Indoor temperature and humidity acquisition system based on distributed architecture

Abstract

The invention relates to the technical field of indoor environment monitoring, and discloses an indoor temperature and humidity acquisition system based on a distributed architecture, which comprises a distributed sensor node cluster, a data aggregation gateway and a central processing unit. The distributed sensor node cluster is deployed in different indoor areas to collect data, the data aggregation gateway preprocesses the data, and the central processing unit divides the areas according to spatial and temporal distribution characteristics of the data, constructs various data collection strategies, calculates node energy consumption estimation, data redundancy and transmission delay, and screens an optimal strategy to control node sampling and communication. The system is further provided with a fault-tolerant module, an edge calculation unit and the like, and can detect faults, predict trends, adjust strategy weights and the like. The system is comprehensive and accurate in data acquisition, good in data processing and transmission performance, intelligent in node management and high in environmental adaptability, has expansibility and collaboration, and can be widely applied to temperature and humidity monitoring of various indoor places.

Description

Technical Field

[0001] The present invention relates to the technical field of indoor environment monitoring, and in particular to an indoor temperature and humidity acquisition system based on a distributed architecture. Background Art

[0002] In modern living and working environments, indoor temperature and humidity have a vital impact on human comfort, equipment operation stability, and the quality of item preservation. For example, in office spaces, suitable temperature and humidity can improve employee work efficiency and reduce health problems caused by environmental discomfort; in electronic equipment rooms, accurate temperature and humidity control is the key to ensuring the normal operation of equipment and extending the service life of equipment; in the storage field, temperature and humidity conditions are directly related to the quality and safety of stored items. Therefore, it is of great significance to accurately and real-time collect and monitor indoor temperature and humidity.

[0003] Traditional indoor temperature and humidity collection systems have many limitations. On the one hand, some systems adopt a centralized architecture and rely on only a few sensors for data collection. This method is difficult to fully and accurately reflect the actual distribution of temperature and humidity in large indoor spaces. Due to the influence of factors such as lighting, ventilation, and human activities in different areas of the room, the temperature and humidity vary greatly. Centralized collection is prone to missing key information, resulting in inaccurate monitoring results. For example, in a large shopping mall, the temperature and humidity of areas close to and far from the vents may differ greatly, and centralized collection cannot accurately obtain these difference data.

[0004] On the other hand, existing systems also have problems in data processing and transmission. Some systems lack effective data preprocessing mechanisms, and the raw data may contain noise, outliers, etc. The direct transmission of these data not only increases the amount of data transmission, but also affects the accuracy of subsequent data analysis. Moreover, in terms of data transmission, some systems do not fully consider transmission delays and data redundancy issues, resulting in low data transmission efficiency and even data loss. In some application scenarios with high real-time requirements, such as temperature and humidity monitoring in medical operating rooms, data transmission delays may affect the smooth progress of the operation.

[0005] In addition, traditional systems are not intelligent enough in the management and maintenance of sensor nodes. Sensor node calibration usually needs to be performed manually on a regular basis, which consumes a lot of manpower and time, and the calibration accuracy is difficult to guarantee. At the same time, when a sensor node fails, the system often lacks an effective fault-tolerant mechanism and cannot automatically adjust in time, resulting in interruption of data collection and affecting the stability of the entire monitoring system.

[0006] Traditional systems also have poor adaptability when facing different indoor environments. Different indoor environments, such as conference rooms, laboratories, warehouses, etc., have different requirements and priorities for temperature and humidity monitoring. However, traditional systems are difficult to flexibly adjust data collection strategies according to specific environmental characteristics and cannot meet diverse application needs. Summary of the invention

[0007] The purpose of the present invention is to provide an indoor temperature and humidity acquisition system based on a distributed architecture to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: an indoor temperature and humidity acquisition system based on a distributed architecture, the system comprising: Distributed sensor node clusters are deployed in different areas indoors. Each sensor node includes a temperature and humidity sensor, a signal conditioning unit, and a wireless communication module. A data aggregation gateway is used to receive the original temperature and humidity data collected by the distributed sensor node cluster and pre-process the data; A central processing unit, for receiving pre-processed data, constructing multiple data collection strategies, dividing the data collection area according to the spatiotemporal distribution characteristics of the original temperature and humidity data, calling the dynamic node scheduling strategy, adaptive calibration parameters and data fusion algorithm based on the type of the data collection area, respectively calculating the node energy consumption estimation, data redundancy and transmission delay under each data collection strategy, selecting the optimal data collection strategy according to the selected optimization strategy, and controlling the sampling and communication behavior of the distributed sensor node cluster based on the optimal data collection strategy; The data collection strategy includes a division method of at least two data collection areas, the central processor pre-stores dynamic node scheduling strategies, adaptive calibration parameters and data fusion algorithms corresponding to multiple standard environment models, and the optimization strategy includes the lowest estimated node energy consumption, the smallest data redundancy and the shortest transmission delay; The central processing unit comprises: A region division module, used to construct an initial data collection region according to the spatial coordinates, timestamp and data fluctuation amplitude of the original temperature and humidity data, and generate a plurality of data collection strategies based on the initial data collection region; Preferably, the area division module includes: A space-time coordinate system construction unit, used to establish a four-dimensional coordinate system including three-dimensional space coordinates and time dimension; A feature extraction unit, used to extract a distribution feature function of the original temperature and humidity data in the four-dimensional coordinate system; Clustering unit, according to the formula And the distribution characteristic function generates dynamic clustering center points, where is the time window length, k is the preset number of clusters, ; The area generation unit divides the data collection area based on the dynamic clustering center point and a preset radius threshold.

[0009] Preferably, the method for generating the dynamic node scheduling strategy includes: Constructing a node state matrix based on historical energy consumption data and signal strength of sensor nodes in the data collection area; Extract node load feature vector based on matrix decomposition algorithm; According to the comparison result between the load characteristic vector and the preset threshold, the sampling frequency or the sleep period of the node is dynamically adjusted.

[0010] Preferably, the method for calculating the adaptive calibration parameters includes: Obtaining a reference measurement value of the temperature and humidity sensor under a preset standard environment; Calculate the sensor drift compensation coefficient based on the data deviation rate within the sliding time window; The drift compensation coefficient is iteratively optimized using a Kalman filter algorithm to generate real-time calibration parameters.

[0011] Preferably, the data fusion algorithm includes: Perform outlier detection on the measurements of multiple sensor nodes in the same data collection area; The effective data is integrated based on the weighted average model, where the weight is determined by the node signal strength, historical accuracy and environmental interference level; The wavelet transform algorithm is used to reduce the noise of the fused data.

[0012] Preferably, the system further comprises: The fault-tolerant module is used to detect communication failures or data anomalies of sensor nodes and perform the following operations: When a node loses data continuously for more than a preset number of times, the neighbor node replacement collection mechanism is triggered; When the abnormal data fluctuation exceeds the threshold, the redundant data cross-validation process is initiated.

[0013] Preferably, the preprocessing method of the data convergence gateway includes: Perform timestamp alignment operation on the received raw data; Use differential encoding algorithm to compress data packet length; Attach a digital signature and encryption tag to each data packet.

[0014] Preferably, the system further comprises: The edge computing unit is deployed in the data aggregation gateway and is used to perform the following operations: Generate short-term temperature and humidity change trends based on time series prediction models; The optimization weight coefficient of the data collection strategy is dynamically adjusted according to the trend.

[0015] Preferably, the central processing unit further comprises: The protocol adapter module is used to parse data frames of different communication protocols and convert them into a unified format; Interface expansion modules support plug-and-play configuration of new sensor nodes, including automatic assignment of node IDs and synchronization of calibration parameters.

[0016] Preferably, the system further comprises: Cloud collaboration module, used to upload historical data and optimization strategies to the cloud server; The global parameters of the standard environment model and the data fusion algorithm are updated based on the federated learning algorithm.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention deploys distributed sensor node clusters in different indoor areas. Each node is equipped with a temperature and humidity sensor, a signal conditioning unit and a wireless communication module, which can collect temperature and humidity data from all indoor areas. Compared with the traditional centralized collection method, it effectively avoids the problem of data omissions caused by regional differences, and ensures that the collected data can more truly and comprehensively reflect the actual distribution of indoor temperature and humidity. For example, in a large library, distributed sensor nodes can cover different locations such as bookshelves, reading areas, and air-conditioning outlets, accurately monitor the temperature and humidity changes in each area, and provide an accurate basis for the library's environmental control.

[0018] The data aggregation gateway preprocesses the received raw temperature and humidity data, including timestamp alignment, data packet compression, and adding digital signatures and encryption tags. Timestamp alignment ensures the accuracy of data time, and the differential coding algorithm compresses the length of the data packet, reduces the amount of data transmission, and improves transmission efficiency; digital signatures and encryption tags ensure the integrity and security of the data during transmission. The central processing unit constructs multiple data acquisition strategies, divides the data acquisition area according to the spatiotemporal distribution characteristics of the original temperature and humidity data, and calls the corresponding dynamic node scheduling strategy, adaptive calibration parameters, and data fusion algorithm based on different area types. After calculating the node energy consumption estimate, data redundancy, and transmission delay under each data acquisition strategy, the optimal strategy is selected, which effectively reduces data redundancy, shortens transmission delay, and improves the accuracy of data processing and the reliability of transmission.

[0019] The central processor in the system can construct a node state matrix based on the historical energy consumption data and signal strength of the sensor nodes in the data collection area, and then extract the node load feature vector to dynamically adjust the node's sampling frequency or sleep cycle. This intelligent management method can reasonably allocate resources according to the actual working status of the node, reduce node energy consumption, extend the service life of the sensor node, and reduce maintenance costs. When a sensor node has a communication failure or data anomaly, the fault-tolerant module will respond in time. If the node loses data continuously for more than the preset number of times, the neighbor node replacement collection mechanism is triggered; if the abnormal data fluctuation exceeds the threshold, the redundant data cross-validation process is started. These measures ensure that the system can continue to collect data stably when problems occur in some nodes, greatly improving the stability and reliability of the system.

[0020] The central processor pre-stores dynamic node scheduling strategies, adaptive calibration parameters and data fusion algorithms corresponding to a variety of standard environmental models. In practical applications, the system can automatically select appropriate strategies, parameters and algorithms according to the characteristics of different indoor environments to achieve accurate monitoring of different environments. In laboratory environments, the accuracy requirements for temperature and humidity are high, and the system can call high-precision calibration parameters and more accurate data fusion algorithms; in warehouse environments, more attention is paid to energy consumption and data stability, and the system will give priority to energy-saving node scheduling strategies. This adaptive capability enables the system to be widely used in various types of indoor places to meet diverse application needs.

[0021] The interface expansion module of the central processor supports plug-and-play configuration of new sensor nodes, automatically assigns node IDs and synchronizes calibration parameters, making it easy for the system to expand according to actual needs. Whether it is necessary to increase the monitoring area or replace faulty nodes, it can be easily achieved, reducing the difficulty and cost of system expansion. The cloud collaboration module uploads historical data and optimization strategies to the cloud server, and updates the global parameters of the standard environmental model and data fusion algorithm based on the federated learning algorithm. In this way, different indoor temperature and humidity acquisition systems can learn from each other and optimize collaboratively, continuously improve the performance and monitoring level of the system, and adapt to changing application scenarios and needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a working principle diagram of the indoor temperature and humidity collection system of the present invention; Figure 2 A graph of steps generated for a dynamic node scheduling strategy; Figure 3 This is the workflow diagram of the data fusion algorithm. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] See also Figure 1-3 The overall implementation scheme of the indoor temperature and humidity acquisition system based on the distributed architecture of the present invention is as follows: The system includes a distributed sensor node cluster, a data aggregation gateway, and a central processor. The distributed sensor node cluster is deployed in different areas indoors, and each sensor node integrates a temperature and humidity sensor, a signal conditioning unit, and a wireless communication module. The temperature and humidity sensor is responsible for real-time collection of temperature and humidity data in the area. The signal conditioning unit amplifies and filters the signals collected by the sensor to ensure stable signal quality. The wireless communication module sends the processed temperature and humidity data.

[0025] The data aggregation gateway receives the raw temperature and humidity data collected by the distributed sensor node cluster and pre-processes the data. Its main function is to preliminarily organize and optimize the data for more efficient subsequent transmission and processing.

[0026] The central processor receives the preprocessed data and constructs multiple data collection strategies. The central processor first divides the data collection area according to the spatiotemporal distribution characteristics of the original temperature and humidity data, and then calls the dynamic node scheduling strategy, adaptive calibration parameters and data fusion algorithm based on the type of data collection area. In this process, the central processor will calculate the node energy consumption estimate, data redundancy and transmission delay under each data collection strategy. The central processor pre-stores dynamic node scheduling strategies, adaptive calibration parameters and data fusion algorithms corresponding to multiple standard environmental models. The selected optimization strategy includes the lowest node energy consumption estimate, the smallest data redundancy and the shortest transmission delay. Finally, the central processor selects the optimal data collection strategy according to the selected optimization strategy, and controls the sampling and communication behavior of the distributed sensor node cluster based on the optimal data collection strategy, thereby realizing efficient and accurate indoor temperature and humidity collection.

[0027] The present invention will be further described below in conjunction with Examples 1 to 5: Embodiment 1: The space-time coordinate system construction unit establishes a four-dimensional coordinate system that includes three-dimensional spatial coordinates and time dimensions. In actual application scenarios, indoor space can be represented by three-dimensional coordinates (x, y, z), and the time dimension can be accurate to seconds or finer time granularity. Through such a four-dimensional coordinate system, the original temperature and humidity data can be located in space and time, providing a basic framework for subsequent analysis.

[0028] The feature extraction unit extracts the distribution characteristic function of the original temperature and humidity data in the four-dimensional coordinate system. For example, by using statistical analysis methods, the mean, variance and other statistics of the temperature and humidity data within a certain time and space range are calculated to obtain the characteristic function that reflects the data distribution law. These characteristic functions can intuitively show the changing trend and fluctuation of temperature and humidity data at different time and space locations.

[0029] The clustering units are classified according to the formula And distribution characteristic function generates dynamic cluster center points, where is the time window length, k is the preset number of clusters, Assume that in a large indoor space, the preset number of clusters k is 5 and the time window length is The clustering unit calculates the dynamic clustering center points at 5 different times according to the formula, and these center points will be dynamically adjusted with the changes of time and data.

[0030] The area generation unit divides the data collection area based on the dynamic cluster center point and the preset radius threshold. For example, through multiple experiments, the preset radius threshold is determined to be 5 meters, and different data collection areas are divided with each dynamic cluster center point as the center and 5 meters as the radius. This division can classify areas with similar temperature and humidity data characteristics into one category, which is convenient for subsequent targeted data processing and analysis.

[0031] Embodiment 2: This embodiment describes in detail the generation method of dynamic node scheduling strategy and its practical application in the system, which helps to reduce node energy consumption and extend the service life of sensor nodes. The generation of dynamic node scheduling strategy is mainly based on the historical energy consumption data and signal strength of sensor nodes in the data collection area.

[0032] First, a node status matrix is ​​constructed based on the historical energy consumption data and signal strength of the sensor nodes in the data collection area. In actual operation, taking a data collection area containing 10 sensor nodes as an example, the energy consumption data and signal strength data of each node in the past period of time (such as 24 hours) are sorted. The energy consumption data can be recorded in milliampere hours (mAh), and the signal strength data is recorded in decibel milliwatts (dBm). These data are arranged in order of node number and time to form a node status matrix with 10 rows (corresponding to 10 nodes) and multiple columns (corresponding to energy consumption and signal strength data at different time points).

[0033] Next, the node load feature vector is extracted based on the matrix decomposition algorithm. The singular value decomposition (SVD) algorithm is used to decompose the node state matrix to obtain vectors that can reflect the node load characteristics. These vectors contain comprehensive information about node energy consumption and signal strength. For example, by analyzing the elements in the vector, it is possible to determine which nodes have higher energy consumption and which nodes have weaker signal strength.

[0034] Finally, based on the comparison between the load characteristic vector and the preset threshold, the sampling frequency or sleep cycle of the node is dynamically adjusted. Assume that the preset energy consumption threshold is 50mAh and the signal strength threshold is -80dBm. If the load characteristic vector of a node shows that its energy consumption exceeds 50mAh and the signal strength is lower than -80dBm, the system will reduce the sampling frequency of the node, such as from once per minute to once every 5 minutes; or increase the sleep cycle of the node, so that it is in a dormant state for a period of time to reduce energy consumption. This can balance the load of the node and improve the stability and energy saving of the entire system.

[0035] Embodiment 3: This embodiment focuses on the calculation method of adaptive calibration parameters, which is crucial to improving the accuracy of temperature and humidity sensor measurements. The adaptive calibration parameters are calculated mainly by obtaining the reference measurement values ​​of the temperature and humidity sensor under a preset standard environment and combining them with the data deviation rate within the sliding time window.

[0036] First, obtain the benchmark measurement value of the temperature and humidity sensor in a preset standard environment. In a laboratory environment, simulate a standard temperature and humidity environment, such as a temperature of 25°C and a relative humidity of 50%. Place the temperature and humidity sensor in this environment, and after it stabilizes for a period of time, record its measurement value as the benchmark measurement value. Assume that after multiple measurements, the benchmark measurement value of temperature is 24.9°C and the benchmark measurement value of relative humidity is 49.8%.

[0037] Then, the sensor drift compensation coefficient is calculated based on the data deviation rate in the sliding time window. Taking the temperature sensor as an example, the sliding time window is set to 10 minutes, and data is collected every 1 minute. In each 10-minute window, the deviation rate between the current measurement value and the reference measurement value is calculated. Assuming that in a certain window, the measurement value in the first minute is 25.1℃, the deviation rate is (25.1 - 24.9) ÷ 24.9 × 100% ≈ 0.8%. The sensor drift compensation coefficient is obtained by weighted averaging of all deviation rates in the window.

[0038] Finally, the Kalman filter algorithm is used to iteratively optimize the drift compensation coefficient to generate real-time calibration parameters. The Kalman filter algorithm can continuously optimize the calibration parameters based on the calibration parameters of the previous moment and the measurement data of the current moment. In practical applications, as time goes by, the temperature and humidity sensors will be affected by various factors, such as environmental changes, component aging, etc. Through the Kalman filter algorithm, the system can dynamically adjust the calibration parameters to ensure the accuracy of sensor measurements. For example, after multiple iterative optimizations, the real-time calibration parameters of the temperature sensor may be adjusted from an initial value to a more accurate value, making the measurement results closer to the actual temperature and humidity values.

[0039] Embodiment 4: This embodiment describes in detail the application process of the data fusion algorithm in the system, which can improve the reliability and accuracy of data and reduce the impact of data noise. The data fusion algorithm mainly includes outlier detection of the measurement values ​​of multiple sensor nodes in the same data collection area, fusion of valid data based on the weighted average model, and noise reduction of the fused data using the wavelet transform algorithm.

[0040] When performing outlier detection on the measured values ​​of multiple sensor nodes in the same data collection area, an outlier detection algorithm based on statistical methods is used. For example, the mean and standard deviation of the data are calculated, and a threshold is set, such as 3 times the standard deviation. If the deviation of the measured value of a certain sensor node from the mean exceeds 3 times the standard deviation, the data is considered to be an outlier. In a data collection area containing 5 sensor nodes, the temperature measurement values ​​at a certain moment are 24.5℃, 24.8℃, 25.0℃, 24.6℃, and 30.0℃ respectively. Through calculation, it is found that the deviation of 30.0℃ from the mean is too large, so it is determined as an outlier and removed.

[0041] The valid data is fused based on the weighted average model, where the weight is determined by the node signal strength, historical accuracy and environmental interference level. For example, a higher weight is given to a node with a stronger signal strength, a higher historical accuracy and a lower environmental interference level. Assume that the signal strength of node A is -70dBm, the historical accuracy is 98%, and the environmental interference level is low; the signal strength of node B is -80dBm, the historical accuracy is 95%, and the environmental interference level is medium. Through a certain calculation method, the weight of node A is determined to be 0.6 and the weight of node B is determined to be 0.4. When fusing temperature measurements, if the measurement value of node A is 24.8℃ and the measurement value of node B is 24.6℃, the fused temperature value is 24.8×0.6 + 24.6×0.4 = 24.72℃.

[0042] The wavelet transform algorithm is used to reduce the noise of the fused data. The wavelet transform algorithm can decompose the data into different frequency scales, remove high-frequency noise components, and retain low-frequency useful signals. For example, the fused temperature and humidity data are subjected to wavelet transform to obtain the coefficients of different frequency components. The high-frequency coefficients are thresholded to remove smaller coefficients, and then the data is reconstructed through inverse wavelet transform to obtain the temperature and humidity data after noise reduction, thereby improving the quality and reliability of the data.

[0043] Embodiment 5: This embodiment covers the specific implementation methods of the fault-tolerant module, data aggregation gateway preprocessing, edge computing unit, protocol adaptation module, interface extension module and cloud collaboration module in the system. These modules jointly ensure the stability, scalability and collaboration of the system.

[0044] The fault-tolerant module is used to detect communication failures or data anomalies of sensor nodes. When a node loses data continuously for more than a preset number of times, the neighbor node replacement collection mechanism is triggered. For example, the preset number of consecutive data losses is 5 times. If a sensor node fails to upload data for 5 consecutive times, the system will automatically detect the fault and start the neighbor node replacement collection mechanism. The system will send instructions to the nodes adjacent to the faulty node and with good signals, allowing them to replace the faulty node for data collection within a period of time to ensure the continuity of data collection. When the abnormal data fluctuation exceeds the threshold, the redundant data cross-validation process is started. Assuming that the normal fluctuation range of temperature data is ±2℃, if the temperature data fluctuates by more than 5℃ at a certain moment, the system will start the redundant data cross-validation process, compare the measurement data of other nodes at similar times, and determine whether the data is abnormal. If the abnormality is confirmed, the corresponding processing will be performed.

[0045] The preprocessing method of the data aggregation gateway includes aligning the timestamp of the received raw data, compressing the length of the data packet using the differential coding algorithm, and attaching a digital signature and encryption tag to each data packet. In the timestamp alignment operation, since the clocks of different sensor nodes may have slight deviations, the data aggregation gateway will uniformly calibrate the timestamps of the received raw data according to its own high-precision clock to ensure the accuracy of the data time. The differential coding algorithm encodes by calculating the difference between adjacent data, thereby compressing the length of the data packet. For example, for a series of temperature data of 24.5℃, 24.6℃, and 24.8℃, differential coding can be used to record them as 24.5℃, 0.1℃, and 0.2℃, reducing the amount of data transmission. Attach a digital signature and encryption tag to each data packet. The digital signature is used to verify the integrity and source reliability of the data, and the encryption tag uses an encryption algorithm (such as the AES algorithm) to encrypt the data to ensure the security of the data during transmission.

[0046] The edge computing unit is deployed in the data aggregation gateway to generate short-term temperature and humidity change trends based on the time series prediction model, and dynamically adjust the optimization weight coefficient of the data collection strategy according to the trend. The time series prediction model can adopt the ARIMA model, etc. For example, by training the temperature and humidity data of the past 24 hours, the ARIMA model can predict the temperature and humidity change trend within the next hour. If the temperature and humidity are predicted to change drastically, the edge computing unit will appropriately increase the weight coefficient of the optimization strategy of the shortest transmission delay in the data collection strategy to collect data more timely; if the temperature and humidity are predicted to change relatively smoothly, the weight coefficient will be reduced, and more attention will be paid to the optimization of node energy consumption and data redundancy.

[0047] The protocol adapter module in the central processing unit is used to parse data frames of different communication protocols and convert them into a unified format. In actual applications, there may be sensor nodes with multiple communication protocols, such as ZigBee, Bluetooth, Wi-Fi, etc. The protocol adapter module can identify the data frame formats of different protocols and convert them into a format that the system can uniformly process through a specific parsing algorithm to ensure smooth transmission and processing of data. The interface expansion module supports plug-and-play configuration of new sensor nodes, including automatic assignment of node IDs and synchronization of calibration parameters. When a new sensor node is connected to the system, the interface expansion module will automatically assign it a unique node ID to avoid ID conflicts. At the same time, the calibration parameters are obtained from the central processing unit and synchronized to the new node to ensure that the new node can accurately collect data.

[0048] The cloud collaboration module is used to upload historical data and optimization strategies to the cloud server, and update the global parameters of the standard environment model and data fusion algorithm based on the federated learning algorithm. The cloud server has powerful storage and computing capabilities, and can store a large amount of historical data, which is convenient for users to perform data analysis and mining. The federated learning algorithm allows each node to collaboratively update the global model parameters without sharing the original data. For example, systems in different indoor environments upload their own model training results to the cloud. The cloud integrates these results based on the federated learning algorithm, updates the global parameters of the standard environment model and data fusion algorithm, and then sends the updated parameters to each system to achieve continuous optimization of system performance.

[0049] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0050] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An indoor temperature and humidity acquisition system based on a distributed architecture, characterized in that: The system comprises: Distributed sensor node clusters are deployed in different areas indoors. Each sensor node includes a temperature and humidity sensor, a signal conditioning unit, and a wireless communication module. A data aggregation gateway is used to receive the original temperature and humidity data collected by the distributed sensor node cluster and pre-process the data; A central processing unit, for receiving pre-processed data, constructing multiple data collection strategies, dividing the data collection area according to the spatiotemporal distribution characteristics of the original temperature and humidity data, calling the dynamic node scheduling strategy, adaptive calibration parameters and data fusion algorithm based on the type of the data collection area, respectively calculating the node energy consumption estimation, data redundancy and transmission delay under each data collection strategy, selecting the optimal data collection strategy according to the selected optimization strategy, and controlling the sampling and communication behavior of the distributed sensor node cluster based on the optimal data collection strategy; The data collection strategy includes a division method of at least two data collection areas, the central processor pre-stores dynamic node scheduling strategies, adaptive calibration parameters and data fusion algorithms corresponding to multiple standard environment models, and the optimization strategy includes the lowest estimated node energy consumption, the smallest data redundancy and the shortest transmission delay; The central processing unit comprises: The area division module is used to construct an initial data collection area according to the spatial coordinates, timestamp and data fluctuation amplitude of the original temperature and humidity data, and generate multiple data collection strategies based on the initial data collection area.

2. The indoor temperature and humidity acquisition system based on a distributed architecture according to claim 1 is characterized in that: The area division module comprises: A space-time coordinate system construction unit, used to establish a four-dimensional coordinate system including three-dimensional space coordinates and time dimension; A feature extraction unit, used to extract a distribution feature function of the original temperature and humidity data in the four-dimensional coordinate system; Clustering unit, according to the formula And the distribution characteristic function generates dynamic clustering center points, where is the time window length, k is the preset number of clusters, ; The area generation unit divides the data collection area based on the dynamic clustering center point and a preset radius threshold.

3. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 2 is characterized in that: The method for generating the dynamic node scheduling strategy includes: Constructing a node state matrix based on historical energy consumption data and signal strength of sensor nodes in the data collection area; Extract node load feature vector based on matrix decomposition algorithm; According to the comparison result between the load characteristic vector and the preset threshold, the sampling frequency or the sleep period of the node is dynamically adjusted.

4. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 3 is characterized in that: The calculation method of the adaptive calibration parameters includes: Obtaining a reference measurement value of the temperature and humidity sensor under a preset standard environment; Calculate the sensor drift compensation coefficient based on the data deviation rate within the sliding time window; The drift compensation coefficient is iteratively optimized using a Kalman filter algorithm to generate real-time calibration parameters.

5. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 4 is characterized in that: The data fusion algorithm includes: Perform outlier detection on the measurements of multiple sensor nodes in the same data collection area; The effective data is integrated based on the weighted average model, where the weight is determined by the node signal strength, historical accuracy and environmental interference level; The wavelet transform algorithm is used to reduce the noise of the fused data.

6. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 5 is characterized in that: The system further comprises: The fault-tolerant module is used to detect communication failures or data anomalies of sensor nodes and perform the following operations: When a node loses data continuously for more than a preset number of times, the neighbor node replacement collection mechanism is triggered; When the abnormal data fluctuation exceeds the threshold, the redundant data cross-validation process is initiated.

7. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 6 is characterized in that: The preprocessing method of the data convergence gateway includes: Perform timestamp alignment operation on the received raw data; Use differential encoding algorithm to compress data packet length; Attach a digital signature and encryption tag to each data packet.

8. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 7 is characterized in that: The system further comprises: The edge computing unit is deployed in the data aggregation gateway and is used to perform the following operations: Generate short-term temperature and humidity change trends based on time series prediction models; The optimization weight coefficient of the data collection strategy is dynamically adjusted according to the trend.

9. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 8 is characterized in that: The central processing unit further comprises: The protocol adapter module is used to parse data frames of different communication protocols and convert them into a unified format; Interface expansion modules support plug-and-play configuration of new sensor nodes, including automatic assignment of node IDs and synchronization of calibration parameters.

10. The indoor temperature and humidity acquisition system based on distributed architecture according to claim 9, characterized in that: The system further comprises: Cloud collaboration module, used to upload historical data and optimization strategies to the cloud server; The global parameters of the standard environment model and the data fusion algorithm are updated based on the federated learning algorithm.

Images