CN113758520A - Internet of things laboratory environment monitoring method and system based on neural network - Google Patents
Internet of things laboratory environment monitoring method and system based on neural network Download PDFInfo
- Publication number
- CN113758520A CN113758520A CN202110948130.8A CN202110948130A CN113758520A CN 113758520 A CN113758520 A CN 113758520A CN 202110948130 A CN202110948130 A CN 202110948130A CN 113758520 A CN113758520 A CN 113758520A
- Authority
- CN
- China
- Prior art keywords
- internet
- things
- preset
- neural network
- neuron
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/048—Activation functions
-
- 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
Abstract
The invention is suitable for the technical field of metering equipment, and provides an Internet of things laboratory environment monitoring method and system based on a neural network, wherein the method comprises the following steps: acquiring sensor data in the Internet of things; combining the sensor data with a preset connection weight factor and a threshold value in a neuron, and calculating a result by adopting a preset transfer function; and forming an execution amount according to the result and outputting the execution amount to an execution mechanism. In the embodiment of the invention, the system has the capabilities of self-perception, self-learning and self-decision by collecting the environmental parameter data of each sensor in the Internet of things, processing, calculating and internally learning the data through the neural network and adjusting the weight coefficient to obtain the expected result.
Description
Technical Field
The invention relates to the technical field of Internet of things, in particular to a method and a system for monitoring the environment of an Internet of things laboratory based on a neural network.
Background
The laboratory is the place for carrying out the test; the laboratory is a scientific cradle which is a base of scientific research and a source of scientific development, and plays a very important role in the scientific development; laboratories can be classified into three categories according to affiliation: the first is a laboratory belonging to or hosted by a university; the second kind of laboratory belongs to national institutions, and some laboratories even belong to international institutions; the third kind of laboratories directly belongs to the industrial enterprise sector and serves the development and research of industrial technologies. The laboratory is a scientific cradle, is a base of scientific research and a source of scientific development, and plays a very important role in the scientific development.
At present, the existing laboratory environment monitoring system in China has the problems that factors such as system parameters and control strategies are set artificially, so that the health of a user is influenced when the user works in a laboratory, the laboratory environment cannot be monitored, the health of the laboratory environment cannot be guaranteed, and the air quality, energy consumption, waste gas and waste water required in the laboratory cannot be effectively monitored.
Therefore, it is necessary to invent a method and a system for monitoring an internet of things laboratory environment based on a neural network to solve the above technical problems.
Disclosure of Invention
The invention aims to provide an Internet of things laboratory environment monitoring method and system based on a neural network, and aims to solve the problems in the background technology.
The embodiment of the invention provides an Internet of things laboratory environment monitoring method based on a neural network, which comprises the following steps:
acquiring sensor data in the Internet of things;
combining the sensor data with a preset connection weight factor and a threshold value in a neuron, and calculating a result by adopting a preset transfer function;
and forming an execution amount according to the result and outputting the execution amount to an execution mechanism.
Further, combining the data with a preset connection weight factor and a threshold value in the neuron, and calculating a result by adopting a preset transfer algorithm, wherein a specific calculation formula is as follows:
wherein Xm (i ═ 1,2, 3.. m) is a plurality of input parameters; yj is a single output; thetaiIs a threshold value; wijIs the connection weight factor from neuron i to neuron j; f () is a transfer function, i.e. f (x) sigmoid (x).
Further, a preset error back-propagation learning algorithm is adopted to calculate the network training error of the training sample (Xk, Yk), and the specific formula is as follows:
wherein E isKFor the network training error, y' lj is the actual output value.
Further, a preset error back-propagation learning algorithm is adopted to calculate the total network training error of the training samples (Xk, Yk), and the specific formula is as follows:
wherein E is the total error.
Further, the preset correction weight algorithm formula is as follows:
where μ is the learning rate, μ ranges from 0.01 to 1.
The embodiment of the invention provides an Internet of things laboratory environment monitoring system based on a neural network, which comprises an Internet of things sensing unit, a neural network server and an execution unit;
the sensing unit of the Internet of things is used for acquiring environmental parameters, converting the environmental parameters into sensor data and outputting the sensor data;
the neural network server is used for acquiring sensor data of the sensing unit of the Internet of things, combining the sensor data with a connection weight factor and a threshold value preset in a neuron and then calculating a result by adopting a preset transfer function;
and the execution unit is used for forming an execution quantity execution action according to the result.
Further, the neural network server comprises a network connection module, a data acquisition module, a neuron processing module and an output module;
the network connection module is used for establishing network connection with the Internet of things sensing unit;
the data acquisition module is used for acquiring sensor data of the sensing unit of the Internet of things;
the neuron processing module is used for combining the sensor data with a preset connection weight factor and a threshold value in a neuron and then calculating a result by adopting a preset transfer algorithm;
and the output module is used for outputting the result to the execution unit.
Further, the neuron processing module further comprises a transfer algorithm module, which is used for calculating a result according to a preset transfer algorithm calculation formula.
Further, the neuron processing module further comprises an error training module, which is used for calculating the network training errors of the training samples by adopting a preset error back-propagation learning algorithm to obtain the network training errors of a single training sample and the total errors of all the training samples.
Furthermore, the error training module further comprises a network weight initial value correction module, which is used for continuously correcting the network weight initial value by adopting a preset correction weight algorithm.
According to the method for monitoring the environment of the Internet of things laboratory based on the neural network, the expected result is obtained by collecting the environmental parameter data of each sensor in the Internet of things, processing, calculating and learning the data through the neural network and adjusting the weight coefficient, and the method has the capabilities of self-perception, self-learning and self-decision.
Drawings
FIG. 1 is a schematic network structure diagram of an Internet of things laboratory environment monitoring system based on a neural network;
FIG. 2 is a flow chart of a neural network-based laboratory environment monitoring method for the Internet of things;
FIG. 3 is a diagram of an example of a neuron structure of a neural network-based Internet of things laboratory environment monitoring method;
FIG. 4 is an exemplary diagram of a neural network structure of a method for monitoring the environment of a laboratory of the Internet of things based on a neural network;
FIG. 5 is a schematic structural diagram of a laboratory environment monitoring system of the Internet of things based on a neural network;
FIG. 6 is a schematic structural diagram of another Internet of things laboratory environment monitoring system based on a neural network;
fig. 7 is a diagram illustrating a structure of another internet of things laboratory environment monitoring system based on a neural network.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the embodiment of the invention, a meter detector can automatically identify meters in the surrounding environment through the mobile terminal and can be selectively connected with the meters to acquire meter data, and then the operation state of the on-site meter is judged to be accurate and the precision of the on-site meter reaches the standard through calculation and comparison of the remote server, so that the manual operation is greatly reduced.
Fig. 1 is a schematic network structure diagram of a neural network-based laboratory environment monitoring system of the internet of things, which is suitable for an embodiment of the present invention, and includes at least one neuron sensing unit 1, a neuron network server 2, and at least one execution unit 3, and for convenience of description, only parts related to the present invention are shown.
In the embodiment of the present invention, the neuron sensing unit 1 and the neuron network server 2, the neuron network server 2 and the execution unit 3 are in communication connection in a wireless or wired manner, specifically, the communication connection may be implemented in a wireless connection manner such as WiFi, bluetooth, infrared, ZIGBEE, LORA, or the like, or in a wired connection manner such as ethernet, optical fiber, modbus, canbus, lonworks, devicenet, sparbus, profibus, or the like, which is not limited in the present invention. The particular communication protocol is also not limited herein. In the embodiment of the present invention, the connection modes that may exist between the neuron sensing unit 1 and the neuron network server 2, and between the neuron network server 2 and the execution unit 3 include one-to-one, one-to-many, many-to-one, or many-to-many.
In the embodiment of the present invention, the neuron sensing unit 1 may be various sensor units, such as an air quality sensor, a differential pressure sensor, a wind speed sensor, a temperature and humidity sensor, an electric energy sensor, a water quality sensor, and the like. The neural network server 2 is a kind of computer, which runs faster and is more highly loaded than a general computer. The neuron network server 2 provides computation or application services for other neuron sensing units 1 in the system. The neuron network server 2 has high-speed CPU computing capability, long-time reliable operation, strong I/O external data throughput capability and better expansibility, and has the capability of bearing response service requests, bearing services and guaranteeing services.
The neuron network server 2 is the core of the overall architecture of the environmental monitoring system of the internet of things laboratory, and mainly solves the problems of how to store, retrieve and use data, data security and privacy protection and the like. The neuron network server 2 is responsible for effectively integrating and utilizing the information collected by the neuron sensing unit 1 through technologies such as big data and cloud computing, and then sending a related decision command to the execution unit 3 for execution. The execution unit 3 comprises a laboratory ventilation unit, a laboratory air conditioning unit, a laboratory wastewater treatment unit, a laboratory power supply unit, a laboratory alarm unit (audible and visual alarm) and the like.
Fig. 2 shows a flow chart of a method for monitoring an environment of a laboratory of the internet of things based on a neural network, which is applicable to an embodiment of the present invention, and specifically includes the following steps:
and S101, acquiring sensor data in the Internet of things.
In an embodiment of the present invention, data of a sensor in the internet of things includes: an air quality sensor for detecting indoor temperature, humidity, PM1.0/PM2.5/PM10, TVOC, CO2、NO2Providing parameters for the laboratory ventilation unit; the pressure difference sensor is used for detecting the relative pressure difference between rooms and providing parameters for the laboratory ventilation unit; the air speed sensor is used for detecting indoor and outdoor air speed and air volume and providing parameters for the laboratory ventilation unit; the temperature and humidity sensor is used for detecting indoor and outdoor temperatures and relative humidity and providing parameters for the laboratory air conditioning unit; the electric energy sensor is used for detecting electric parameters such as voltage, current, power consumption, power factors, frequency and the like of an indoor power supply system and an experimental instrument. The water quality sensor detects the discharge of indoor waste water, mainly detects waste water parameters such as COD, BOD, SS, organic solvent, bacteria detection, PH value, heavy metal content in the waste water, provides the parameter for laboratory waste water treatment unit.
Preferably, the internet of things sensor can be replaced by an internet of things monitoring unit, such as an entrance guard unit, a video monitoring unit and the like, as long as the parameters of behaviors and changes generated in the environment or human production and living activities can be acquired.
And S102, combining the sensor data with a preset connection weight factor and a threshold value in the neuron, and calculating a result by adopting a preset transfer function.
In the embodiment of the invention, the acquired sensor data is combined with a preset connection weight factor and a threshold value in a neuron, and a result is calculated by adopting a preset transfer function. The sensor data may be one data source or a plurality of data sources, and each sensor data Xm (i ═ 1,2, 3.. m) is associated with a respective associated connection weight factor W as shown in fig. 3iAnd j is combined and then calculated through a transfer function, and then a related result can be obtained. The specific calculation formula is as follows：
Wherein Xm (i ═ 1,2, 3.. m) is a plurality of input parameters;
yj is a single output;
θiis a threshold value;
Wij is the connection weight factor from neuron i to neuron j;
f () is a transfer function, i.e. f (x) sigmoid (x).
Preferably, as shown in fig. 4, the number of the neurons is one or more, and is not limited herein, a plurality of neurons form a neural network, there is a data connection between neurons, a result output by one neuron can be used as an input of a plurality of other neurons, and a neuron can also receive a result output by a plurality of neurons. Data connections between different neurons require a weight factor W with the connectionsij, the connection weight factor, i.e. the weight coefficient, may be adjusted in order to obtain the desired output result.
The neural network collects environmental parameters of each sensor in the Internet of things as input data through the sensing unit of the Internet of things, the execution quantity formed by the result processed and calculated by the neural network is output as the execution quantity of the execution unit of the neural network, and the expected result is obtained by adjusting the weight coefficient through internal learning and processing of the neural network.
Preferably, in order to obtain more accurate results, the neural network also needs to perform network training on each sample. In the embodiment of the invention, an error back-propagation learning algorithm is adopted to calculate the network training error of each training sample (Xk, Yk), and the specific formula is as follows:
wherein E isKFor the network training error, y' lj is the actual output value.
The network training error of each training sample can be calculated through the error back-propagation learning algorithm.
Preferably, for a plurality of input training samples, the total error of network training of all the training samples needs to be calculated, and for this purpose, a preset error back-propagation learning algorithm is adopted to calculate the total error of network training of the training samples (Xk, Yk), and the specific formula is as follows:
wherein E is the total error.
Preferably, the total error E obtained by the error back-propagation learning algorithm may have different values, i.e., may or may not meet the requirement, and therefore, the weight connection factor needs to be continuously corrected. Continuously correcting the initial value of the network weight by adopting a preset correction weight algorithm, wherein the preset correction weight algorithm formula is as follows:
where μ is the learning rate, μ ranges from 0.01 to 1.
The total error obtained after calculation by the weight correction algorithm can meet the requirement.
And step S103, forming an execution amount according to the result and outputting the execution amount to an execution mechanism.
In the embodiment of the invention, the result calculated by the neural network processing further needs to be converted into the execution amount which can be executed by the execution unit to finally complete the control of the execution unit, and the result is converted differently according to different sensors and execution units of the internet of things to form the execution amount matched with the execution unit as the execution amount of the execution unit to be output by the neural network.
In the embodiment of the invention, the air quality in a laboratory can be monitored by an air quality sensor, the environmental parameters are converted and transmitted to a wireless transmission module by a processor by detecting various environment parameters such as temperature, humidity, PM1.0/PM2.5/PM10, TVOC, CO2, NO2 and the like in indoor air, an edge gateway receives the environment parameters and transmits the environment to a neuron network server, and after the environment is processed by the neuron network server, a relevant decision command is sent to a laboratory ventilation unit which comprises an exhaust fan and a fresh air system.
The air pressure difference outside, inside and between laboratories is monitored through differential pressure sensor, air velocity transducer, and the treater converts the air pressure difference parameter and transmits for wireless transmission module, and the marginal gateway receives the air pressure difference parameter, sends neuron network server to, and neuron network server handles the back, sends relevant decision-making command for laboratory ventilation unit, and ventilation unit contains exhaust fan, new trend system.
Monitoring the air temperature and humidity outside, inside and between laboratories through temperature and humidity sensors, converting and transmitting air temperature and humidity parameters to a wireless transmission module by a processor, receiving the air temperature and humidity parameters by an edge gateway, transmitting the air temperature and humidity parameters to a neural network server, and sending a relevant decision command to a laboratory air conditioning unit and a laboratory ventilation unit after the processing by the neural network server, wherein the laboratory air conditioning unit comprises a humidifier, a dehumidifier and an air conditioner; the ventilation unit comprises an exhaust fan and a fresh air system.
The electric parameters of the laboratory power supply end and the instrument end are monitored through the electric energy sensor, the electric parameters comprise voltage, current, power consumption, power factors, frequency and the like, the processor converts the electric parameters and transmits the electric parameters to the wireless transmission module, the edge gateway receives the electric parameters and transmits the electric parameters to the neural network server, and after the neural network server processes the electric parameters, the relevant decision-making command is sent to the laboratory power supply end and the instrument end. And carrying out related power-off or power-on and power-off operations.
Waste water in the laboratory is monitored through the water quality sensor, water quality parameters comprise COD, BOD, SS, organic solvent, bacteria detection, PH value, heavy metal content and the like, the processor converts the water quality parameters and transmits the water quality parameters to the wireless transmission module, the edge gateway receives the water quality parameters and transmits the water quality parameters to the neural network server, and after the neural network server processes the water quality parameters, a relevant decision-making command is sent to the laboratory waste water treatment unit. The laboratory wastewater treatment unit comprises a direct discharge valve, a wastewater collection tank and the like;
the access control unit is used for recording and verifying the identity of personnel in a laboratory and the identity of the personnel in the laboratory, the processor converts the recording and verifying information and transmits the information to the wireless transmission module, the edge gateway receives the recording and verifying information and transmits the information to the neural network server, and the neural network server performs identity verification, registration and recording;
personnel and the environment in the laboratory are monitored through the video monitoring unit, the video stream is transmitted to the neural network server through the network equipment, and the neural network server records and stores the video stream in real time.
In summary, the system has the capabilities of self-perception, self-learning and self-decision by collecting the environmental parameter data of each sensor in the internet of things, processing, calculating and learning the data through the neural network and adjusting the weight coefficient to obtain the expected result.
Fig. 5 is a block diagram illustrating a structure of a neural network-based internet of things laboratory environment monitoring system suitable for an embodiment of the present invention, including an internet of things sensing unit 1, a neural network server 2, and an execution unit;
the sensing unit 1 of the internet of things is used for acquiring environmental parameters, converting the environmental parameters into sensor data and outputting the sensor data;
the sensing unit of the internet of things in the embodiment of the invention is the basis of the whole system architecture and comprises an information acquisition module, a personnel monitoring module, a signal conversion processing module and a network connection module. The information acquisition module comprises various sensors of the internet of things, such as: an air quality sensor for detecting indoor temperature, humidity, PM1.0/PM2.5/PM10, TVOC, CO2、NO2Providing parameters for the laboratory ventilation unit; the pressure difference sensor is used for detecting the relative pressure difference between rooms and providing parameters for the laboratory ventilation unit; the air speed sensor is used for detecting indoor and outdoor air speed and air volume and providing parameters for the laboratory ventilation unit; temperature and humidity sensorA sensor for detecting indoor and outdoor temperatures and relative humidity and providing parameters for a laboratory air conditioning unit; the electric energy sensor is used for detecting electric parameters such as voltage, current, power consumption, power factors, frequency and the like of an indoor power supply system and an experimental instrument. The water quality sensor detects the discharge of indoor waste water, mainly detects waste water parameters such as COD, BOD, SS, organic solvent, bacteria detection, PH value, heavy metal content in the waste water, provides the parameter for laboratory waste water treatment unit.
The personnel monitoring module includes: the system comprises an access control unit, a video monitoring unit and the like, and the access control unit, the video monitoring unit and the like can be used as long as parameters of behaviors and changes generated in the environment or human production and living activities can be acquired.
The signal conversion processing module is used for converting the information acquired by the information acquisition module and the personnel monitoring module into a signal capable of being remotely transmitted for output.
The network connection module is used for establishing network connection with the neural network server 2 and transmitting data, and can be wireless transmission connection or wired network transmission connection. The wireless transmission part is composed of ZIGBEE wireless sub-networks, LORA wireless sub-networks and other sub-networks and is responsible for transmitting signals collected and processed by the sensing unit 1 of the internet of things to the neural network server 2. The wired network transmission part transmits the acquisition information of the laboratory entrance guard control unit and the video stream of the video monitoring to the neural network server 2.
The neural network server 2 is used for acquiring sensor data of the sensing unit 1 of the Internet of things, combining the sensor data with a preset connection weight factor and a threshold value in a neuron, and calculating a result by adopting a preset transfer function;
in the embodiment of the present invention, the neural network server 2 combines the acquired sensor data with a connection weight factor and a threshold value preset in the neuron, and calculates a result by using a preset transfer function. The sensor data may be one data source or a plurality of data sources, and each sensor data Xm (i ═ 1,2, 3.. m) is associated with a respective associated connection weight factor W as shown in fig. 3iAnd j is combined and then calculated through a transfer function, and then a related result can be obtained. Concrete computing deviceThe formula is as follows:
wherein Xm (i ═ 1,2, 3.. m) is a plurality of input parameters;
yj is a single output;
θiis a threshold value;
Wij is the connection weight factor from neuron i to neuron j;
f () is a transfer function, i.e. f (x) sigmoid (x).
The number of the neurons is one or more, and is not limited herein, a plurality of neurons form a neural network, data connection exists between the neurons and the neurons, the output result of one neuron can be used as the input of a plurality of other neurons, and one neuron can also receive the output result of a plurality of neurons. Data connections between different neurons require a weight factor W with the connectionsij, the connection weight factor, i.e. the weight coefficient, may be adjusted in order to obtain the desired output result.
The neural network server 2 collects environmental parameters of each sensor in the internet of things as input data through the internet of things sensing unit 1, an execution quantity formed by processing a calculated result by the neural network server 2 is used as an execution quantity of the execution unit and is output by the neural network, and an expected result is obtained by adjusting weight coefficients through internal learning and processing of the neural network.
And the execution unit 3 is used for forming an execution quantity execution action according to the result.
In the embodiment of the present invention, the result calculated by the neural network server 2 further needs to be converted into the execution amount that can be executed by the execution unit 3 to finally complete the control of the execution unit 3, and the result is converted differently according to different internet of things sensing units 1 and execution units 3 to form the execution amount matched with the execution unit 3 as the execution amount of the execution unit 3, which is the output of the neural network.
In the embodiment of the invention, the air quality in a laboratory can be monitored by an air quality sensor, the environmental parameters are converted and transmitted to a wireless transmission module by a processor by detecting various environment parameters such as temperature, humidity, PM1.0/PM2.5/PM10, TVOC, CO2, NO2 and the like in indoor air, an edge gateway receives the environment parameters and transmits the environment to a neuron network server, and after the environment is processed by the neuron network server, a relevant decision command is sent to a laboratory ventilation unit which comprises an exhaust fan and a fresh air system.
The air pressure difference outside, inside and between laboratories is monitored through differential pressure sensor, air velocity transducer, and the treater converts the air pressure difference parameter and transmits for wireless transmission module, and the marginal gateway receives the air pressure difference parameter, sends neuron network server to, and neuron network server handles the back, sends relevant decision-making command for laboratory ventilation unit, and ventilation unit contains exhaust fan, new trend system.
Monitoring the air temperature and humidity outside, inside and between laboratories through temperature and humidity sensors, converting and transmitting air temperature and humidity parameters to a wireless transmission module by a processor, receiving the air temperature and humidity parameters by an edge gateway, transmitting the air temperature and humidity parameters to a neural network server, and sending a relevant decision command to a laboratory air conditioning unit and a laboratory ventilation unit after the processing by the neural network server, wherein the laboratory air conditioning unit comprises a humidifier, a dehumidifier and an air conditioner; the ventilation unit comprises an exhaust fan and a fresh air system.
The electric parameters of the laboratory power supply end and the instrument end are monitored through the electric energy sensor, the electric parameters comprise voltage, current, power consumption, power factors, frequency and the like, the processor converts the electric parameters and transmits the electric parameters to the wireless transmission module, the edge gateway receives the electric parameters and transmits the electric parameters to the neural network server, and after the neural network server processes the electric parameters, the relevant decision-making command is sent to the laboratory power supply end and the instrument end. And carrying out related power-off or power-on and power-off operations.
Waste water in the laboratory is monitored through the water quality sensor, water quality parameters comprise COD, BOD, SS, organic solvent, bacteria detection, PH value, heavy metal content and the like, the processor converts the water quality parameters and transmits the water quality parameters to the wireless transmission module, the edge gateway receives the water quality parameters and transmits the water quality parameters to the neural network server, and after the neural network server processes the water quality parameters, a relevant decision-making command is sent to the laboratory waste water treatment unit. The laboratory wastewater treatment unit comprises a direct discharge valve, a wastewater collection tank and the like;
the access control unit is used for recording and verifying the identity of personnel in a laboratory and the identity of the personnel in the laboratory, the processor converts the recording and verifying information and transmits the information to the wireless transmission module, the edge gateway receives the recording and verifying information and transmits the information to the neural network server, and the neural network server performs identity verification, registration and recording;
personnel and the environment in the laboratory are monitored through the video monitoring unit, the video stream is transmitted to the neural network server through the network equipment, and the neural network server records and stores the video stream in real time.
In summary, the system has the capabilities of self-perception, self-learning and self-decision by collecting the environmental parameter data of each sensor in the internet of things, processing, calculating and learning the data through the neural network and adjusting the weight coefficient to obtain the expected result.
Further, the neural network server 2 includes a network connection module 21, a data acquisition module 22, a neuron processing module 23 and an output module 24;
the network connection module 21 is used for establishing network connection with the internet of things sensing unit 1;
the data acquisition module 22 is used for acquiring sensor data of the sensing unit 1 of the internet of things;
the neuron processing module 23 is configured to combine the sensor data with a preset connection weight factor and a threshold in a neuron, and then calculate a result by using a preset transfer algorithm;
an output module 24, configured to output the result to the execution unit 3.
Further, fig. 6 shows a block diagram of another neural network-based internet of things laboratory environment monitoring system suitable for the embodiment of the present invention, where the neuron processing module 23 includes a transfer algorithm module 231;
and a transfer algorithm module 231, configured to calculate a result according to a preset transfer algorithm calculation formula, where the transfer algorithm is as described above.
Further, the neuron processing module 23 includes an error training module 232;
and the error training module 232 is configured to calculate the network training errors of the training samples by using a preset error back-propagation learning algorithm, so as to obtain the network training errors of a single training sample and the total errors of all the training samples.
In the embodiment of the present invention, in order to obtain a more accurate result, the neural network server 2 further needs to perform network training on each sample by using the error training module 232. Calculating the network training error of each training sample (Xk, Yk) by adopting an error back propagation learning algorithm, wherein the specific formula is as follows:
wherein E isKFor the network training error, y' lj is the actual output value.
The network training error of each training sample can be calculated through the error back-propagation learning algorithm.
Preferably, for a plurality of input training samples, the total error of network training of all the training samples needs to be calculated, and for this purpose, a preset error back-propagation learning algorithm is adopted to calculate the total error of network training of the training samples (Xk, Yk), and the specific formula is as follows:
wherein E is the total error.
Further, the error training module 232 includes a network weight initial value modification module 2321;
the network weight initial value correcting module 2321 is configured to continuously correct the network weight initial value by using a preset correction weight algorithm.
In the embodiment of the present invention, the total error E obtained by the error back-propagation learning algorithm may have different values, that is, may meet the requirement, or may not meet the requirement, so that the weight connection factor needs to be continuously corrected by the network weight initial value correcting module 2321. Continuously correcting the initial value of the network weight by adopting a preset correction weight algorithm, wherein the preset correction weight algorithm formula is as follows:
where μ is the learning rate, μ ranges from 0.01 to 1.
It should be understood that, although the steps in the flowcharts of the embodiments of the present invention are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in various embodiments may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. A laboratory environment monitoring method for the Internet of things based on a neural network is characterized by comprising the following steps:
acquiring sensor data in the Internet of things;
combining the sensor data with a preset connection weight factor and a threshold value in a neuron, and calculating a result by adopting a preset transfer function;
and forming an execution amount according to the result and outputting the execution amount to an execution mechanism.
2. The method for monitoring the environment of the internet of things laboratory based on the neural network as claimed in claim 1, wherein the data is combined with a preset connection weight factor and a threshold in the neuron, and a result is calculated by adopting a preset transfer algorithm, wherein the specific calculation formula is as follows:
wherein Xm (i ═ 1,2, 3.. m) is a plurality of input parameters; y isjIs a single output; thetaiIs a threshold value; wijIs the connection weight factor from neuron i to neuron j; f () is a transfer function, i.e. f (x) sigmoid (x).
3. The method for monitoring the environment of the internet of things laboratory based on the neural network as claimed in claim 2, wherein the combining the data with the connection weight factor and the threshold value preset in the neuron and calculating the result by adopting a preset transfer algorithm further comprises:
using a preset error back-propagation learning algorithm to train the sample (X)k,Yk) The network training error of (2) is calculated, and the specific formula is as follows:
wherein E isKFor the network training error, y' lj is the actual output value.
4. The method for monitoring the environment of the internet of things laboratory based on the neural network as claimed in claim 3, wherein the combining the data with the connection weight factor and the threshold value preset in the neuron and calculating the result by adopting a preset transfer algorithm further comprises:
using a preset error back-propagation learning algorithm to train the sample (X)k,Yk) The network training total error is calculated by the following specific formula:
wherein E is the total error.
5. The method for monitoring the environment of the Internet of things laboratory based on the neural network as claimed in claim 3 or 4, wherein the training samples (X) are subjected to a preset error back-propagation learning algorithmk,Yk) The calculating of the total error of the network training specifically further comprises:
continuously correcting the initial value of the network weight by adopting a preset correction weight algorithm, wherein the preset correction weight algorithm formula is as follows:
where μ is the learning rate, μ ranges from 0.01 to 1.
6. An Internet of things laboratory environment monitoring system based on a neural network is characterized by comprising an Internet of things sensing unit, a neural network server and an execution unit;
the sensing unit of the Internet of things is used for acquiring environmental parameters, converting the environmental parameters into sensor data and outputting the sensor data;
the neural network server is used for acquiring sensor data of the sensing unit of the Internet of things, combining the sensor data with a connection weight factor and a threshold value preset in a neuron and then calculating a result by adopting a preset transfer function;
and the execution unit is used for forming an execution quantity execution action according to the result.
7. The laboratory environment monitoring system for the internet of things based on the neural network as claimed in claim 6, wherein the neural network server comprises a network connection module, a data acquisition module, a neuron processing module and an output module;
the network connection module is used for establishing network connection with the Internet of things sensing unit;
the data acquisition module is used for acquiring sensor data of the sensing unit of the Internet of things;
the neuron processing module is used for combining the sensor data with a preset connection weight factor and a threshold value in a neuron and then calculating a result by adopting a preset transfer algorithm;
and the output module is used for outputting the result to the execution unit.
8. The system of claim 7, wherein the neuron processing module further comprises:
and the transfer algorithm module is used for calculating a result according to a preset transfer algorithm calculation formula.
9. The system of claim 7, wherein the neuron processing module further comprises:
and the error training module is used for calculating the network training errors of the training samples by adopting a preset error back-propagation learning algorithm to obtain the network training errors of a single training sample and the total errors of all the training samples.
10. The system of claim 9, wherein the error training module further comprises:
and the network weight initial value correcting module is used for continuously correcting the network weight initial value by adopting a preset correction weight algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110948130.8A CN113758520A (en) | 2021-08-18 | 2021-08-18 | Internet of things laboratory environment monitoring method and system based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110948130.8A CN113758520A (en) | 2021-08-18 | 2021-08-18 | Internet of things laboratory environment monitoring method and system based on neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113758520A true CN113758520A (en) | 2021-12-07 |
Family
ID=78790282
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110948130.8A Pending CN113758520A (en) | 2021-08-18 | 2021-08-18 | Internet of things laboratory environment monitoring method and system based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113758520A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015172560A1 (en) * | 2014-05-16 | 2015-11-19 | 华南理工大学 | Central air conditioner cooling load prediction method based on bp neural network |
| CN106765959A (en) * | 2016-12-27 | 2017-05-31 | 武汉虹信技术服务有限责任公司 | Heat-air conditioner energy-saving control method based on genetic algorithm and depth B P neural network algorithms |
| CN106951960A (en) * | 2017-03-02 | 2017-07-14 | 平顶山学院 | A kind of learning method of neutral net and the neutral net |
| CN109632649A (en) * | 2018-12-13 | 2019-04-16 | 云南电网有限责任公司电力科学研究院 | SF based on artificial neural network6Gas fiber laser arrays quantitative analysis method |
| CN109816094A (en) * | 2019-01-03 | 2019-05-28 | 山东省科学院海洋仪器仪表研究所 | Optical dissolved oxygen sensor non-linear temperature compensation method based on neural network L-M algorithm |
| CN111464793A (en) * | 2020-05-15 | 2020-07-28 | 哈尔滨工业大学 | Method and system for monitoring and early warning experimental device based on artificial intelligence |
| CN112034766A (en) * | 2020-09-15 | 2020-12-04 | 广州邦禾检测技术有限公司 | Laboratory safety management system based on Internet of things |
-
2021
- 2021-08-18 CN CN202110948130.8A patent/CN113758520A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015172560A1 (en) * | 2014-05-16 | 2015-11-19 | 华南理工大学 | Central air conditioner cooling load prediction method based on bp neural network |
| CN106765959A (en) * | 2016-12-27 | 2017-05-31 | 武汉虹信技术服务有限责任公司 | Heat-air conditioner energy-saving control method based on genetic algorithm and depth B P neural network algorithms |
| CN106951960A (en) * | 2017-03-02 | 2017-07-14 | 平顶山学院 | A kind of learning method of neutral net and the neutral net |
| CN109632649A (en) * | 2018-12-13 | 2019-04-16 | 云南电网有限责任公司电力科学研究院 | SF based on artificial neural network6Gas fiber laser arrays quantitative analysis method |
| CN109816094A (en) * | 2019-01-03 | 2019-05-28 | 山东省科学院海洋仪器仪表研究所 | Optical dissolved oxygen sensor non-linear temperature compensation method based on neural network L-M algorithm |
| CN111464793A (en) * | 2020-05-15 | 2020-07-28 | 哈尔滨工业大学 | Method and system for monitoring and early warning experimental device based on artificial intelligence |
| CN112034766A (en) * | 2020-09-15 | 2020-12-04 | 广州邦禾检测技术有限公司 | Laboratory safety management system based on Internet of things |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Occupancy prediction through Markov based feedback recurrent neural network (M-FRNN) algorithm with WiFi probe technology | |
| Ha et al. | Sensing data fusion for enhanced indoor air quality monitoring | |
| CN107942960B (en) | A kind of intelligentized information processing system | |
| CN104820072B (en) | Based on the monitoring method of the Electronic Nose air-quality monitoring system of cloud computing | |
| Jiang et al. | Indoor occupancy estimation from carbon dioxide concentration | |
| Massana et al. | Short-term load forecasting in a non-residential building contrasting models and attributes | |
| Tanasiev et al. | Enhancing environmental and energy monitoring of residential buildings through IoT | |
| US9933320B2 (en) | Systems, methods and devices for measurement of rate of heat exchange of airflow systems | |
| Pradityo et al. | Indoor air quality monitoring and controlling system based on IoT and fuzzy logic | |
| Chen et al. | Environmental sensors-based occupancy estimation in buildings via IHMM-MLR | |
| CN113687609A (en) | Intelligent monitoring system and monitoring method for Internet of things applied to abnormal environment | |
| KR102313465B1 (en) | Multi-sensor based air quality status mobile notification system and method | |
| Rastogi et al. | Context-aware monitoring and control of ventilation rate in indoor environments using internet of things | |
| Yao et al. | Establishing an ai model on data sensing and prediction for smart home environment control based on labview | |
| Zhang et al. | Design of building environment detection system for architectures based on internet of things | |
| Saini et al. | A novel application of fuzzy inference system optimized with particle swarm optimization and genetic algorithm for PM10 prediction | |
| Mathur et al. | High resolution air pollution mapping using wireless sensor nodes | |
| Pitt et al. | A sensor network for predicting and maintaining occupant comfort | |
| CN113758520A (en) | Internet of things laboratory environment monitoring method and system based on neural network | |
| CN112070317A (en) | Hotel air conditioner energy consumption prediction method | |
| Vesselenyi et al. | Research equipment and computation techniques for Intelligent Building applications | |
| Ramanujam et al. | Improving Indoor occupancy estimation using a hybrid CNN-LSTM approach | |
| Wolf et al. | CO2-based grey-box model to estimate airflow rate and room occupancy | |
| Shamang et al. | Carbon Dioxide Sensor Application In Building Energy Efficiency | |
| Yang et al. | Prediction of pollutant diffusion trend using CNN-GRU hybrid network in IoT air monitoring 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 |