CN103389719A - Intelligent home monitoring system and method based on cloud computing - Google Patents

Abstract

The invention provides an intelligent home monitoring system and method based on cloud computing. The system comprises a compound sensor, centralized device and a cloud server, wherein the compound sensor comprise at least one sensor and is connected with various monitored equipment or located in a monitored area; an output end of the compound sensor is connected with an input end of the centralized device; and an output end of the centralized device is connected with an input end of the cloud server. By the aid of the compound sensor, various types of monitoring information in the monitored area can be monitored comprehensively, the system cost is saved, wiring is simplified, and the system stability is also improved.

Description

Intelligent home monitoring system and method based on cloud computing

Technical field

The invention belongs to communication technical field, be specifically related to a kind of Intelligent home monitoring system based on cloud computing and method.

Background technology

Smart Home is take house as platform, utilize comprehensive wiring technology, the network communications technology, automatic control technology and audio frequency and video technology etc. that the facility relevant with life staying idle at home is integrated, thereby build the management system of efficient house facility and family's schedule affairs, promote house security, convenience, comfortableness, artistry, and realize the living environment of environmental protection and energy saving.

The development of the Smart Home technology history of existing more than 10 years.Smart Home has adopted multiple sensors and network structure.But, in prior art, in monitoring of environmental, need to arrange a plurality of independent sensors, each sensor detects respectively different facility state parameters, has that integrated level is low, somewhat expensive, wiring complexity and a problem such as stability is limited.

Summary of the invention

Defect for the prior art existence, the invention provides a kind of Intelligent home monitoring system based on cloud computing and method, by compound sensor can the overall monitor monitoring area the multiclass monitor message, namely saved system cost, simplify wiring, also improved system stability.

The technical solution used in the present invention is as follows:

The invention provides a kind of Intelligent home monitoring system based on cloud computing, comprising: compound sensor, concentrating equipment and cloud server; Described compound sensor comprises at least a sensor, described sensor is connected with all kinds of monitored devices or is positioned at monitoring area, the output terminal of described compound sensor is connected with the input end of described concentrating equipment, and the output terminal of described concentrating equipment is connected with the input end of described cloud server.

Preferably, described monitored device comprises one or more in gas-cooker, TV, refrigerator, air-conditioning, lamp and water heater.

Preferably, described compound sensor comprises processor, optical sensor, gas sensor, humidity sensor, temperature sensing, proximity transducer, infrared sensor and communication interface; Described processor is connected with described optical sensor, described gas sensor, described humidity sensor, described temperature sensing, described proximity transducer, described infrared sensor and the communication interface of being connected respectively.

Preferably, described communication interface is bluetooth communication interface or zigbee communication interface.

Preferably, described concentrating equipment comprises arm processor, wifi communication interface, zigbee communication interface, rs485 communication interface, rs232 communication interface and usb communication interface; Described arm processor is connected with described wifi communication interface, described zigbee communication interface, described rs485 communication interface, described rs232 communication interface and the usb communication interface of being connected respectively.

Preferably, described cloud server comprises that database, individual/family data manager, data course manager, Authorization Manager, data sensitive degree check that manager, sensitive data analysis management device, equipment manager, User Manager, log manager, metering and accounting management device, security manager, service managerZ-HU and service provide manager;

Described User Manager is used for supervisor authority user and relevant file thereof; Wherein, described file comprise individual subscriber file, station address file and with user-dependent telephone number file;

Described individual/family data manager is for the protection of the individual subscriber file of the described authorized user of described user management management;

Described Authorization Manager is used for to the designated user granted rights, makes described designated user access or shares other users' data;

Described data course manager is used for monitoring the activity of the storage data of described individual/family data manager mandate, specifically comprises: access historical record in described database the access time point, access content and the corresponding relation of executable operations;

Described service managerZ-HU is used for to user's Push Service catalogue, and described service catalogue is enumerated service content that described Intelligent home monitoring system provides to the user and the pay imformation corresponding with described service content;

Described service provides manager to be used for the access of automatic management user to described cloud server;

Described metering and accounting management device are used for calculating the expense situation of collecting to each authorized user;

Described log manager is used for recording described cloud server event or movable content and corresponding time point;

Described equipment manager is used for recording the operation information of each authorized user monitored device;

Described data sensitive degree checks that manager is used for checking user's request service information, the memory location of service data in database of judgement user request;

Described sensitive data analysis management device is used for, based on neural network, the monitor message that receives is carried out analysis-by-synthesis, by reference to the same tutorial historical data of storing in database, at first infer people's behavioural habits, then judge whether described monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message;

Described security manager is used to service that authentication, mandate and key management are provided.

The present invention also provides a kind of method for supervising of applying above-mentioned Intelligent home monitoring system based on cloud computing, comprises the following steps:

S1, install respectively described compound sensor in each monitoring area; Wherein, described compound sensor comprises at least a sensor, and described sensor is connected with all kinds of monitored devices or is positioned at monitoring area;

S2, described compound sensor acquisition monitoring information, then send to described concentrating equipment with described monitor message; Wherein, described monitor message comprises the equipment operational factor of all kinds of monitored devices that are connected with self and the environmental information of monitoring area;

S3, described concentrating equipment receives described monitor message, and described monitor message is encrypted, and obtains encrypting monitor message; Then described encryption monitor message is sent to described cloud server;

S4, the described encryption monitor message that described cloud server deciphering receives, reduction obtains described monitor message, and described monitor message is stored in database;

S5, described cloud server employing Analysis of Neural Network Method is stored in the data in database, judges whether described monitor message is abnormal,, if abnormal, to the cell-phone number corresponding with described monitoring area of reserving, sends prompting message.

Preferably, S5 is specially:

Described cloud server employing Analysis of Neural Network Method is stored in the historical data in database, at first infer people's behavioural habits, then judge whether described monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message.

Preferably, S5 is specially:

Described cloud server uses neural network learning people's daily behavior custom, adopts the Multilayer Perception backpropagation train and revise, and sets up the behavioural habits neural network model;

Described cloud server is according to the described behavioural habits neural network model of setting up, prediction and analyst's future behaviour.

Preferably, described cloud server uses neural network learning people's daily behavior custom, adopts the Multilayer Perception backpropagation train and revise, and sets up the behavioural habits neural network model and is specially:

S51, the topological structure of establishment feedforward neural network, described feedforward neural network comprises input layer, hidden layer and output layer; Described input layer has N input variable, and described hidden layer has M=N, N+1, N+2 ... individual hidden unit; Described output layer has P output variable;

Wherein, input variable comprise building room number, floor number, dormitory number, room number and when this room to room in the operation information carried out of application-specific;

Output variable is: when application-specific is abnormal in drawing room, to the cellphone subscriber corresponding with this room, send abnormal reminder message;

S52, use historical data as the sample data of neural metwork training, with back-propagation algorithm, neural network trained, and obtains described behavioural habits neural network model; Wherein, with back-propagation algorithm, neural network is trained and is specially: follow these steps to constantly update each layer connection weight value:

S521, employing formula 1 initialization all-network connection weight value;

$W_{i,j}^1;i=\stackrel{\‾}{1,N};j=\stackrel{\‾}{1,M}$ $W_{i,j}^2;i=\stackrel{\‾}{1,M}-$ -------------------------------------Shi 1

This weighted value is [N/2, N/2] interval interior random digit;

S522, if

Carry out S523-S526;

Wherein: E (W) represents error function;

t _kThe expression desired value;

o _kThe expression output valve;

T represents threshold values;

S523, to network input example

And by formula 2, calculate output

$\stackrel{\‾}{O}=\stackrel{\‾}{X}\·\stackrel{\‾}{W^1}\·\stackrel{\‾}{W^2}.$ -------------------------------------Shi 2

Wherein:

Expression: the sample of input;

Expression: weight;

Expression: weight;

S524,, for each network output unit k, during k=1, calculate its error term δ by formula 3 _k

δ _k=o _k(1-o _k) (t _k-o _k)----------------------------------Shi 3

S525, for each hidden unit H, H=1, M, calculate its error delta by formula 4 _h

${\mathrm{\δ}}_h=o_h(1-o_h){\mathrm{\Σ}}_{k\∈\mathrm{outputs}\left(P\right)}{W}_{k,h.}^2{\mathrm{\δ}}_k.$ --------------------------------formula 4

Wherein,

K the weight of expression output valve h;

S526, by the connection weight value W of formula 5, formula 6 each network of renewal _{I, j}

W _{I, j}=W _{I, j}+ Δ W _{I, j}--------------------------------formula 5

Δ W _{I, j}=α. δ ₁X _{I, j}--------------------------------formula 6

Wherein, α represents the step number of learning.

Beneficial effect of the present invention is as follows:

(1) cross the multiclass monitor message that compound sensor can the overall monitor monitoring area, namely saved system cost, simplified wiring, also improved system stability.

(2) concentrating equipment has data encryption feature, thereby has improved the security of supervisory system;

(3) cloud server adopts a plurality of assemblies, has ensured the security that is stored in the cloud server data;

(4) cloud server adopts nerual network technique analyst's behavioural habits, and when the behavioural habits of finding Monitoring Data and people are not inconsistent, to associated user's mobile phone, sends prompting message, has promoted user's experience.

Description of drawings

Fig. 1 is the structural representation of the Intelligent home monitoring system based on cloud computing provided by the invention;

Fig. 2 is the structural representation of compound sensor provided by the invention;

Fig. 3 is the structural representation of cloud server provided by the invention;

Fig. 4 is the structural representation of neural network model provided by the invention.

Embodiment

The present invention is described in detail below in conjunction with accompanying drawing:

As shown in Figure 1, the invention provides a kind of Intelligent home monitoring system based on cloud computing, comprising: compound sensor, concentrating equipment and cloud server; Compound sensor comprises at least a sensor, and sensor is connected with all kinds of monitored devices or is positioned at monitoring area, and the output terminal of compound sensor is connected with the input end of concentrating equipment, and the output terminal of concentrating equipment is connected with the input end of cloud server.Wherein, monitored device comprises one or more in gas-cooker, TV, refrigerator, air-conditioning, lamp and water heater.

Apply the method for supervising of this Intelligent home monitoring system, comprise the following steps:

S1, install respectively compound sensor in each monitoring area; Wherein, compound sensor comprises at least a sensor, and sensor is connected with all kinds of monitored devices or is positioned at monitoring area;

S2, compound sensor acquisition monitoring information, then send to concentrating equipment with monitor message; Wherein, monitor message comprises the equipment operational factor of all kinds of monitored devices that are connected with self and the environmental information of monitoring area;

S3, concentrating equipment receives monitor message, and monitor message is encrypted, and obtains encrypting monitor message; Then will encrypt monitor message and send to cloud server;

S4, the encryption monitor message that the cloud server deciphering receives, reduction obtains monitor message, and monitor message is stored in database;

S5, cloud server employing Analysis of Neural Network Method is stored in the data in database, judges whether monitor message is abnormal,, if abnormal, to the cell-phone number corresponding with monitoring area of reserving, sends prompting message.

Concrete, cloud server employing Analysis of Neural Network Method is stored in the historical data in database, at first infers people's behavioural habits, judges then whether monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message.Follow-uply in the detailed process of analyst's behavioural habits introduce in detail while introducing cloud server, do not repeat them here.

Below compound sensor provided by the invention, concentrating equipment and cloud server are introduced respectively in detail:

(1) compound sensor

Compound sensor is the sensor that comprises multiple assembly, specifically can comprise processor, optical sensor, gas sensor, humidity sensor, temperature sensing, proximity transducer, infrared sensor, communication interface and; Processor respectively with optical sensor, gas sensor, humidity sensor, temperature sensing, proximity transducer, infrared sensor, communication interface.Wherein, communication interface includes but not limited to bluetooth communication interface or zigbee communication interface.

Compound sensor and all kinds of domestic. applications equipment connection, such as: lamp, TV, gas-cooker, water heater etc.Compound sensor detects data from user's various application apparatuss.The data that then will detect send to concentrating equipment.

(2) concentrating equipment

The system that the most use of present gateway has existed, in the present invention, for improving the security of data transmission, concentrating equipment has data encryption feature, at first the data from the compound sensor transmission is encrypted, and then encrypted data transmission is arrived cloud server.

Concrete, concentrating equipment comprises arm processor, wifi communication interface, zigbee communication interface, rs485 communication interface, rs232 communication interface and usb communication interface; Described arm processor is connected with described wifi communication interface, described zigbee communication interface, described rs485 communication interface, described rs232 communication interface and the usb communication interface of being connected respectively.

(3) cloud server

Cloud server adopts cloud computing technology, shares the Various types of data resource on the network.Due in cloud computing environment, a large amount of personal data of cloud server storage, therefore, need to take effective measures, and ensures the data security of cloud server, prevents that personal data are stolen by others or abuse.Cloud computing refers to uses remote computation resource (hardware and software) to realize large data calculating.Name derives from cloud type symbol and represents the system infrastructure that cloud computing is compound.Cloud computing offers the long-range data of user, software and calculation services.

In order to protect secure user data and to be user's service of offering the best, cloud server has adopted multiple assembly, concrete, cloud server comprises that database, individual/family data manager, data course manager, Authorization Manager, data sensitive degree check that manager, sensitive data analysis management device, equipment manager, User Manager, log manager, metering and accounting management device, security manager, service managerZ-HU and service provide manager;

User Manager is used for supervisor authority user and relevant file thereof; Wherein, described file comprise individual subscriber file, station address file and with user-dependent telephone number file;

Described individual/family data manager is for the protection of the individual subscriber file of the described authorized user of described user management management;

Described Authorization Manager is used for to the designated user granted rights, makes described designated user access or shares other users' data;

Described data course manager is used for monitoring the activity of the storage data of described individual/family data manager mandate, specifically comprises: access historical record in described database the access time point, access content and the corresponding relation of executable operations;

Described service managerZ-HU is used for to user's Push Service catalogue, and described service catalogue is enumerated service content that described Intelligent home monitoring system provides to the user and the pay imformation corresponding with described service content;

Described service provides manager to be used for the access of automatic management user to described cloud server;

Described metering and accounting management device are used for calculating the expense situation of collecting to each authorized user;

Described log manager is used for recording described cloud server event or movable content and corresponding time point;

Described equipment manager is used for recording the operation information of each authorized user monitored device;

Described security manager is used to service that authentication, mandate and key management are provided, and uses Single Sign-On Technology Used to provide convenience as the user.According to the difference of service content, key management needs different cryptographic systems.

The data sensitive degree checks that manager is an assembly, is used for checking user's request service information, and the memory location of service data in database of judgement user request, for example, store the table position of this service data etc.For example, there are a large amount of building data to be stored in high in the clouds, cross Internet of Things when user's request service information exchange and encrypt terminal while sending to high in the clouds, this component inspection user request service information from building position, room number etc., then judge whether to exist the list item of this building position of storage and room number.

Described sensitive data analysis management device is used for, based on neural network, the monitor message that receives is carried out analysis-by-synthesis, by reference to the same tutorial historical data of storing in database, at first infer people's behavioural habits, then judge whether described monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message.

Below introduce in detail sensitive data analysis management device and the monitor message that receives carried out the process of analysis-by-synthesis:

The cloud server reception, from the enciphered data of concentrating equipment transmission, after the enciphered data deciphering, is adopted Analysis of Neural Network Method, and the data of analysis comprise: the data relevant to lamp, pressure data, the data of with TV, being correlated with, temperature data etc.For example: pressure transducer detects gas tank top hole pressure information in kitchen, and the pressure information that then will detect sends to high in the clouds; Whether high in the clouds, by pressure information is analyzed, judges this pressure information lower than threshold value, if so, draws the conclusion that gas tank leaks, and to the user mobile phone that uses this gas tank, sends alarm message.Same,, by the analytic process in compound sensor and high in the clouds, can also monitor user ' whether fall, the room information such as breaking out of fire whether.

In addition, sensitive data analysis management device carries out analysis-by-synthesis based on neural network to the monitor message that receives, and infers people's behavioural habits, judges that then whether monitor message meets people's behavioural habits, is exemplified below:

Compound sensor is all monitored the lighting time of lamp A in room A every day and is turned off the light the time, then the lighting time of lamp A and the time of turning off the light is uploaded to high in the clouds; For example, cloud server received respectively following information in 5 days: first day, lighting time are 6:10 in morning, and the time of turning off the light is 9:05 in evening; Second day, lighting time are 5:52 in morning, and the time of turning off the light is 9:08 in evening; The 3rd day, lighting time was 5:55 in morning, and the time of turning off the light is 8:56 in evening; The 4th day, lighting time was 6:09 in morning, and the time of turning off the light is 8:55 in evening; The 5th day, lighting time was 6:04 in morning, and the time of turning off the light is 8:56 in evening; Cloud server uses the data that the nerual network technique analysis receives, and draws the behavioural habits information with the servant: depart from 6:00 AM and carried out the operation of turning on light in scope in 10 minutes; Depart from scope in 9: 10 minutes evenings and carry out the operation of turning off the light.Work as some day, high in the clouds does not receive the turning off the light during monitor message of compound sensor transmission 9: 10 minutes evenings departing from scope, and high in the clouds sends the reminder message of " please turn off the light " to this user mobile phone.

One large advantage of neural network is can be according to the sample self-teaching.The method that tradition is dealt with problems, need to elaborate the model that uses, and then provides the algorithm of a succession of Solve problems.But the problem that classic method exists is mainly: it is very difficult or even impossible setting up a definite algorithm.

In neural network model, do not need directly to provide the algorithm that solves a certain problem, only need to carry out training study for neural network provides some samples, neural network is acquired information from training sample, and oneself just can set up the model of dealing with problems.The tremendous potential of neural network, make and use various dimensions sample solving practical problems to have boundless space.

There is certain behavior pattern usually in the mankind, yet the mankind interrupt sometimes and change this behavior pattern even fully.The object of the invention is: use neural network, study people's daily behavior custom, set up the behavioural habits neural network model.Then according to the behavioural habits neural network model of setting up, the behavior in prediction and analyst's future.Adopt the Multilayer Perception backpropagation to train and correction model.

In the present invention, as shown in Figure 4, be the structural representation of neural network model provided by the invention, select to have Multilayer Perception and the back propagation learning algorithm of a hidden layer.Room number and various domestic. applications adopt binary coding to facilitate storage, and preferred select time complexity is log ₂The encoding by bit of N (N is the entry in neural network), thus be conducive to later expansion.

For example, suppose that a certain residential quarter has 25 buildings, the building room number uses B1, B2..., B25 coding.5 layers, each building, floor number use I, II, III, IV, V coding.Every floor has two dormitories, and dormitory number is encoded with D1 and D2.Each dormitory has 4 rooms, and room number is encoded with R1, R2, R3, R4.

Compound sensor is placed in each dormitory, and image data in concentrating equipment, and sends to cloud server with data.

Input layer: the input data comprise building number, floor number, dormitory number, room number and when this room to room in the operation of a certain application.

Illustrate, sometime, No. 23 building, 3rd floors, in the 4th room of dormitory 1, someone has carried out a certain operation, as turns on light.

<23><III><D1><R4><00001><19.00><ON>

The present invention selects 5 bits to encode, and obtains

10111.00011.00001.00100.00001.10011.00000.00001

Start, use sample to train system, then according to the data of compound sensor collection, people's behavioural habits are predicted and analyzed.

Hidden layer: change the number (M) of hidden layer neuron, from N, N+1, N+2 ... (N represents the number of input layer) starts.

If the data of input are as shown below:

First day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.00＜ON 〉

10111.00011.0001.00100.00001.10011.00000.01

Second day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.10＜ON 〉

10111.00011.0001.00100.00001.10011.01010.01

The 3rd day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.30＜ON 〉

10111.00011.0001.00100.00001.10011.11110.01

The 4th day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.20＜ON 〉

10111.00011.0001.00100.00001.10011.10100.01

The 5th day:＜23〉＜III〉＜D1〉＜R4〉＜01＜18.00＜ON 〉

10111.00011.0001.00100.00001.10010.00000.01

The 6th day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.05＜ON 〉

10111.00011.0001.00100.00001.10011.00101.01

The 7th day:＜23〉＜III〉＜D1〉＜R4〉＜01＜19.30＜ON 〉

10111.00011.0001.00100.00001.10011.11110.01

If use the data of above 7 days to train system,, by the calculating of hidden layer, just can predict the applicable value in this building, floor, dormitory, room, or according to the applicable value in the building that obtains, floor, dormitory, room, analyze.

Output layer: neural network obtains Output rusults by output layer.This output valve, can remind people should carry out which type of action at this time point.For example,, by neural network, predict the 8th day, should carry out the action of turning off the light 18: 50 to 19: 30.In practice, if the time has surpassed 17: 30, system sends a note will for this people, and whether the reminding user inspection turns off the light.

In the same way, use neural network, can also predict other behavioural habits of people in Smart Home, and abnormal in analysis and discovery behavior, in time send and remind and early warning information.

Neural network adopts the following methods training:

In the process of training/study, use following algorithm:

1, create feedforward neural network, have N input, M=N, N+1, N+2 ... an individual hidden unit and P output unit.

2, initialization all-network weight

$W_{i,j}^1;i=\stackrel{\&OverBar;}{1,N};j=\stackrel{\&OverBar;}{1,M}$ $W_{i,j}^2;i=\stackrel{\&OverBar;}{1,M}$ -------------------------------Shi 1

This weight is [N/2, N/2] interval interior random digit.

3. if $E\left(W\right)=\frac{1}{2}{\mathrm{\&Sigma;}}_{k\&Element;\mathrm{outputs}\left(P\right)}{(t_k-o_k)}^2\&le;T,$

Wherein: E (W) represents error function;

t _kThe expression desired value;

o _kThe expression output valve;

T represents threshold values;

Carry out;

3.1 to network input example

And calculate and export

$\stackrel{\&OverBar;}{O}=\stackrel{\&OverBar;}{X}\&CenterDot;\stackrel{\&OverBar;}{W^1}\&CenterDot;\stackrel{\&OverBar;}{W^2}$ --------------------------------------formula 2

Wherein:

Expression: the sample of input;

Expression: weight;

Expression: weight;

3.2., for each network output unit k, during k=1, calculate its error term δ _k

δ _k=o _k(1-o _k) (t _k-o _k)-----------------------------formula 3

3.3 for each hidden unit H, H=1, M, calculate its error delta _h ${\mathrm{\&delta;}}_h=o_h(1-o_h){\mathrm{\&Sigma;}}_{k\&Element;\mathrm{outputs}\left(P\right)}{W}_{k,h.}^2{\mathrm{\&delta;}}_k.$ --------------------------------------formula 4

3.4 upgrade the weights W i of each network, j

W _{I, j}=W _{I, j}+ Δ W _{I, j}--------------------------------------formula 5

Δ W _{I, j}=α. δ ₁X _{I, j}--------------------------------------formula 6

Wherein, α is the step number of study.Weight is interval interior by random initializtion at [2/N, 2/N].N is the neuron number of input layer., in order to obtain better result, use following activation function:

F (X)=(2/ (1+e ^-x))-1--------------------------------------formula 7

X represents input value.

In the present invention, use neural network dynamic study, one of difference of dynamic learning and Static Learning is in the dynamic learning process, based on the result of feedforward step, to predict.This means,, if predicted value is correct, only make once backward process., if predicted value is incorrect, before obtaining the correct Prediction result, carry out backward operation always.In this case, the dynamic learning process starts with the weight that the muscle-setting exercise process obtains.

In sum, the invention provides a kind of Intelligent home monitoring system based on cloud computing and method, have the following advantages:

(1) cross the multiclass monitor message that compound sensor can the overall monitor monitoring area, namely saved system cost, simplified wiring, also improved system stability.

(2) concentrating equipment has data encryption feature, thereby has improved the security of supervisory system;

(3) cloud server adopts a plurality of assemblies, has ensured the security that is stored in the cloud server data;

(4) cloud server adopts nerual network technique analyst's behavioural habits, and when the behavioural habits of finding Monitoring Data and people are not inconsistent, to associated user's mobile phone, sends prompting message, has promoted user's experience.

The above is only the preferred embodiment of the present invention; should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be looked protection scope of the present invention.

Claims (10)

1. the Intelligent home monitoring system based on cloud computing, is characterized in that, comprising: compound sensor, concentrating equipment and cloud server; Described compound sensor comprises at least a sensor, described sensor is connected with all kinds of monitored devices or is positioned at monitoring area, the output terminal of described compound sensor is connected with the input end of described concentrating equipment, and the output terminal of described concentrating equipment is connected with the input end of described cloud server.

2. the Intelligent home monitoring system based on cloud computing according to claim 1, is characterized in that, described monitored device comprises one or more in gas-cooker, TV, refrigerator, air-conditioning, lamp and water heater.

3. the Intelligent home monitoring system based on cloud computing according to claim 1, it is characterized in that, described compound sensor comprises processor, optical sensor, gas sensor, humidity sensor, temperature sensing, proximity transducer, infrared sensor and communication interface; Described processor is connected with described optical sensor, described gas sensor, described humidity sensor, described temperature sensing, described proximity transducer, described infrared sensor and the communication interface of being connected respectively.

4. the Intelligent home monitoring system based on cloud computing according to claim 3, is characterized in that, described communication interface is bluetooth communication interface or zigbee communication interface.

5. the Intelligent home monitoring system based on cloud computing according to claim 1, it is characterized in that, described concentrating equipment comprises arm processor, wifi communication interface, zigbee communication interface, rs485 communication interface, rs232 communication interface and usb communication interface; Described arm processor is connected with described wifi communication interface, described zigbee communication interface, described rs485 communication interface, described rs232 communication interface and the usb communication interface of being connected respectively.

6. the Intelligent home monitoring system based on cloud computing according to claim 1, it is characterized in that, described cloud server comprises that database, individual/family data manager, data course manager, Authorization Manager, data sensitive degree check that manager, sensitive data analysis management device, equipment manager, User Manager, log manager, metering and accounting management device, security manager, service managerZ-HU and service provide manager;

Described User Manager is used for supervisor authority user and relevant file thereof; Wherein, described file comprise individual subscriber file, station address file and with user-dependent telephone number file;

Described individual/family data manager is for the protection of the individual subscriber file of the described authorized user of described user management management;

Described Authorization Manager is used for to the designated user granted rights, makes described designated user access or shares other users' data;

Described data course manager is used for monitoring the activity of the storage data of described individual/family data manager mandate, specifically comprises: access historical record in described database the access time point, access content and the corresponding relation of executable operations;

Described service managerZ-HU is used for to user's Push Service catalogue, and described service catalogue is enumerated service content that described Intelligent home monitoring system provides to the user and the pay imformation corresponding with described service content;

Described service provides manager to be used for the access of automatic management user to described cloud server;

Described metering and accounting management device are used for calculating the expense situation of collecting to each authorized user;

Described log manager is used for recording described cloud server event or movable content and corresponding time point;

Described equipment manager is used for recording the operation information of each authorized user monitored device;

Described data sensitive degree checks that manager is used for checking user's request service information, the memory location of service data in database of judgement user request;

Described sensitive data analysis management device is used for, based on neural network, the monitor message that receives is carried out analysis-by-synthesis, by reference to the same tutorial historical data of storing in database, at first infer people's behavioural habits, then judge whether described monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message;

Described security manager is used to service that authentication, mandate and key management are provided.

7. the method for supervising of the described Intelligent home monitoring system based on cloud computing of application rights requirement 1-6 any one, is characterized in that, comprises the following steps:

S1, install respectively described compound sensor in each monitoring area; Wherein, described compound sensor comprises at least a sensor, and described sensor is connected with all kinds of monitored devices or is positioned at monitoring area;

S2, described compound sensor acquisition monitoring information, then send to described concentrating equipment with described monitor message; Wherein, described monitor message comprises the equipment operational factor of all kinds of monitored devices that are connected with self and the environmental information of monitoring area;

S3, described concentrating equipment receives described monitor message, and described monitor message is encrypted, and obtains encrypting monitor message; Then described encryption monitor message is sent to described cloud server;

S4, the described encryption monitor message that described cloud server deciphering receives, reduction obtains described monitor message, and described monitor message is stored in database;

S5, described cloud server employing Analysis of Neural Network Method is stored in the data in database, judges whether described monitor message is abnormal,, if abnormal, to the cell-phone number corresponding with described monitoring area of reserving, sends prompting message.

8. method for supervising according to claim 7, is characterized in that, S5 is specially:

Described cloud server employing Analysis of Neural Network Method is stored in the historical data in database, at first infer people's behavioural habits, then judge whether described monitor message meets people's behavioural habits,, if do not meet, to the corresponding phone number of reserving, send alert notification message.

9. method for supervising according to claim 7, is characterized in that, S5 is specially:

Described cloud server uses neural network learning people's daily behavior custom, adopts the Multilayer Perception backpropagation train and revise, and sets up the behavioural habits neural network model;

Described cloud server is according to the described behavioural habits neural network model of setting up, prediction and analyst's future behaviour.

10. method for supervising according to claim 9, is characterized in that, described cloud server uses neural network learning people's daily behavior custom, adopts the Multilayer Perception backpropagation train and revise, and sets up the behavioural habits neural network model and is specially:

S51, the topological structure of establishment feedforward neural network, described feedforward neural network comprises input layer, hidden layer and output layer; Described input layer has N input variable, and described hidden layer has M=N, N+1, N+2 ... individual hidden unit; Described output layer has P output variable;

Wherein, input variable comprise building room number, floor number, dormitory number, room number and when this room to room in the operation information carried out of application-specific;

Output variable is: when application-specific is abnormal in drawing room, to the cellphone subscriber corresponding with this room, send abnormal reminder message;

S52, use historical data as the sample data of neural metwork training, with back-propagation algorithm, neural network trained, and obtains described behavioural habits neural network model; Wherein, with back-propagation algorithm, neural network is trained and is specially: follow these steps to constantly update each layer connection weight value:

S521, employing formula 1 initialization all-network connection weight value;

$W_{i,j}^1;i=\stackrel{\&OverBar;}{1,N};j=\stackrel{\&OverBar;}{1,M}$ $W_{i,j}^2;i=\stackrel{\&OverBar;}{1,M}.$ ----------------------------------Shi 1

This weighted value is [N/2, N/2] interval interior random digit;

S522, if

Carry out S523-S526;

Wherein: E (W) represents error function;

t _kThe expression desired value;

o _kThe expression output valve;

T represents threshold values;

S523, to network input example

And by formula 2, calculate output

$\stackrel{\&OverBar;}{O}=\stackrel{\&OverBar;}{X}\&CenterDot;\stackrel{\&OverBar;}{W^1}\&CenterDot;\stackrel{\&OverBar;}{W^2}$ ----------------------------------Shi 2

Wherein:

Expression: the sample of input;

Expression: weight;

Expression: weight;

S524,, for each network output unit k, during k=1, calculate its error term δ by formula 3 _k

δ _k=o _k(1-o _k) (t _k-o _k)---------------------------------formula 3

S525, for each hidden unit H, H=1, M, calculate its error delta by formula 4 _h

${\mathrm{\&delta;}}_h=o_h(1-o_h){\mathrm{\&Sigma;}}_{k\&Element;\mathrm{outputs}\left(P\right)}{W}_{k,h.}^2{\mathrm{\&delta;}}_k.$ ----------------------------------Shi 4

Wherein,

K the weight of expression output valve h;

S526, by the connection weight value W of formula 5, formula 6 each network of renewal _{I, j}

W _{I, j}=W _{I, j}+ Δ W _{I, j}---------------------------------formula 5

Δ W _{I, j}=α. δ ₁X _{I, j}----------------------------------Shi 6

Wherein, α represents the step number of learning.

Images