurality of predetermined time points are obtained, wherein the plurality of sensors include a temperature sensor, a smoke detector and a carbon monoxide detector, and after the measurement values of the plurality of sensors at the predetermined time points are respectively arranged as input vectors, the measurement values are respectively passed through a time sequence encoder comprising a one-dimensional convolutional layer and a fully-connected layer to obtain a first eigenvector, a second eigenvector and a third eigenvector. As described above, in order to avoid determining whether a decision is abnormal only by data detected by a single sensor, and further avoid a loss of a determination result, in the technical solution of the present application, it is desirable to perform accurate and timely early warning on a fire at home by using a multi-sensor data fusion method, and consider that when performing multi-sensor information fusion, extracted features need to be effectively fused in a high-dimensional feature space according to data properties of the sensors, so as to obtain features for classification, which are expressed accurately.
That is, specifically, in the technical solution of the present application, first, the measured values of the respective sensors are acquired at a plurality of predetermined time points by a plurality of sensors disposed in a home, where the plurality of sensors include a temperature sensor, a smoke detector, and a carbon monoxide detector. Then, after the measured values of the plurality of predetermined time points are respectively arranged into input vectors, the input vectors are respectively encoded in a time sequence encoder comprising a one-dimensional convolutional layer and a full-link layer, so as to respectively extract high-dimensional associated features of the data of the temperature, the smoke and the carbon monoxide at each time point, and thus a first feature vector, a second feature vector and a third feature vector are obtained.
Specifically, in this embodiment of the present application, the process of arranging the measured values of the sensors at the predetermined time points into input vectors and then obtaining a first feature vector, a second feature vector, and a third feature vector by passing through a time-sequence encoder including a one-dimensional convolutional layer and a fully-connected layer, respectively, includes: firstly, the input vector is respectively coded by using a full-connection layer of the time sequence coder according to the following formula to extract high-dimensional implicit characteristics of characteristic values of all positions in the input vector, wherein the formula is
Figure DEST_PATH_IMAGE029
Wherein
Figure 747223DEST_PATH_IMAGE030
Is the input vector of the input vector,
Figure 955350DEST_PATH_IMAGE032
is the output vector of the output vector,
Figure DEST_PATH_IMAGE033
is a matrix of the weights that is,
Figure 190022DEST_PATH_IMAGE034
is a vector of the offset to the offset,
Figure DEST_PATH_IMAGE035
representing a matrix multiplication. Then, using a one-dimensional convolution layer of the time sequence encoder to perform one-dimensional convolution encoding on the input vector by using the following formula so as to extract high-dimensional implicit association features of association among feature values of all positions in the input vector, wherein the formula is as follows:
Figure DEST_PATH_IMAGE037
wherein the content of the first and second substances,ais a convolution kernel inxA width in the direction,FIs a convolution kernel parameter vector,GIs a local vector matrix that operates with a convolution kernel,wthe size of the convolution kernel.
In step S130, first to third gaussian density maps of the first to third eigenvectors are constructed, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector. It should be understood that, considering that the first to third feature vectors each correspond to a feature distribution manifold in the high-dimensional feature space, and these feature distribution manifolds are due to their irregular shapes and scattering positions, if classification feature vectors are obtained by cascading only the first to third feature vectors, it would be equivalent to simply superimposing these feature distribution manifolds in the original positions and shapes, so that the boundaries of the newly obtained feature distribution manifolds become very irregular and complex, and when an optimal point is found by gradient descent, it is very easy to fall into a local extreme point and a global optimal point cannot be obtained. Therefore, it is necessary to further perform appropriate fusion on the first to third feature vectors so that the respective feature distributions can be converged on the profile with respect to each other, thereby improving the accuracy of subsequent classification.
That is, in the technical solution of the present application, for the first to third feature vectors, first to third gaussian density maps of the first to third feature vectors are constructed
Figure 922224DEST_PATH_IMAGE024
In which
Figure 925952DEST_PATH_IMAGE025
Is the desired vector, i.e. the feature vector itself, and
Figure 39401DEST_PATH_IMAGE027
is an autocovariance matrix, is the variance between the eigenvalues of each two positions of the eigenvector. It should be understood that, when the high-dimensional features of the convolutional neural network are extracted and fused at the high-dimensional feature level, based on the fact that the data distribution conforms to the gaussian distribution itself and the feature vectors have great difference in the feature domain, the feature fusion is performed in an iterative manner of mean and oblique square difference of the gaussian density map with respect to the characteristic that the gaussian density map can be used as the learning target of the convolutional neural network.
In step S140, the first to third gaussian density maps are fused based on an iterative manner of likelihood maximization to obtain a final gaussian density map, where the iterative manner of likelihood maximization represents that a mean vector of the fused gaussian density maps is a dot product between mean vectors of two gaussian density maps to be fused, and a covariance matrix of the fused gaussian density maps is a product of a covariance matrix of the first gaussian density map to be fused and a tensor of a covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fusedThe product of the two is multiplied by the transposition of the tensor product of the covariance matrix of the second gaussian density map to be fused and the mean vector of the second gaussian density map to be fused. That is, in the technical solution of the present application, in order to fuse the high-dimensional characteristic information of the temperature, the smoke, and the carbon monoxide at each time point and to integrate the three information to accurately determine and detect a home fire, it is further necessary to fuse the first to third gaussian density maps in an iterative manner based on likelihood maximization to obtain a final gaussian density map
Figure 230211DEST_PATH_IMAGE028
. It should be appreciated that by an iterative approach based on likelihood maximization, the target density map can be dynamically updated to modify feature domain shifts to help the convolutional neural network learn to a consistent feature representation and is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region.
Specifically, in this embodiment of the present application, a process of fusing the first to third gaussian density maps based on an iterative manner of likelihood maximization to obtain a final gaussian density map includes: fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization according to the following formula to obtain a final Gaussian density map;
wherein the formula is:
Figure 985678DEST_PATH_IMAGE012
Figure 329066DEST_PATH_IMAGE014
wherein
Figure 878996DEST_PATH_IMAGE015
It is shown that the dot-product,
Figure 557102DEST_PATH_IMAGE008
express tensor multiplication, and
Figure 850680DEST_PATH_IMAGE016
，
Figure 563421DEST_PATH_IMAGE017
。
in steps S150 and S160, gaussian discretizing the gaussian distribution of each position in the final gaussian density map to obtain a classification matrix, and passing the classification matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire alarm is generated. That is, in the technical solution of the present application, after a final gaussian density map fused with feature information is obtained, gaussian discretization is further performed on a gaussian distribution of each position in the final gaussian density map to obtain a classification matrix. Accordingly, in a specific example, first, the gaussian distribution of each position in the final gaussian density map is subjected to gaussian discretization to convert the gaussian distribution of each position into a one-dimensional vector; and then, two-dimensionally arranging the one-dimensional vectors corresponding to the Gaussian distribution of each position in the final Gaussian density map to obtain the classification matrix. Then, the classification matrix is passed through a classifier to obtain a classification result for indicating whether a fire early warning is generated, and particularly, a fire early warning prompt is issued in response to the classification result being that a fire early warning is generated.
Specifically, in an embodiment of the present application, the process of passing the classification matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether a fire alarm is generated includes: processing the classification matrix using the classifier to generate the classification result with the formula:
Figure 284252DEST_PATH_IMAGE018
in which
Figure 698922DEST_PATH_IMAGE019
Representing the projection of the classification matrix as a vector,
Figure 796191DEST_PATH_IMAGE020
to
Figure 97859DEST_PATH_IMAGE021
Is a weight matrix of the fully connected layers of each layer,
Figure 255171DEST_PATH_IMAGE022
to is that
Figure 173448DEST_PATH_IMAGE023
A bias matrix representing the layers of the fully connected layer.
In summary, the method for internet of things smart home based on the embodiment of the application is elucidated, and when the time sequence encoder is used for performing high-dimensional associated feature extraction on the measured values of the sensors at the plurality of predetermined time points and fusing the measured values at the high-dimensional feature level, as the data distribution per se conforms to gaussian distribution and the feature vectors have great difference in the feature domain, feature fusion is performed in an iterative manner of mean and oblique square difference of the gaussian density map with respect to the characteristic that the gaussian density map can be used as a learning target of the convolutional neural network. In this way, the target density map can be dynamically updated to modify the feature domain shifts to help the convolutional neural network learn to a consistent feature representation, and which is robust to scale variations, the response region can reflect the scale features of the fused features, and the covariance matrix is iteratively adjusted to fit the response region, again in an iterative manner based on likelihood maximization. Therefore, the fire disaster in the house can be accurately and timely warned.
Exemplary System
Fig. 4 illustrates a block diagram of an internet of things smart home system according to an embodiment of the application. As shown in fig. 4, the internet of things smart home system 400 according to the embodiment of the present application includes: a measurement value acquisition unit 410 for acquiring measurement values of a plurality of sensors disposed in a home at a plurality of predetermined time points, wherein the plurality of sensors include a temperature sensor, a smoke detector, and a carbon monoxide detector; an encoding unit 420, configured to arrange the measurement values of the sensors obtained by the measurement value obtaining units 410 at the multiple predetermined time points into input vectors respectively, and then obtain a first feature vector, a second feature vector, and a third feature vector through a time sequence encoder including a one-dimensional convolutional layer and a full connection layer, respectively; a gaussian density map constructing unit 430, configured to construct first to third gaussian density maps of the first to third eigenvectors obtained by the encoding unit 420, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector; a likelihood maximization fusion unit 440, configured to fuse the first to third gaussian density maps obtained by the gaussian density map construction unit 430 based on an iterative manner of likelihood maximization to obtain a final gaussian density map, where the iterative manner of likelihood maximization indicates that a mean vector of the fused gaussian density map is a dot product between mean vectors of two gaussian density maps to be fused, and a covariance matrix of the fused gaussian density map is a transpose of a product between a covariance matrix of the first gaussian density map to be fused and a tensor product of a covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused and a tensor product of the covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused; a gaussian discretization unit 450, configured to perform gaussian discretization on the gaussian distribution at each position in the final gaussian density map obtained by the likelihood maximization fusion unit 440 to obtain a classification matrix; and a classification unit 460 for passing the classification matrix obtained by the gaussian discretization unit 450 through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire early warning is generated.
In an example, in the internet of things smart home system 400, the encoding unit 420 includes: a full-connection layer feature extraction unit, configured to use a full-connection layer of the time-series encoder to encode the input vector with a formula to extract high-dimensional implicit features of feature values of each position in the input vector, where the formula is
Figure 293982DEST_PATH_IMAGE001
Wherein
Figure 715736DEST_PATH_IMAGE003
Is a function of the input vector or vectors,
Figure 778370DEST_PATH_IMAGE004
is the output vector of the digital video signal,
Figure 918365DEST_PATH_IMAGE005
is a matrix of the weights that is,
Figure 357436DEST_PATH_IMAGE007
is a vector of the offset to the offset,
Figure 633697DEST_PATH_IMAGE008
represents a matrix multiplication; and a one-dimensional convolutional layer feature extraction unit, configured to perform one-dimensional convolutional coding on the input vector by using a one-dimensional convolutional layer of the time-series encoder according to the following formula to extract a high-dimensional implicit correlation feature of a correlation between feature values of each position in the input vector, where the formula is:
Figure 382079DEST_PATH_IMAGE010
wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernel inxA width in the direction,FIs a convolution kernel parameter vector,GOperated on convolution kernelsA matrix of local vectors is formed by a matrix of local vectors,wthe size of the convolution kernel.
In an example, in the internet of things smart home system 400, the likelihood maximization fusion unit 440 is further configured to: fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization to obtain a final Gaussian density map;
wherein the formula is:
Figure 743790DEST_PATH_IMAGE012
Figure 720973DEST_PATH_IMAGE014
wherein
Figure 117320DEST_PATH_IMAGE015
Which represents a dot product of the light beam,
Figure 787335DEST_PATH_IMAGE008
express tensor multiplication, and
Figure 652654DEST_PATH_IMAGE016
，
Figure 167949DEST_PATH_IMAGE017
。
in an example, in the internet of things smart home system 400, the gaussian discretization unit 450 is further configured to: performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to convert the Gaussian distribution of each position into a one-dimensional vector; and two-dimensionally arranging the one-dimensional vectors corresponding to the Gaussian distribution of each position in the final Gaussian density map to obtain the classification matrix.
In an example, in the foregoing internet of things smart home system 400, the classifying unit 460 is further configured to: processing the classification matrix using the classifier with the following formulaTo generate the classification result, wherein the formula is:
Figure 418802DEST_PATH_IMAGE018
in which
Figure 259719DEST_PATH_IMAGE019
Representing the projection of the classification matrix as a vector,
Figure 596022DEST_PATH_IMAGE020
to is that
Figure 180588DEST_PATH_IMAGE021
Is a weight matrix of the fully connected layers of each layer,
Figure 535214DEST_PATH_IMAGE022
to is that
Figure 547033DEST_PATH_IMAGE023
A bias matrix representing the layers of the fully connected layer.
In an example, in the foregoing internet of things smart home system 400, the internet of things smart home method further includes sending a fire warning prompt in response to the classification result being that a fire warning is generated.
Here, it can be understood by those skilled in the art that the specific functions and operations of the respective units and modules in the internet of things smart home system 400 have been described in detail in the above description of the internet of things smart home method with reference to fig. 1 to 3, and thus, a repeated description thereof will be omitted.
As described above, the internet of things smart home system 400 according to the embodiment of the present application may be implemented in various terminal devices, for example, a server of an internet of things smart home algorithm. In an example, the internet of things smart home system 400 according to the embodiment of the present application may be integrated into a terminal device as a software module and/or a hardware module. For example, the internet of things smart home system 400 may be a software module in an operating system of the terminal device, or may be an application developed for the terminal device; of course, the internet of things smart home system 400 may also be one of many hardware modules of the terminal device.
Alternatively, in another example, the internet of things smart home system 400 and the terminal device may also be separate devices, and the internet of things smart home system 400 may be connected to the terminal device through a wired and/or wireless network and transmit the interaction information according to an agreed data format.
Exemplary electronic device
Next, an electronic apparatus according to an embodiment of the present application is described with reference to fig. 5. As shown in fig. 5, the electronic device 10 includes one or more processors 11 and memory 12. The processor 11 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.
The memory 12 may include at least one computer program product that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, read Only Memory (ROM), hard disk, flash memory, etc. At least one computer program instruction may be stored on the computer readable storage medium, and the processor 11 may execute the program instruction to implement the internet of things smart home method of the various embodiments of the present application described above and/or other desired functions. Various contents such as the first gaussian density map, the classification matrix, etc. may also be stored in the computer-readable storage medium.
In one example, the electronic device 10 may further include: an input device 13 and an output device 14, which are interconnected by a bus system and/or other form of connection mechanism (not shown).
The input device 13 may include, for example, a keyboard, a mouse, and the like.
The output device 14 can output various information including the classification result to the outside. The output devices 14 may include, for example, a display, speakers, a printer, and a communication network and its connected remote output devices, among others.
Of course, for simplicity, only some of the components of the electronic device 10 relevant to the present application are shown in fig. 5, and components such as buses, input/output interfaces, and the like are omitted. In addition, the electronic device 10 may include any other suitable components depending on the particular application.
Exemplary computer program product and computer-readable storage Medium
In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the internet of things smart home method according to various embodiments of the present application described in the "exemplary methods" section of this specification, above.
The computer program product may be written with program code for performing the operations of embodiments of the present application in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the steps in the internet of things smart home method according to various embodiments of the present application described in the "exemplary methods" section above in this specification.
The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having at least one wire, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. An Internet of things smart home method is characterized by comprising the following steps:
obtaining measurements at a plurality of predetermined points in time from a plurality of sensors deployed in a home, wherein the plurality of sensors comprises a temperature sensor, a smoke detector, and a carbon monoxide detector;
respectively arranging the measured values of the sensors at the plurality of preset time points into input vectors, and then respectively passing through a time sequence encoder comprising a one-dimensional convolutional layer and a full-connection layer to obtain a first feature vector, a second feature vector and a third feature vector;
constructing first to third gaussian density maps of the first to third eigenvectors, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector;
fusing the first to third Gaussian density maps based on an iterative manner of likelihood maximization to obtain a final Gaussian density map, wherein the iterative manner of likelihood maximization represents that a mean vector of the fused Gaussian density map is a point product between mean vectors of two Gaussian density maps to be fused, and a covariance matrix of the fused Gaussian density map is a product between a covariance matrix of the first Gaussian density map to be fused and a covariance matrix of the second Gaussian density map to be fused and a tensor product of the mean vector of the second Gaussian density map to be fused, and then is multiplied by a transpose of the product of the covariance matrix of the second Gaussian density map to be fused and the tensor product of the mean vector of the second Gaussian density map to be fused;
performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map to obtain a classification matrix; and
and passing the classification matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a fire early warning is generated or not.
2. The internet-of-things smart home method according to claim 1, wherein the step of arranging the measured values of the sensors at the predetermined time points into input vectors and then obtaining a first feature vector, a second feature vector and a third feature vector by a time sequence encoder comprising a one-dimensional convolutional layer and a full connection layer comprises the steps of:
respectively encoding the input vector by using a full connection layer of the time sequence encoder according to the following formula to extract high-dimensional implicit characteristics of characteristic values of all positions in the input vector, wherein the formula is
Figure 465180DEST_PATH_IMAGE001
Wherein
Figure 587856DEST_PATH_IMAGE002
Is the input vector of the input vector,
Figure 593727DEST_PATH_IMAGE003
is an outputThe vector of the vector is then calculated,
Figure 42026DEST_PATH_IMAGE004
is a matrix of the weights that is,
Figure 352922DEST_PATH_IMAGE005
is a vector of the offset to be used,
Figure 30022DEST_PATH_IMAGE006
represents a matrix multiplication; and
performing one-dimensional convolutional encoding on the input vector by using a one-dimensional convolutional layer of the time sequence encoder according to the following formula to extract high-dimensional implicit associated features of association among feature values of all positions in the input vector, wherein the formula is as follows:
Figure 375553DEST_PATH_IMAGE007
wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernelxWidth in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wthe size of the convolution kernel.
3. The internet of things smart home method of claim 2, wherein fusing the first to third gaussian density maps based on an iterative manner of likelihood maximization to obtain a final gaussian density map comprises:
fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization to obtain a final Gaussian density map;
wherein the formula is:
Figure 729174DEST_PATH_IMAGE008
Figure 527365DEST_PATH_IMAGE009
wherein
Figure 506691DEST_PATH_IMAGE010
It is shown that the dot-product,
Figure 706729DEST_PATH_IMAGE006
express tensor multiplication, and
Figure 965672DEST_PATH_IMAGE011
，
Figure 251159DEST_PATH_IMAGE012
。
4. the Internet of things smart home method according to claim 3, wherein the Gaussian discretization of the Gaussian distribution of each position in the final Gaussian density map to obtain a classification matrix comprises:
performing Gaussian discretization on the Gaussian distribution of each position in the final Gaussian density map so as to convert the Gaussian distribution of each position into a one-dimensional vector; and
and two-dimensionally arranging one-dimensional vectors corresponding to the Gaussian distribution of each position in the final Gaussian density map to obtain the classification matrix.
5. The Internet of things smart home method according to claim 4, wherein the classifying matrix is passed through a classifier to obtain a classification result, and the classification result is used for indicating whether a fire early warning is generated or not, and comprises the following steps:
processing the classification matrix using the classifier to generate the classification result with the formula:
Figure 535641DEST_PATH_IMAGE013
wherein
Figure 324606DEST_PATH_IMAGE014
Representing the projection of the classification matrix as a vector,
Figure 20029DEST_PATH_IMAGE015
to
Figure 792813DEST_PATH_IMAGE016
Is a weight matrix of the fully connected layers of each layer,
Figure 864675DEST_PATH_IMAGE017
to
Figure 22992DEST_PATH_IMAGE018
A bias matrix representing the layers of the fully connected layer.
6. The Internet of things smart home method according to claim 5, further comprising sending a fire early warning prompt in response to the classification result being that a fire early warning is generated.
7. The utility model provides an thing networking smart home systems which characterized in that includes:
a measurement value acquisition unit for acquiring measurement values of a plurality of sensors deployed in a home at a plurality of predetermined time points, wherein the plurality of sensors include a temperature sensor, a smoke detector and a carbon monoxide detector;
the encoding unit is used for respectively arranging the measured values of the sensors obtained by the measured value obtaining unit at the plurality of preset time points into input vectors and then respectively passing through a time sequence encoder comprising a one-dimensional convolution layer and a full connection layer to obtain a first characteristic vector, a second characteristic vector and a third characteristic vector;
a gaussian density map construction unit configured to construct first to third gaussian density maps of the first to third eigenvectors obtained by the encoding unit, wherein a mean vector of the first gaussian density map is the first eigenvector, a value of each position in a covariance matrix of the first gaussian density map is a variance between eigenvalues of the corresponding two positions in the first eigenvector, a mean vector of the second gaussian density map is the second eigenvector, a value of each position in a covariance matrix of the second gaussian density map is a variance between eigenvalues of the corresponding two positions in the second eigenvector, a mean vector of the third gaussian density map is the third eigenvector, and a value of each position in a covariance matrix of the third gaussian density map is a variance between eigenvalues of the corresponding two positions in the third eigenvector;
a likelihood maximization fusion unit, configured to fuse the first to third gaussian density maps obtained by the gaussian density map construction unit based on an iterative manner of likelihood maximization to obtain a final gaussian density map, where the iterative manner of likelihood maximization indicates that a mean vector of the fused gaussian density map is a dot product between mean vectors of two gaussian density maps to be fused, and a covariance matrix of the fused gaussian density map is a transpose of a product between a covariance matrix of the first gaussian density map to be fused and a tensor product of a covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused and a tensor product of the covariance matrix of the second gaussian density map to be fused and a mean vector of the second gaussian density map to be fused;
a gaussian discretization unit for performing gaussian discretization on the gaussian distribution of each position in the final gaussian density map obtained by the likelihood maximization fusion unit to obtain a classification matrix; and
and the classification unit is used for enabling the classification matrix obtained by the Gaussian discretization unit to pass through a classifier so as to obtain a classification result, and the classification result is used for indicating whether fire early warning is generated or not.
8. The internet of things smart home system of claim 7, wherein the encoding unit comprises:
a full connection layer feature extraction unit for using the time sequence encoderThe full-connection layer respectively encodes the input vector by the following formula to extract high-dimensional implicit characteristics of characteristic values of all positions in the input vector, wherein the formula is
Figure 154896DEST_PATH_IMAGE001
In which
Figure 149397DEST_PATH_IMAGE002
Is the input vector of the input vector,
Figure 759370DEST_PATH_IMAGE003
is the output vector of the output vector,
Figure 257348DEST_PATH_IMAGE004
is a matrix of the weights that is,
Figure 310885DEST_PATH_IMAGE005
is a vector of the offset to the offset,
Figure 792682DEST_PATH_IMAGE006
represents a matrix multiplication; and
a one-dimensional convolutional layer feature extraction unit, configured to perform one-dimensional convolutional coding on the input vector by using a one-dimensional convolutional layer of the time-series encoder according to the following formula to extract a high-dimensional implicit correlation feature of a correlation between feature values of each position in the input vector, where the formula is:
Figure 940767DEST_PATH_IMAGE007
wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernelxA width in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wthe size of the convolution kernel.
9. The internet of things smart home system of claim 7, wherein the likelihood maximization fusion unit is further configured to:
fusing the first to third Gaussian density maps in an iterative manner based on likelihood maximization according to the following formula to obtain a final Gaussian density map;
wherein the formula is:
Figure 824409DEST_PATH_IMAGE008
Figure 32537DEST_PATH_IMAGE009
wherein
Figure 250897DEST_PATH_IMAGE010
Which represents a dot product of the light beam,
Figure 202673DEST_PATH_IMAGE006
express tensor multiplication, and
Figure 675242DEST_PATH_IMAGE011
，
Figure 319850DEST_PATH_IMAGE012
。
10. an electronic device, comprising:
a processor; and
a memory having stored therein computer program instructions that, when executed by the processor, cause the processor to perform the internet of things smart home method of any of claims 1-6.
CN202210222195.9A 2022-03-09 2022-03-09 Internet of things smart home method and system and electronic equipment Pending CN115171317A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210222195.9A CN115171317A (en) 2022-03-09 2022-03-09 Internet of things smart home method and system and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210222195.9A CN115171317A (en) 2022-03-09 2022-03-09 Internet of things smart home method and system and electronic equipment

Publications (1)

Publication Number Publication Date
CN115171317A true CN115171317A (en) 2022-10-11

Family

ID=83484016

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210222195.9A Pending CN115171317A (en) 2022-03-09 2022-03-09 Internet of things smart home method and system and electronic equipment

Country Status (1)

Country Link
CN (1) CN115171317A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115861246A (en) * 2022-12-09 2023-03-28 马鞍山远昂科技有限公司 Product quality abnormity detection method and system applied to industrial Internet
CN115905980A (en) * 2022-12-09 2023-04-04 马鞍山玖帆科技有限公司 Equipment monitoring system for wind power energy and monitoring method thereof
CN115935261A (en) * 2022-12-30 2023-04-07 马鞍山思凯科技有限公司 Group equipment non-absolute forward feedback method based on industrial Internet
CN116258947A (en) * 2023-03-07 2023-06-13 浙江研几网络科技股份有限公司 Industrial automatic processing method and system suitable for home customization industry
CN116321620A (en) * 2023-05-11 2023-06-23 杭州行至云起科技有限公司 Intelligent lighting switch control system and method thereof
CN116820052A (en) * 2023-07-13 2023-09-29 滁州优胜高分子材料有限公司 PBT material production equipment and control method thereof

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115861246A (en) * 2022-12-09 2023-03-28 马鞍山远昂科技有限公司 Product quality abnormity detection method and system applied to industrial Internet
CN115905980A (en) * 2022-12-09 2023-04-04 马鞍山玖帆科技有限公司 Equipment monitoring system for wind power energy and monitoring method thereof
CN115861246B (en) * 2022-12-09 2024-02-27 唐山旭华智能科技有限公司 Product quality abnormality detection method and system applied to industrial Internet
CN115935261A (en) * 2022-12-30 2023-04-07 马鞍山思凯科技有限公司 Group equipment non-absolute forward feedback method based on industrial Internet
CN116258947A (en) * 2023-03-07 2023-06-13 浙江研几网络科技股份有限公司 Industrial automatic processing method and system suitable for home customization industry
CN116258947B (en) * 2023-03-07 2023-08-18 浙江研几网络科技股份有限公司 Industrial automatic processing method and system suitable for home customization industry
CN116321620A (en) * 2023-05-11 2023-06-23 杭州行至云起科技有限公司 Intelligent lighting switch control system and method thereof
CN116321620B (en) * 2023-05-11 2023-08-11 杭州行至云起科技有限公司 Intelligent lighting switch control system and method thereof
CN116820052A (en) * 2023-07-13 2023-09-29 滁州优胜高分子材料有限公司 PBT material production equipment and control method thereof
CN116820052B (en) * 2023-07-13 2024-02-02 滁州优胜高分子材料有限公司 PBT material production equipment and control method thereof

Similar Documents

Publication Publication Date Title
CN115171317A (en) Internet of things smart home method and system and electronic equipment
CN114724386B (en) Short-time traffic flow prediction method and system under intelligent traffic and electronic equipment
CN114548246A (en) Control method and system of intelligent device and electronic device
CN115620303B (en) Personnel file intelligent management system
CN116189865A (en) Hospital reservation registration management system
CN116247824B (en) Control method and system for power equipment
CN116015837A (en) Intrusion detection method and system for computer network information security
CN115783923A (en) Elevator fault mode identification system based on big data
CN114092972A (en) Intelligent peep-proof electronic equipment display method and system and electronic equipment
CN115392320A (en) Transformer with anti-theft function and method thereof
CN116245513A (en) Automatic operation and maintenance system and method based on rule base
CN115324843A (en) Wind generating set fault diagnosis system and method based on monitoring data
CN115471216A (en) Data management method of intelligent laboratory management platform
CN116866054A (en) Public information safety monitoring system and method thereof
CN115128663A (en) Mine safety production detection system based on Internet of things and detection method thereof
CN115374822A (en) Fault diagnosis method and system based on multi-level feature fusion
CN115146676A (en) Circuit fault detection method and system
CN116055293A (en) Remote fault monitoring method of router and router
CN116975728B (en) Safety management method and system for coal bed methane drilling engineering
CN113837302B (en) City safety management monitoring method and system based on Internet of things and electronic equipment
CN116467485B (en) Video image retrieval construction system and method thereof
CN112800217A (en) Vector relevance matrix-based intelligent assessment method for government affair transaction processing report
CN116698283A (en) Intelligent monitoring system and method for load gravity center deviation
CN115861969A (en) Intelligent driving vision system and method based on artificial intelligence
CN115510900A (en) Automatic circuit fault diagnosis method and system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination