Public building intelligent control system based on LoRa communication technology

Abstract

The invention discloses a public building intelligent control system based on a LoRa communication technology, and relates to the technical field of intelligent control, the public building intelligent control system comprises a LoRa transmission data sensing module, a system performance evaluation and optimization module, a public building intelligent control module and an execution module, the LoRa transmission data sensing module is used for collecting LoRa transmission data, and the system performance evaluation and optimization module is used for evaluating and optimizing the LoRa transmission data; the system performance evaluation and optimization module is used for designing a LoRa communication intelligent control system and evaluating and optimizing the system performance, the public building intelligent control module is used for intelligently controlling the energy management, safety monitoring, environment monitoring and intelligent parking process of a public building, and the execution module is used for executing the LoRa communication intelligent control system. The system is used for realizing the final intelligent control of the public building, optimizing the problems of system transmission delay, operation stability, network interference and the like, achieving the intelligent control of energy consumption, safety state, indoor environment quality and parking resource scheduling, and promoting the development of the system in the green, energy-saving, efficient, safe, comfortable and convenient directions.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an intelligent control system for public buildings based on LoRa communication technology. Background Technology

[0002] With the advancement of urbanization, the management of public buildings faces increasing challenges, such as growing demands for energy efficiency management, security, and environmental monitoring. Traditional building management systems often rely on wired networks and centralized control methods, leading to complex wiring, high maintenance costs, and poor real-time response capabilities. To address these issues, wireless intelligent control systems based on LoRa communication technology are gaining attention. LoRa is a low-power, long-range wireless communication technology characterized by long-distance transmission, high anti-interference capabilities, and low energy consumption. In public buildings, LoRa technology can effectively enable remote control of intelligent devices and data acquisition. Reducing the high costs and complexity of traditional wiring, LoRa technology, combined with emerging technologies such as the Internet of Things and artificial intelligence, enables smart buildings to not only achieve remote management but also optimize internal environmental control and energy consumption through data analysis. For example, by combining data collected by LoRa sensors with AI algorithms, intelligent systems can automatically adjust lighting, air conditioning, and security equipment in the building based on real-time conditions, thereby improving the building's energy efficiency and living comfort. Against this backdrop, intelligent control systems for public buildings based on LoRa technology have demonstrated their enormous potential, providing practical technical support for the green, energy-saving, and intelligent management of future smart buildings.

[0003] Traditional intelligent control technologies for public buildings largely rely on manual control, resulting in poor flexibility, incomplete control, and low levels of intelligence. This invention designs an intelligent control system for public buildings based on LoRa communication technology. It continuously monitors and improves the system to address potential issues such as data transmission delays, unstable long-term equipment operation, network interference, and command loss caused by adverse weather conditions affecting the reference public building environment. This enables intelligent control of the reference public building in areas such as energy management, security monitoring, environmental monitoring, and intelligent parking. Summary of the Invention

[0004] To achieve the above objectives, the present invention is implemented through the following technical solution: a public building intelligent control system based on LoRa communication technology, comprising a LoRa data transmission sensing module, a system performance evaluation and optimization module, a public building intelligent control module, and an execution module, wherein the various modules in the public building intelligent control system based on LoRa communication technology are communicatively connected; The LoRa transmission data sensing module is used to collect LoRa transmission data, which includes system performance monitoring data, public building meteorological data, energy management data, security monitoring data, environmental monitoring data, and smart parking data, enabling real-time and accurate collection of data from all dimensions. The system performance evaluation and optimization module designs an intelligent control system for LoRa communication. Based on system performance monitoring data and meteorological data of public buildings, it uses a neural network algorithm to construct a system performance evaluation model, outputs system performance evaluation coefficients, and adjusts and optimizes the performance of the intelligent control system for LoRa communication according to the system performance evaluation coefficients, providing an intelligent approach to ensure the long-term stable and efficient operation of the LoRa communication system. The intelligent control module for public buildings is used to intelligently control the energy management, security monitoring, environmental monitoring, and intelligent parking processes of public buildings. The intelligent control module for public buildings is divided into an energy management unit, a security monitoring unit, an environmental monitoring unit, and a comprehensive command distribution unit, realizing integrated and refined intelligent management and control of public buildings in multiple scenarios. The energy management unit constructs an energy management monitoring model using energy management data and a random forest algorithm, and obtains an energy management reference index, providing a scientific and accurate basis for decision-making on energy-efficient utilization. The safety monitoring unit uses a convolutional neural network to extract features from safety monitoring data to obtain safety monitoring feature data. It then uses a convolutional neural network algorithm to construct a safety monitoring analytical model, obtain a safety monitoring assessment index, and realize the identification and quantitative assessment of safety risks in public buildings. The environmental monitoring unit combines recurrent neural network algorithms with environmental monitoring data to construct an indoor environmental diagnostic model and output an indoor environmental diagnostic index, providing comprehensive and reliable diagnostic support for optimizing indoor environmental quality. The integrated command allocation unit, based on energy management reference index, safety monitoring and assessment index, indoor environment diagnosis index and intelligent parking data, uses ant colony algorithm to construct intelligent control model of public buildings, and then generates intelligent control commands for public buildings, realizing the accurate output of control commands driven by multi-dimensional data; The execution module integrates and deploys the constructed model and generated instructions to obtain the final intelligent control system for public buildings based on LoRa communication technology. It then executes the intelligent control instructions for public buildings to achieve intelligent control of public buildings, thus realizing closed-loop management and efficient implementation of intelligent control of public buildings based on LoRa communication technology.

[0005] A further improvement to the technical solution of this invention lies in that, in the LoRa transmission data sensing module, the LoRa transmission data acquisition process includes: Deploy different types of data acquisition devices to collect LoRa transmitted data. These devices include network probes, network splitters, Tuptime tools, temperature acquisition devices, LoRa IO acquisition and control terminals, RHF4T003 signal quality testing instruments, LoRa transceivers, temperature and humidity sensors, ultrasonic wind speed sensors, piezoelectric rain sensors, beta-ray absorption dust monitors, light intensity sensors, sound intensity meters, smart meters, smart water meters, smart gas meters, full-color cameras, smart temperature fire detectors, smart smoke sensors, pipeline temperature and humidity sensors, particulate matter sensors, formaldehyde sensors, volatile organic compound sensors, nitrogen oxide sensors, ozone sensors, carbon dioxide sensors, noise sensors, space sensors, flow monitoring sensors, GPS receivers, and parking space sensing devices. System performance monitoring data includes the number of retransmissions of each data packet in the network layer, the total number of data packets transmitted, the number of lost packets, network bandwidth utilization, the running time and real-time temperature of the acquisition equipment, the number of failures of the LoRa communication intelligent control system, the RSSI value, SNR value and LQI value of the LoRa gateway, and the receiving sensitivity of the LoRa node. Among them, the RSSI value, SNR value and LQI value of the LoRa gateway are the signal strength, signal-to-noise ratio and link quality of the LoRa gateway, respectively. Meteorological data for public buildings includes real-time outdoor temperature, humidity, wind speed, precipitation, and dust concentration; energy management data includes real-time indoor light intensity, temperature, humidity, sound intensity, actual electricity consumption, water supply, and gas supply; security monitoring data includes real-time indoor monitoring images, flame temperature, smoke concentration, and water pipe surface humidity; environmental monitoring data includes real-time indoor air concentrations of PM2.5, PM10, formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide, real-time noise intensity and noise source frequency, total indoor area, and traffic flow; smart parking data includes the latitude and longitude coordinates, number, and real-time usage status of each parking space, as well as the real-time latitude and longitude coordinates of the customer's vehicle. The real-time usage status of each parking space includes whether it is vacant or busy. The system performs image denoising and image enhancement on the real-time monitoring images of the public building interiors. It also performs data cleaning and normalization on the collected system performance monitoring data, public building meteorological data, energy management data, smoke concentration, real-time humidity of water pipe surfaces, environmental monitoring data, and smart parking data. Finally, it integrates the pre-processed LoRa transmission data to generate a public building intelligent control dataset.

[0006] A further improvement to the technical solution of this invention lies in the fact that, in the system performance evaluation and optimization module, the process of designing a LoRa communication intelligent control system, constructing a system performance evaluation model, and outputting system performance evaluation coefficients includes: The system performance evaluation coefficients include the transmission delay coefficient, operational stability coefficient, and network interference coefficient of the LoRa communication intelligent control system; Design an intelligent control system for LoRa communication, which consists of a perception layer, a network layer, a platform layer, and an application layer; The LoRa data transmission sensing module is used as the sensing layer, and a LoRa wireless module is provided for the sensing layer. A network layer is set up, and a LoRa gateway and LoRaWAN protocol are provided for the network layer. The intelligent control module for public buildings is used as the platform layer, and the execution module is used as the application layer. The LoRa transmission data collected by the sensing layer is sent to the LoRa gateway via the LoRa wireless module and then transmitted to the platform layer via the LoRaWAN protocol. The platform layer stores, processes, analyzes, and outputs instructions to the LoRa transmission data, and the output instructions are executed by the application layer. Calculate the percentage of lost packets in the total number of packets to obtain the packet loss rate of the transmitted data. Use the packet loss rate of the transmitted data to replace the total number of transmitted data packets and the number of lost packets in the system performance monitoring data, thereby realizing the update of the intelligent control dataset for public buildings. System performance monitoring data and meteorological data of public buildings are extracted from the intelligent control dataset of public buildings and converted into a first training set and a first test set, wherein the ratio of the first training set data to the first test set data is 8:2. Using a neural network algorithm, the first training set data is used as input, and the system performance evaluation coefficients are used as output. The relationships between the first neural network, the second neural network, and the third neural network are learned to obtain a trained system performance evaluation model. The first test set data is input into the system performance evaluation model. An adaptive moment estimator optimizer is used to adjust the parameters of the system performance evaluation model, optimize the performance of the system performance evaluation model, and obtain the final system performance evaluation model. The current system performance monitoring data and public building meteorological data are input into the final system performance evaluation model to obtain the transmission delay coefficient, operation stability coefficient and network interference coefficient of the LoRa communication intelligent control system.

[0007] A further improvement to the technical solution of this invention lies in that, in the system performance evaluation and optimization module, the adjustment and optimization process of the LoRa communication intelligent control system in terms of performance includes: A transmission delay coefficient threshold is set. When the transmission delay coefficient of the LoRa intelligent communication control system is lower than or equal to the transmission delay coefficient threshold, no adjustment or optimization is performed. When the transmission delay coefficient of the LoRa intelligent communication control system is higher than the transmission delay coefficient threshold, the transmission delay of the LoRa intelligent communication control system is adjusted and optimized by increasing the channel bandwidth of the LoRa device, reducing the size of each transmitted data packet, and optimizing the network topology. A threshold for the operational stability coefficient is set. When the operational stability coefficient of the LoRa intelligent communication control system is higher than or equal to the threshold, no adjustment or optimization is performed. When the operational stability coefficient of the LoRa intelligent communication control system is lower than the threshold, adjustments and optimizations are made to the operational stability of the LoRa intelligent communication control system by optimizing network load, adjusting the transmission power of LoRa nodes, managing signal strength, introducing an automatic retransmission mechanism, and designing a self-recovery mechanism for nodes in the network. A network interference coefficient threshold is set. When the network interference coefficient of the LoRa intelligent communication control system is lower than or equal to the threshold, no adjustment or optimization is performed. When the network interference coefficient of the LoRa intelligent communication control system is higher than the threshold, LoRaWAN frequency hopping technology is used to achieve rapid switching between multiple frequencies, effectively avoiding frequency interference and improving the anti-interference capability of communication. Signal redundancy is increased to enhance anti-interference capability. The transmit power and receive sensitivity of LoRa devices are adjusted to ensure that the devices can still successfully receive and send data in a low interference environment. Network density is reduced to avoid power frequency interference, and appropriate modulation methods are selected to achieve adjustment and optimization of the LoRa intelligent communication control system in terms of network interference.

[0008] A further improvement to the technical solution of this invention lies in that, within the energy management unit, the process of constructing an energy management monitoring model and obtaining an energy management reference index includes: The energy management reference index includes the smart lighting management index, the air conditioning and heating control index, and the energy balance monitoring index; Energy management data was extracted from the intelligent control dataset of public buildings and divided into a second training set and a second test set, with the ratio of the second training set data to the second test set data being 7:3. Using the random forest algorithm, the second training set data is set as input and the energy management reference index is set as output. The first random forest relationship, the second random forest relationship and the third random forest relationship are learned to obtain the trained energy management monitoring model. The second test set data is input into the trained energy management monitoring model. The parameters of the energy management monitoring model are adjusted using a stochastic gradient descent optimizer to optimize the performance of the energy management monitoring model and obtain the final energy management monitoring model. By combining current energy management data with the final energy management monitoring model, a corresponding energy management reference index is output.

[0009] A further improvement to the technical solution of this invention lies in the fact that, in the security monitoring unit, the process of extracting features from security monitoring data to obtain security monitoring feature data, constructing a security monitoring analytical model, and obtaining a security monitoring evaluation index includes: Safety monitoring feature data includes the flame area and flame RGB value inside public buildings, the water leakage area and water level inside public buildings, and the frequency of people entering and exiting and the time spent in real-time monitoring images inside public buildings, providing core basis for safety monitoring inside public buildings. Convolutional neural networks are used to extract features from real-time monitoring images of public building interiors. The convolutional neural network includes convolutional layers, pooling layers, and fully connected layers, which builds an efficient image feature extraction technology framework that is suitable for security monitoring scenarios. Among them, the convolutional layer uses several convolutional kernels to extract the edges, corners and textures of real-time monitoring images of public building interiors, and automatically learns to identify relevant features of security monitoring feature data to obtain security monitoring feature maps, thereby achieving accurate capture and preliminary screening of key visual information of real-time monitoring images of public building interiors. The pooling layer reduces the size of the security monitoring feature map by using average pooling, thus preserving the important information of the security monitoring image. Using a fully connected layer as the last layer of a convolutional neural network, and based on the processing of convolutional and pooling layers, flame and water leakage areas in real-time monitoring images of public building interiors are identified and classified. Mask images of flame and water leakage areas are obtained respectively. Combined with OpenCV contour detection method, the area of ​​flame, RGB value of flame, area of ​​water leakage and water level of public building interiors are obtained. Then, using target detection algorithm, people in real-time monitoring images of public building interiors are identified. By analyzing the changes in the position of people in several image frames and combining with tracking algorithm, the frequency of people entering and leaving and the dwell time of people in real-time monitoring images of public building interiors are obtained. The convolutional neural network obtains the final safety monitoring feature data through several convolutional layers, pooling layers, and fully connected layers, and integrates the safety monitoring feature data into the intelligent control dataset of public buildings, achieving accurate extraction of multi-dimensional safety monitoring feature data and providing comprehensive and reliable data support for indoor safety monitoring of Honggong buildings.

[0010] A further improvement to the technical solution of this invention lies in that, in the security monitoring unit, the process of constructing a security monitoring analysis model and obtaining a security monitoring evaluation index includes: The safety monitoring assessment index includes the fire monitoring index, access control monitoring index, and water leakage monitoring index; The safety monitoring feature data, real-time flame temperature, smoke concentration and water pipe surface humidity of the public building interior were extracted from the intelligent control dataset of public buildings and converted into a third training set and a third test set in a 6:4 ratio. Using a convolutional neural network algorithm, the third training set data is used as input and the security monitoring evaluation index is used as output. The relationship between the first, second and third convolutional neural networks is learned. Through learning the relationship between the convolutional neural networks, a trained security monitoring analysis model is obtained. The third test set data is input into the trained security monitoring and parsing model, the parameters of the security monitoring and parsing model are adjusted, the performance of the security monitoring and parsing model is optimized, and the final security monitoring and parsing model is obtained. The current safety monitoring characteristic data, real-time flame temperature, smoke concentration, and real-time humidity of water pipe surfaces in public buildings are input into the final safety monitoring analysis model to obtain the safety monitoring evaluation index.

[0011] A further improvement to the technical solution of this invention lies in that, in the environmental monitoring unit, the process of constructing an indoor environmental diagnostic model and outputting an indoor environmental diagnostic index includes: The indoor environmental diagnostic index includes the air quality diagnostic index and the noise diagnostic index. Environmental monitoring data is extracted from the intelligent control dataset of public buildings and divided into a fourth training set and a fourth test set, with the ratio of the fourth training set to the fourth test set being 7:3. Using a recurrent neural network algorithm, environmental monitoring data is set as input data and indoor environmental diagnostic index is set as output data. The relationship between the first recurrent neural network and the relationship between the second recurrent neural network are learned to obtain a trained indoor environmental diagnostic model. The fourth test set data is input into the trained indoor environment diagnostic model. The parameters of the indoor environment diagnostic model are adjusted to optimize its performance and obtain the final indoor environment diagnostic model. Combined with the current environmental monitoring data, the indoor environment diagnostic index is output.

[0012] A further improvement to the technical solution of this invention lies in that, in the integrated instruction allocation unit, the process of constructing a public building intelligent control model and generating intelligent control instructions for public buildings includes: The intelligent control commands for public buildings include energy management commands, safety management commands, indoor environment optimization commands, and intelligent parking guidance commands. Among them, energy management commands consist of intelligent lighting commands, air conditioning and heating control commands, and energy balance allocation commands; safety management commands consist of fire control commands, access control and security control commands, and water leakage control commands; and indoor environment optimization commands consist of air quality optimization commands and noise control commands. The energy management reference index, safety monitoring and assessment index, indoor environment diagnosis index and smart parking data are used as input data, and the intelligent control commands for public buildings are used as output data. The ant colony algorithm is used to initialize the number and initial position of individual ants, set decision nodes related to intelligent control commands for public buildings, and initialize pheromones at each decision node. Each individual ant represents a command output, and the pheromone represents the priority of each path selection decision. Each public building's intelligent control command and intelligent parking data sets a path selection standard. Based on the set path selection standard, ants leave pheromones at each decision node. According to the pheromone concentration and heuristic function, the optimal decision path is selected to complete the construction of the public building intelligent control model. Based on the objective function in the intelligent control model of public buildings, the quality of each better decision path is evaluated. Through multiple iterations, the pheromone on each decision node is updated so that the better decision path gradually approaches the optimal solution. The best decision path is then selected to obtain the intelligent control command for public buildings.

[0013] A further improvement to the technical solution of this invention lies in that, in the execution module, the process of obtaining the final intelligent control system for public buildings based on LoRa communication technology and realizing intelligent control of public buildings includes: The system performance evaluation model, energy management monitoring model, safety monitoring analysis model, indoor environment diagnosis model, and public building intelligent control model are integrated and deployed into the platform layer of the LoRa communication intelligent control system to obtain the final public building intelligent control system based on LoRa communication technology. Based on the LoRa transmission data acquired by the perception layer, the platform layer adjusts and optimizes the performance of the LoRa communication intelligent control system. Through the model deployed in the platform layer, the control commands in the energy management, security monitoring, environmental monitoring and intelligent parking processes of public buildings are analyzed, and finally the intelligent control commands of public buildings are output. Then, the intelligent control commands of public buildings are wirelessly transmitted to the application layer through the network layer. The application layer executes the intelligent control commands of each public building to realize the intelligent control of public buildings.

[0014] This invention presents an intelligent control system for public buildings based on LoRa communication technology. Compared to traditional intelligent control systems for public buildings, this system integrates LoRa wireless communication technology, neural network algorithms, random forest algorithms, convolutional neural network algorithms, recurrent neural network algorithms, and ant colony algorithms with modern information technology. Through various acquisition devices in the LoRa data transmission sensing module, it accurately captures multi-dimensional data, further obtaining energy management reference indices, safety monitoring assessment indices, indoor environmental diagnostic indices, and intelligent control commands for public buildings. This achieves real-time and comprehensive monitoring of energy consumption, safety status, indoor environmental quality, and parking resource scheduling in public buildings. During this process, a system performance evaluation and optimization module constructs relevant models to specifically optimize issues such as transmission delay, operational stability, and network interference. Finally, the execution module completes the entire process. The implementation of this method solves the problems of complex wiring, high maintenance costs, and poor real-time response capabilities caused by traditional systems relying on wired networks, as well as the lack of flexibility, incomplete control, and low level of intelligence in manual control. Furthermore, it addresses the difficulties in coping with data transmission delays, unstable equipment operation, network interference and command loss caused by severe weather. This ensures that the method in this invention can refine the dynamic monitoring standards for intelligent control systems of public buildings based on LoRa communication technology within a more precise range. Under the same conditions, the monitored system performance data, energy data, safety data, environmental data, and parking data become more accurate indicators. The development and application of this method significantly enhances the intelligence level in the intelligent control process of public buildings based on LoRa communication technology, realizing integrated and refined intelligent management and control of public buildings across multiple scenarios, and gradually promoting the upgrading of building management towards green, energy-saving, efficient, safe, comfortable, and convenient directions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a block diagram of the intelligent control system for public buildings based on LoRa communication technology according to the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, the present invention provides an intelligent control system for public buildings based on LoRa communication technology, including a LoRa data transmission sensing module, a system performance evaluation and optimization module, an intelligent control module for public buildings, and an execution module, wherein the various modules in the intelligent control system for public buildings based on LoRa communication technology are communicatively connected. The LoRa transmission data sensing module is used to collect LoRa transmission data, which includes system performance monitoring data, public building meteorological data, energy management data, security monitoring data, environmental monitoring data, and smart parking data. It realizes the comprehensive collection and integration of data related to the intelligent control of public buildings, providing complete and reliable data source support for subsequent modules. The system performance evaluation and optimization module designs an intelligent control system for LoRa communication. Based on system performance monitoring data and meteorological data of public buildings, a neural network algorithm is used to construct a system performance evaluation model and output system performance evaluation coefficients. Based on the system performance evaluation coefficients, the performance of the intelligent control system for LoRa communication is adjusted and optimized. Ultimately, the adaptive optimization of LoRa communication in the intelligent control process of complex public buildings is realized, ensuring that the overall network is in a stable and efficient operating state for a long time. The intelligent control module for public buildings is used to intelligently control the energy management, security monitoring, environmental monitoring, and intelligent parking processes of public buildings. The intelligent control module for public buildings is divided into an energy management unit, a security monitoring unit, an environmental monitoring unit, and a comprehensive command distribution unit, realizing integrated and refined intelligent management and control of public buildings in multiple scenarios. The energy management unit constructs an energy management monitoring model using energy management data and a random forest algorithm, and obtains an energy management reference index. This provides a scientific basis for decision-making to achieve on-demand allocation, peak-shifting regulation, and balance monitoring of building energy, thereby maximizing energy utilization efficiency. The safety monitoring unit uses convolutional neural networks to extract features from safety monitoring data to obtain safety monitoring feature data. It further employs convolutional neural network algorithms to construct a safety monitoring analytical model and obtain a safety monitoring evaluation index, thereby realizing the transformation from passively receiving alarms to actively identifying hidden dangers and achieving accurate quantitative assessment and early warning of safety risks. The environmental monitoring unit combines recurrent neural network algorithms with environmental monitoring data to construct an indoor environmental diagnostic model and output an indoor environmental diagnostic index, providing comprehensive and reliable diagnostic support for ensuring the health and comfort of the indoor environment. The integrated command allocation unit, based on energy management reference index, safety monitoring and assessment index, indoor environment diagnosis index and intelligent parking data, uses ant colony algorithm to construct intelligent control model of public buildings, and then generates intelligent control commands for public buildings to achieve coordinated response to multi-scenario needs and ensure the scientific nature of control commands; The execution module integrates and deploys the constructed model and generated instructions to obtain the final intelligent control system for public buildings based on LoRa communication technology. It then executes the intelligent control instructions for public buildings to achieve intelligent control of public buildings, thus promoting the efficient implementation of intelligent control of public buildings based on LoRa communication technology from scheme design to practical application.

[0019] In some embodiments, the LoRa transmission data acquisition process in the LoRa transmission data sensing module includes: Deploy different types of data acquisition devices to collect LoRa transmitted data. These devices include network probes, network splitters, Tuptime tools, temperature acquisition devices, LoRa IO acquisition and control terminals, RHF4T003 signal quality testing instruments, LoRa transceivers, temperature and humidity sensors, ultrasonic wind speed sensors, piezoelectric rain sensors, beta-ray absorption dust monitors, light intensity sensors, sound intensity meters, smart meters, smart water meters, smart gas meters, full-color cameras, smart temperature fire detectors, smart smoke sensors, pipeline temperature and humidity sensors, particulate matter sensors, formaldehyde sensors, volatile organic compound sensors, nitrogen oxide sensors, ozone sensors, carbon dioxide sensors, noise sensors, space sensors, flow monitoring sensors, GPS receivers, and parking space sensing devices. System performance monitoring data includes the number of retransmissions of each data packet in the network layer, the total number of data packets transmitted, the number of lost packets, network bandwidth utilization, the running time and real-time temperature of the acquisition equipment, the number of failures of the LoRa communication intelligent control system, the RSSI value, SNR value and LQI value of the LoRa gateway, and the receiving sensitivity of the LoRa node. Among them, the RSSI value, SNR value and LQI value of the LoRa gateway are the signal strength, signal-to-noise ratio and link quality of the LoRa gateway, respectively. Meteorological data for public buildings includes real-time outdoor temperature, humidity, wind speed, precipitation, and dust concentration; energy management data includes real-time indoor light intensity, temperature, humidity, sound intensity, actual electricity consumption, water supply, and gas supply; security monitoring data includes real-time indoor monitoring images, flame temperature, smoke concentration, and water pipe surface humidity; environmental monitoring data includes real-time indoor air concentrations of PM2.5, PM10, formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide, real-time noise intensity and noise source frequency, total indoor area, and traffic flow; smart parking data includes the latitude and longitude coordinates, number, and real-time usage status of each parking space, as well as the real-time latitude and longitude coordinates of the customer's vehicle. The real-time usage status of each parking space includes whether it is vacant or busy. Specifically, network probes are used to collect the retransmission count of each data packet in the network layer. A network splitter is used to collect the total number of transmitted data packets, the number of lost packets, and the network bandwidth utilization. Tuptime tool and temperature acquisition device are used to obtain the running time and real-time temperature of the acquisition device. LoRa IO acquisition control terminal is used to record the number of failures of LoRa communication intelligent control system. Integrated signal quality testing instruments and LoRa transceivers are used to collect the RSSI value, SNR value, LQI value of LoRa gateway and the receiving sensitivity of LoRa node. Specifically, by combining temperature and humidity sensors, ultrasonic wind speed sensors, piezoelectric rain sensors, and beta-ray absorption dust monitors, real-time temperature, real-time humidity, real-time wind speed, real-time precipitation, and real-time dust concentration of the outdoor environment of public buildings are collected. Specifically, by combining full-color cameras, intelligent temperature fire detectors, intelligent smoke sensors, and pipe-type temperature and humidity sensors, real-time monitoring images, real-time flame temperature, smoke concentration, and real-time humidity of water pipe surfaces are collected inside public buildings. Specifically, particulate matter sensors are used to collect real-time concentrations of PM2.5 and PM10 in the indoor air of public buildings. Formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide sensors are used to collect real-time concentrations of these components in the indoor air of public buildings. Noise sensors, spatial sensors, and flow monitoring sensors are used to collect real-time noise intensity and frequency of noise sources in public buildings, as well as the total indoor area and traffic flow in the public building environment. Specifically, each parking space is assigned a number, and a GPS receiver is used to collect the latitude and longitude coordinates of each parking space and the real-time latitude and longitude coordinates of the customer's vehicle. Parking space sensing devices are used to collect the real-time usage status of each parking space. The system performs image denoising and image enhancement on the real-time monitoring images of the public building interiors. It also performs data cleaning and normalization on the collected system performance monitoring data, public building meteorological data, energy management data, smoke concentration, real-time humidity of water pipe surfaces, environmental monitoring data, and smart parking data. Finally, it integrates the pre-processed LoRa transmission data to generate a public building intelligent control dataset.

[0020] In some embodiments, the process of designing a LoRa communication intelligent control system, constructing a system performance evaluation model, and outputting system performance evaluation coefficients in the system performance evaluation and optimization module includes: The system performance evaluation coefficients include the transmission delay coefficient, operational stability coefficient, and network interference coefficient of the LoRa communication intelligent control system; Design an intelligent control system for LoRa communication, which consists of a perception layer, a network layer, a platform layer, and an application layer; The LoRa data transmission sensing module is used as the sensing layer, and a LoRa wireless module is provided for the sensing layer. A network layer is set up, and a LoRa gateway and LoRaWAN protocol are provided for the network layer. The intelligent control module for public buildings is used as the platform layer, and the execution module is used as the application layer. The LoRa transmission data collected by the sensing layer is sent to the LoRa gateway via the LoRa wireless module and then transmitted to the platform layer via the LoRaWAN protocol. The platform layer stores, processes, analyzes, and outputs instructions to the LoRa transmission data, and the output instructions are executed by the application layer. Calculate the percentage of lost packets in the total number of packets to obtain the packet loss rate of the transmitted data. Use the packet loss rate of the transmitted data to replace the total number of transmitted data packets and the number of lost packets in the system performance monitoring data, thereby realizing the update of the intelligent control dataset for public buildings. System performance monitoring data and meteorological data of public buildings are extracted from the intelligent control dataset of public buildings and converted into a first training set and a first test set, wherein the ratio of the first training set data to the first test set data is 8:2. A neural network algorithm is used, taking the replaced system performance monitoring data and public building meteorological data as input, and the transmission delay coefficient, operational stability coefficient, and network interference coefficient of the LoRa communication intelligent control system as output. The algorithm learns the first, second, and third neural network relationships to obtain a trained system performance evaluation model. The first neural network relationship represents the nonlinear relationship between public building meteorological data, the number of retransmissions of each data packet in the network layer, the packet loss rate of transmitted data, and the transmission delay coefficient of the LoRa communication intelligent control system. The second neural network relationship represents the nonlinear relationship between public building meteorological data, the runtime and real-time temperature of the acquisition equipment, and the operational stability coefficient of the LoRa communication intelligent control system. The third neural network relationship represents the nonlinear relationship between public building meteorological data, the number of failures in the LoRa communication intelligent control system, the RSSI, SNR, and LQI values ​​of the LoRa gateway, the receiving sensitivity of the LoRa node, the packet loss rate of transmitted data, and the network interference coefficient of the LoRa communication intelligent control system. The first test set data is input into the system performance evaluation model. An adaptive moment estimator optimizer is used to adjust the parameters of the system performance evaluation model, optimize the performance of the system performance evaluation model, and obtain the final system performance evaluation model. The current system performance monitoring data and public building meteorological data are input into the final system performance evaluation model to obtain the transmission delay coefficient, operation stability coefficient and network interference coefficient of the LoRa communication intelligent control system.

[0021] In some embodiments, the performance adjustment and optimization process of the LoRa communication intelligent control system in the system performance evaluation and optimization module includes: A transmission delay coefficient threshold is set. When the transmission delay coefficient of the LoRa intelligent communication control system is lower than or equal to the transmission delay coefficient threshold, no adjustment or optimization is performed. When the transmission delay coefficient of the LoRa intelligent communication control system is higher than the transmission delay coefficient threshold, the transmission delay of the LoRa intelligent communication control system is adjusted and optimized by increasing the channel bandwidth of the LoRa device, reducing the size of each transmitted data packet, and optimizing the network topology. A threshold for the operational stability coefficient is set. When the operational stability coefficient of the LoRa intelligent communication control system is higher than or equal to the threshold, no adjustment or optimization is performed. When the operational stability coefficient of the LoRa intelligent communication control system is lower than the threshold, adjustments and optimizations are made to the operational stability of the LoRa intelligent communication control system by optimizing network load, adjusting the transmission power of LoRa nodes, managing signal strength, introducing an automatic retransmission mechanism, and designing a self-recovery mechanism for nodes in the network. A network interference coefficient threshold is set. When the network interference coefficient of the LoRa intelligent communication control system is lower than or equal to the threshold, no adjustment or optimization is performed. When the network interference coefficient of the LoRa intelligent communication control system is higher than the threshold, LoRaWAN frequency hopping technology is used to achieve rapid switching between multiple frequencies, effectively avoiding frequency interference and improving the anti-interference capability of communication. Signal redundancy is increased to enhance anti-interference capability. The transmit power and receive sensitivity of LoRa devices are adjusted to ensure that the devices can still successfully receive and send data in a low interference environment. Network density is reduced to avoid power frequency interference, and appropriate modulation methods are selected to achieve adjustment and optimization of the LoRa intelligent communication control system in terms of network interference.

[0022] In some embodiments, the process of constructing an energy management monitoring model and obtaining an energy management reference index in the energy management unit includes: The energy management reference index includes the smart lighting management index, the air conditioning and heating control index, and the energy balance monitoring index; Energy management data was extracted from the intelligent control dataset of public buildings and divided into a second training set and a second test set, with the ratio of the second training set data to the second test set data being 7:3. Using the random forest algorithm, the second training set data is set as input, and the energy management reference index is set as output. The first random forest relationship, the second random forest relationship, and the third random forest relationship are learned to obtain the trained energy management monitoring model. The first random forest relationship is the nonlinear relationship between the real-time light intensity and real-time sound intensity in the public building and the intelligent lighting management index; the second random forest relationship is the nonlinear relationship between the real-time temperature and real-time humidity in the public building and the air conditioning and heating control index; and the third random forest relationship is the nonlinear relationship between the actual electricity consumption, actual water supply, and actual gas supply in the public building and the energy balance monitoring index. The second test set data is input into the trained energy management monitoring model. The parameters of the energy management monitoring model are adjusted using a stochastic gradient descent optimizer to optimize the performance of the energy management monitoring model and obtain the final energy management monitoring model. By combining current energy management data with the final energy management monitoring model, a corresponding energy management reference index is output.

[0023] In some embodiments, the process of extracting features from security monitoring data to obtain security monitoring feature data, constructing a security monitoring analysis model, and obtaining a security monitoring evaluation index in the security monitoring unit includes: Safety monitoring feature data includes the flame area and flame RGB value inside public buildings, the water leakage area and water level inside public buildings, and the frequency of people entering and exiting and the time spent in real-time monitoring images inside public buildings, providing core basis for safety monitoring inside public buildings. Convolutional neural networks are used to extract features from real-time monitoring images of public building interiors. The convolutional neural network includes convolutional layers, pooling layers, and fully connected layers, which builds an efficient image feature extraction technology framework that is suitable for security monitoring scenarios. Among them, the convolutional layer uses several convolutional kernels to extract the edges, corners and textures of real-time monitoring images of public building interiors, and automatically learns to identify relevant features of security monitoring feature data to obtain security monitoring feature maps, thereby achieving accurate capture and preliminary screening of key visual information of real-time monitoring images of public building interiors. The pooling layer reduces the size of the security monitoring feature map by using average pooling, thus preserving the important information of the security monitoring image. Using a fully connected layer as the last layer of a convolutional neural network, and based on the processing of convolutional and pooling layers, flame and water leakage areas in real-time monitoring images of public building interiors are identified and classified. Mask images of flame and water leakage areas are obtained respectively. Combined with OpenCV contour detection method, the area of ​​flame, RGB value of flame, area of ​​water leakage and water level of public building interiors are obtained. Then, using target detection algorithm, people in real-time monitoring images of public building interiors are identified. By analyzing the changes in the position of people in several image frames and combining with tracking algorithm, the frequency of people entering and leaving and the dwell time of people in real-time monitoring images of public building interiors are obtained. The convolutional neural network obtains the final safety monitoring feature data through several convolutional layers, pooling layers, and fully connected layers, and integrates the safety monitoring feature data into the intelligent control dataset of public buildings, achieving accurate extraction of multi-dimensional safety monitoring feature data and providing comprehensive and reliable data support for indoor safety monitoring of Honggong buildings.

[0024] In some embodiments, the process of constructing a security monitoring analysis model and obtaining a security monitoring evaluation index in the security monitoring unit includes: The safety monitoring assessment index includes the fire monitoring index, access control monitoring index, and water leakage monitoring index; The safety monitoring feature data, real-time flame temperature, smoke concentration and water pipe surface humidity of the public building interior were extracted from the intelligent control dataset of public buildings and converted into a third training set and a third test set in a 6:4 ratio. A convolutional neural network (CNN) algorithm is employed, using the third training set as input and the safety monitoring assessment index as output. It learns the relationships between the first, second, and third CNNs. Through learning these CNN relationships, a trained safety monitoring analytical model is obtained. Specifically, the first CNN relationship describes the nonlinear relationship between the flame area, flame RGB value, real-time flame temperature, and smoke concentration inside the public building and the fire monitoring index; the second CNN relationship describes the nonlinear relationship between the frequency of personnel entering and exiting and their dwell time in real-time monitoring images inside the public building and the access control monitoring index; and the third CNN relationship describes the nonlinear relationship between the water leakage area, water level, and real-time humidity of water pipe surfaces inside the public building and the water leakage monitoring index. The third test set data is input into the trained security monitoring and parsing model, the parameters of the security monitoring and parsing model are adjusted, the performance of the security monitoring and parsing model is optimized, and the final security monitoring and parsing model is obtained. The current safety monitoring characteristic data, real-time flame temperature, smoke concentration, and real-time humidity of water pipe surfaces in public buildings are input into the final safety monitoring analysis model to obtain the safety monitoring evaluation index.

[0025] In some embodiments, the process of constructing an indoor environmental diagnostic model and outputting an indoor environmental diagnostic index in the environmental monitoring unit includes: The indoor environmental diagnostic index includes the air quality diagnostic index and the noise diagnostic index. Environmental monitoring data is extracted from the intelligent control dataset of public buildings and divided into a fourth training set and a fourth test set, with the ratio of the fourth training set to the fourth test set being 7:3. A recurrent neural network algorithm is used to take the real-time concentrations of PM2.5, PM10, formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide in the indoor air of public buildings, the real-time noise intensity and the real-time frequency of noise sources, the total indoor area of ​​public buildings, and the traffic flow in the public building environment as input data, and the air quality diagnostic index and noise diagnostic index as output data. The algorithm learns the first recurrent neural network relationship and the second recurrent neural network relationship to obtain a trained indoor environment diagnostic model. The first recurrent neural network relationship is the nonlinear relationship between the real-time concentrations of PM2.5, PM10, formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide in the indoor air of public buildings, the total indoor area of ​​public buildings, and the air quality diagnostic index. The second recurrent neural network relationship is the nonlinear relationship between the real-time noise intensity and the real-time frequency of noise sources in public buildings, the traffic flow in the public building environment, and the noise diagnostic index. The fourth test set data is input into the trained indoor environment diagnostic model. The parameters of the indoor environment diagnostic model are adjusted to optimize its performance and obtain the final indoor environment diagnostic model. Combined with the current environmental monitoring data, the indoor environment diagnostic index is output.

[0026] In some embodiments, the process by which the integrated instruction allocation unit constructs a public building intelligent control model and generates intelligent control instructions for the public building includes: The intelligent control commands for public buildings include energy management commands, safety management commands, indoor environment optimization commands, and intelligent parking guidance commands. Among them, energy management commands consist of intelligent lighting commands, air conditioning and heating control commands, and energy balance allocation commands; safety management commands consist of fire control commands, access control and security control commands, and water leakage control commands; and indoor environment optimization commands consist of air quality optimization commands and noise control commands. The energy management reference index, safety monitoring and assessment index, indoor environment diagnosis index and smart parking data are used as input data, and the intelligent control commands for public buildings are used as output data. The ant colony algorithm is used to initialize the number and initial position of individual ants, set decision nodes related to intelligent control commands for public buildings, and initialize pheromones at each decision node. Each individual ant represents a command output, and the pheromone represents the priority of each path selection decision. Path selection criteria were set for each public building's intelligent control commands and intelligent parking data. Based on these criteria, ants left pheromones at each decision node. According to the pheromone concentration and heuristic function, the optimal decision path was selected to complete the construction of the public building's intelligent control model. The specific path selection criteria are as follows: for intelligent lighting commands and air conditioning / heating control commands, the path selection criterion is low energy consumption; for energy balance allocation commands, the path selection criterion is energy balance supply and demand balance; for fire control commands, the path selection criterion is rapid fire control, efficient evacuation, and prevention of spread; for access control and security control commands, the path selection criterion is accurate identification, hierarchical management, and prevention of intrusion and misjudgment; for water leakage control commands, the path selection criterion is rapid location, accurate source interruption, and reduction of diffusion and loss; for air quality optimization commands, the path selection criterion is rapid achievement of standards, zone adaptation, and energy consumption balance; for noise control commands, the path selection criterion is sound source localization and hierarchical noise reduction; based on intelligent parking data, the optimal decision path was selected based on maximizing parking space utilization, shortening the time drivers spend searching for parking spaces, and avoiding parking area congestion. Based on the objective function in the intelligent control model of public buildings, the quality of each better decision path is evaluated. Through multiple iterations, the pheromone on each decision node is updated so that the better decision path gradually approaches the optimal solution. The best decision path is then selected to obtain the intelligent control command for public buildings.

[0027] In some embodiments, the execution module acquires the final intelligent control system for public buildings based on LoRa communication technology, and the process of realizing intelligent control of public buildings includes: The system performance evaluation model, energy management monitoring model, safety monitoring analysis model, indoor environment diagnosis model, and public building intelligent control model are integrated and deployed into the platform layer of the LoRa communication intelligent control system to obtain the final public building intelligent control system based on LoRa communication technology. Based on the LoRa transmission data acquired by the perception layer, the platform layer adjusts and optimizes the performance of the LoRa communication intelligent control system. Through the model deployed in the platform layer, the control commands in the energy management, security monitoring, environmental monitoring and intelligent parking processes of public buildings are analyzed, and finally the intelligent control commands of public buildings are output. Then, the intelligent control commands of public buildings are wirelessly transmitted to the application layer through the network layer. The application layer executes the intelligent control commands of each public building to realize the intelligent control of public buildings.

[0028] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A public building intelligent control system based on LoRa communication technology, comprising a LoRa data transmission sensing module, a system performance evaluation and optimization module, a public building intelligent control module, and an execution module, wherein, The communication connections between the various modules of the intelligent control system for public buildings based on LoRa communication technology are characterized in that, The LoRa transmission data sensing module is used to collect LoRa transmission data, which includes system performance monitoring data, public building meteorological data, energy management data, security monitoring data, environmental monitoring data, and smart parking data. The system performance evaluation and optimization module designs a LoRa communication intelligent control system. Based on the system performance monitoring data and the meteorological data of the public buildings, it uses a neural network algorithm to construct a system performance evaluation model, outputs system performance evaluation coefficients, and adjusts and optimizes the performance of the LoRa communication intelligent control system according to the system performance evaluation coefficients. The intelligent control module for public buildings is used for intelligent control of energy management, security monitoring, environmental monitoring, and intelligent parking processes in public buildings; the intelligent control module for public buildings is divided into an energy management unit, a security monitoring unit, an environmental monitoring unit, and a comprehensive command distribution unit; The energy management unit constructs an energy management monitoring model and obtains an energy management reference index by using the energy management data and a random forest algorithm. The security monitoring unit uses a convolutional neural network to extract features from the security monitoring data to obtain security monitoring feature data. It further uses a convolutional neural network algorithm to construct a security monitoring analysis model and obtain a security monitoring evaluation index. The environmental monitoring unit, combining the recurrent neural network algorithm with the environmental monitoring data, constructs an indoor environmental diagnostic model and outputs an indoor environmental diagnostic index. The integrated command allocation unit, based on the energy management reference index, the safety monitoring and evaluation index, the indoor environment diagnosis index, and the smart parking data, uses the ant colony algorithm to construct a public building intelligent control model, and then generates intelligent control commands for the public building. The execution module integrates and deploys the constructed model and the generated instructions to obtain the final intelligent control system for public buildings based on LoRa communication technology, and executes the intelligent control instructions for public buildings to achieve intelligent control of public buildings.

2. The intelligent control system for public buildings based on LoRa communication technology according to claim 1, characterized in that, The process of acquiring LoRa transmission data includes: Deploy different types of acquisition devices to collect the LoRa transmitted data. These different types of acquisition devices include network probes, network splitters, Tuptime tools, temperature acquisition devices, LoRa IO acquisition and control terminals, signal quality testing instruments, LoRa transceivers, temperature and humidity sensors, ultrasonic wind speed sensors, piezoelectric rain sensors, beta-ray absorption dust monitors, light intensity sensors, sound intensity meters, smart meters, smart water meters, smart gas meters, full-color cameras, smart temperature fire detectors, smart smoke sensors, pipeline temperature and humidity sensors, particulate matter sensors, formaldehyde sensors, volatile organic compound sensors, nitrogen oxide sensors, ozone sensors, carbon dioxide sensors, noise sensors, space sensors, flow monitoring sensors, GPS receivers, and parking space sensing devices. The system performance monitoring data includes the number of retransmissions of each data packet in the network layer, the total number of data packets transmitted, the number of lost packets, network bandwidth utilization, the running time and real-time temperature of the acquisition device, the number of failures of the LoRa communication intelligent control system, the RSSI value, SNR value and LQI value of the LoRa gateway, and the receiving sensitivity of the LoRa node. The meteorological data for public buildings includes real-time outdoor temperature, humidity, wind speed, precipitation, and dust concentration; the energy management data includes real-time indoor light intensity, temperature, humidity, sound intensity, actual electricity consumption, water supply, and gas supply; the safety monitoring data includes real-time indoor monitoring images, flame temperature, smoke concentration, and water pipe surface humidity; the environmental monitoring data includes real-time indoor air concentrations of PM2.5, PM10, formaldehyde, volatile organic compounds, nitrogen oxides, ozone, and carbon dioxide, real-time noise intensity and noise source frequency, total indoor area, and traffic flow; and the smart parking data includes the latitude and longitude coordinates, number, and real-time usage status of each parking space, as well as the real-time latitude and longitude coordinates of the customer's vehicle. The collected real-time monitoring images of the public building's interior are subjected to image denoising and image enhancement processing. The collected system performance monitoring data, public building meteorological data, energy management data, smoke concentration, real-time humidity of water pipe surfaces, environmental monitoring data, and smart parking data are subjected to data cleaning and data normalization processing. The pre-processed LoRa transmission data is then integrated to generate a public building intelligent control dataset.

3. The intelligent control system for public buildings based on LoRa communication technology according to claim 2, characterized in that, The process of designing a LoRa communication intelligent control system, constructing a system performance evaluation model, and outputting system performance evaluation coefficients includes: The system performance evaluation coefficients include the transmission delay coefficient, operational stability coefficient, and network interference coefficient of the LoRa communication intelligent control system; Design a LoRa communication intelligent control system, which consists of a perception layer, a network layer, a platform layer, and an application layer; The LoRa data transmission sensing module is used as the sensing layer, and the sensing layer is equipped with a LoRa wireless module. The network layer is set up, and the network layer is equipped with a LoRa gateway and LoRaWAN protocol. The intelligent control module for public buildings is used as the platform layer, and the execution module is used as the application layer. The LoRa transmission data collected by the sensing layer is sent to the LoRa gateway through the LoRa wireless module, and then transmitted to the platform layer through the LoRaWAN protocol. The platform layer stores, processes, analyzes, and outputs instructions to the LoRa transmission data, and the output instructions are executed by the application layer. Calculate the percentage of lost packets in the total number of packets in the transmitted data to obtain the packet loss rate of the transmitted data. Use the packet loss rate of the transmitted data to replace the total number of packets and the number of lost packets in the system performance monitoring data, thereby updating the intelligent control dataset of the public building. Extract the system performance monitoring data and the meteorological data of the public building from the intelligent control dataset of the public building, and convert them into a first training set and a first test set; Using the neural network algorithm, the first training set data is taken as input, and the system performance evaluation coefficients are taken as output. The first neural network relationship, the second neural network relationship, and the third neural network relationship are learned to obtain the trained system performance evaluation model. The first test set data is input into the system performance evaluation model. An adaptive moment estimator optimizer is used to adjust the parameters of the system performance evaluation model, optimize the performance of the system performance evaluation model, and obtain the final system performance evaluation model. The current system performance monitoring data and the meteorological data of the public buildings are input into the final system performance evaluation model to obtain the transmission delay coefficient, operation stability coefficient and network interference coefficient of the LoRa communication intelligent control system.

4. The intelligent control system for public buildings based on LoRa communication technology according to claim 3, characterized in that, The performance adjustment and optimization process of the LoRa communication intelligent control system includes: A transmission delay coefficient threshold is set. When the transmission delay coefficient of the LoRa intelligent communication control system is lower than or equal to the transmission delay coefficient threshold, no adjustment or optimization is performed. When the transmission delay coefficient of the LoRa intelligent communication control system is higher than the transmission delay threshold, the transmission delay of the LoRa intelligent communication control system is adjusted and optimized by increasing the channel bandwidth of the LoRa device, reducing the size of each transmitted data packet, and optimizing the network topology. A threshold for the operating stability coefficient is set. When the operating stability coefficient of the LoRa intelligent communication control system is higher than or equal to the threshold, no adjustment or optimization is performed. When the operating stability coefficient of the LoRa intelligent communication control system is lower than the threshold, adjustments and optimizations are made to the operating stability of the LoRa intelligent communication control system by optimizing network load, adjusting the transmission power of LoRa nodes, managing signal strength, introducing an automatic retransmission mechanism, and designing a self-recovery mechanism for nodes in the network. A network interference coefficient threshold is set. When the network interference coefficient of the LoRa communication intelligent control system is lower than or equal to the network interference coefficient threshold, no adjustment or optimization is performed. When the network interference coefficient of the LoRa communication intelligent control system is higher than the network interference coefficient threshold, adjustments and optimizations are made to the network interference of the LoRa communication intelligent control system by using LoRaWAN frequency hopping technology, increasing signal redundancy, adjusting the transmit power and receive sensitivity of LoRa devices, reducing network density, and selecting appropriate modulation methods.

5. The intelligent control system for public buildings based on LoRa communication technology according to claim 4, characterized in that, The process of constructing an energy management monitoring model and obtaining an energy management reference index includes: The energy management reference index includes the intelligent lighting management index, the air conditioning and heating control index, and the energy balance monitoring index; Extract the energy management data from the intelligent control dataset of the public building and divide it into a second training set and a second test set; Using the random forest algorithm, the second training set data is set as input, and the energy management reference index is set as output. The first random forest relationship, the second random forest relationship, and the third random forest relationship are learned to obtain the trained energy management monitoring model. The second test set data is input into the trained energy management monitoring model. The parameters of the energy management monitoring model are adjusted using a stochastic gradient descent optimizer to optimize the performance of the energy management monitoring model and obtain the final energy management monitoring model. By combining the current energy management data with the final energy management monitoring model, the corresponding energy management reference index is output.

6. The intelligent control system for public buildings based on LoRa communication technology according to claim 5, characterized in that, The process of extracting features from the security monitoring data to obtain security monitoring feature data, constructing a security monitoring analysis model, and obtaining a security monitoring evaluation index includes: The safety monitoring feature data includes the flame area and flame RGB value inside the public building, the water leakage area and water level inside the public building, and the frequency of people entering and leaving and the stay time in the real-time monitoring images inside the public building. The convolutional neural network is used to extract features from real-time monitoring images of the interior of the public building. The convolutional neural network includes convolutional layers, pooling layers, and fully connected layers. The convolutional layer uses several convolutional kernels to extract the edges, corners, and textures of the real-time monitoring images inside the public building, and automatically learns to recognize the relevant features of the security monitoring feature data to obtain a security monitoring feature map. The pooling layer reduces the size of the security monitoring feature map through average pooling; Using the fully connected layer as the last layer of the convolutional neural network, the flame area and water leakage area in the real-time monitoring image of the public building are identified and classified based on the processing of the convolutional layer and pooling layer. The mask images of the flame area and water leakage area are obtained respectively. Combined with the OpenCV contour detection method, the flame area, flame RGB value, water leakage area and water level of the public building are obtained. Then, the target detection algorithm is used to identify the people in the real-time monitoring image of the public building. By analyzing the changes in the position of people in several image frames and combining the tracking algorithm, the entry and exit frequency and stay time of people in the real-time monitoring image of the public building are obtained. The convolutional neural network obtains the final safety monitoring feature data through several convolutional layers, pooling layers, and fully connected layers, and integrates the safety monitoring feature data into the intelligent control dataset of the public building.

7. The intelligent control system for public buildings based on LoRa communication technology according to claim 6, characterized in that, The process of constructing a security monitoring analysis model and obtaining a security monitoring evaluation index includes: The safety monitoring and assessment index includes the fire monitoring index, the access control monitoring index, and the water leakage monitoring index; Extract the safety monitoring feature data, real-time flame temperature, smoke concentration and real-time humidity of water pipe surface in the public building's indoor area from the intelligent control dataset of the public building, and convert them into a third training set and a third test set; Using the convolutional neural network algorithm, the third training set data is taken as input, and the security monitoring evaluation index is taken as output. The first convolutional neural network relationship, the second convolutional neural network relationship, and the third convolutional neural network relationship are learned. Through learning the convolutional neural network relationship, the trained security monitoring analysis model is obtained. The third test set data is input into the trained security monitoring and parsing model, the parameters of the security monitoring and parsing model are adjusted, the performance of the security monitoring and parsing model is optimized, and the final security monitoring and parsing model is obtained. The current safety monitoring feature data, the real-time flame temperature, smoke concentration, and real-time humidity of the water pipe surface in the public building are input into the final safety monitoring analysis model to obtain the safety monitoring evaluation index.

8. The intelligent control system for public buildings based on LoRa communication technology according to claim 7, characterized in that, The process of constructing an indoor environment diagnostic model and outputting an indoor environment diagnostic index includes: The indoor environment diagnostic index includes the air quality diagnostic index and the noise diagnostic index. The environmental monitoring data in the intelligent control dataset of the public building is extracted and divided into a fourth training set and a fourth test set. Using the recurrent neural network algorithm, the environmental monitoring data is set as the input data, the indoor environmental diagnostic index is set as the output data, the first recurrent neural network relationship and the second recurrent neural network relationship are learned, and the trained indoor environmental diagnostic model is obtained. The fourth test set data is input into the trained indoor environment diagnostic model, the parameters of the indoor environment diagnostic model are adjusted, the performance of the indoor environment diagnostic model is optimized, the final indoor environment diagnostic model is obtained, and the indoor environment diagnostic index is output by combining the current environmental monitoring data.

9. The intelligent control system for public buildings based on LoRa communication technology according to claim 8, characterized in that, The process of constructing a smart control model for public buildings and generating smart control commands for public buildings includes: The intelligent control commands for public buildings include energy management commands, safety management commands, indoor environment optimization commands, and intelligent parking guidance commands. The energy management commands consist of intelligent lighting commands, air conditioning and heating control commands, and energy balance allocation commands. The safety management commands consist of fire control commands, access control and security control commands, and water leakage control commands. The indoor environment optimization commands consist of air quality optimization commands and noise control commands. The energy management reference index, the safety monitoring and assessment index, the indoor environment diagnosis index, and the smart parking data are used as input data, and the intelligent control command for public buildings is used as output data. The ant colony algorithm is used to initialize the number and initial position of individual ants, set decision nodes related to the intelligent control commands of the public building, and initialize pheromones at each decision node; Each of the intelligent control commands for public buildings and the intelligent parking data is given a path selection standard. Based on the set path selection standard, the ant leaves pheromones at each decision node. According to the pheromone concentration and the heuristic function, a better decision path is selected to complete the construction of the intelligent control model for the public buildings. Based on the objective function in the intelligent control model of the public building, the quality of each of the better decision paths is evaluated. Through multiple iterations, the pheromones on each decision node are updated, and the best decision path is further selected to obtain the intelligent control command for the public building.

10. The intelligent control system for public buildings based on LoRa communication technology according to claim 9, characterized in that, The process of obtaining the final intelligent control system for public buildings based on LoRa communication technology, and realizing intelligent control of public buildings, includes: The system performance evaluation model, the energy management monitoring model, the security monitoring analysis model, the indoor environment diagnosis model, and the public building intelligent control model are integrated and deployed into the platform layer of the LoRa communication intelligent control system to obtain the final public building intelligent control system based on LoRa communication technology. Based on the LoRa transmission data acquired by the perception layer, the platform layer performs performance adjustments and optimizations on the LoRa communication intelligent control system. Through the model deployed in the platform layer, the control commands in the energy management, security monitoring, environmental monitoring, and intelligent parking processes of the public building are parsed, and finally, the intelligent control commands of the public building are output. Then, the intelligent control commands of the public building are wirelessly transmitted to the application layer through the network layer. The application layer executes the intelligent control commands of each public building to realize the intelligent control of the public building.

Images