Six constant environment parameter real-time monitoring and ai intelligent regulation system based on internet of things

Abstract

The application relates to the technical field of control or regulation systems, and discloses a six-constant environment parameter real-time monitoring and AI intelligent regulation and control system based on the Internet of Things, which comprises the following steps: presetting a state tolerance band through an edge control node and calculating a time vector of environment parameters in real time, combining the time vector control gate to realize dynamic regulation and control decision, preventing the actuator from moving when the parameters deviate from the tolerance band but tend to return spontaneously, and triggering fine adjustment control when the parameters are within the tolerance band but tend to break through the boundary. Through the intelligent sensing and trend prediction mechanism of environment dynamic inertia, the application significantly reduces the performance consumption in the confrontation with the law of natural thermodynamics, realizes accurate identification of transient interference with the help of micro-disturbance interrogation, combines adaptive closed-loop calibration of control efficiency, and builds an intelligent regulation and control system with response speed and long-term stability.

Description

Technical Field

[0001] This invention relates to a real-time monitoring and AI-powered intelligent control system for six constant environmental parameters based on the Internet of Things (IoT). Specifically, it relates to an information system integration service that utilizes artificial intelligence and computers to provide environmental control for smart buildings, smart homes, and high-precision industrial production environments, belonging to the field of control or regulation system technology. Background Technology

[0002] Currently, the field of building environment control generally adopts error-driven reactive control strategies. These strategies use fixed target values ​​as a benchmark and trigger actuator actions by continuously monitoring parameter deviations. In high-precision constant temperature and humidity scenarios such as laboratories, data centers, precision manufacturing workshops, and biopharmaceutical cleanrooms, this type of method is prone to causing significant energy consumption fluctuations due to excessive intervention. When environmental parameters naturally shift due to diurnal temperature differences or equipment start-up and shutdown, traditional systems often ignore their inherent inertial characteristics. The forced return to the set point operation mode leads to frequent actuator start-up and shutdown, which not only wastes energy but also accelerates equipment aging.

[0003] Specifically, existing technologies have three core defects: 1. The conflict between static target point setting and dynamic thermodynamic laws leads to a vicious cycle of ineffective energy consumption in the control system; 2. Due to the rigidity of information processing and decision-making logic, the discrete threshold judgment mechanism adopted by existing computer control systems cannot identify transient disturbances and steady-state evolution trends, and is prone to misjudgment control under complex working conditions; 3. Existing systems lack integrated online perception of actuator status, resulting in a mismatch between control strength and equipment status during long-term operation.

[0004] To address the aforementioned challenges, recent studies have attempted to introduce predictive algorithms to optimize control timing. However, their over-reliance on edge computing resources and the decision-making delays caused by complex models have weakened the system's robustness in real-time scenarios. Therefore, how to construct an intelligent control mechanism that combines trend prediction capabilities with dynamic adaptive characteristics, achieving fundamental energy consumption optimization while ensuring environmental stability, has become the technical problem this invention aims to solve. Summary of the Invention

[0005] This invention provides a real-time monitoring and AI-powered intelligent control system for six constant environmental parameters based on the Internet of Things. Its main purpose is to solve the problems of ineffective energy consumption accumulation and misjudgment of transient interference caused by existing control strategies in smart home, building and industrial production applications due to neglecting the dynamic inertia of the environment.

[0006] To achieve the above objectives, the present invention provides a real-time monitoring and AI-powered intelligent control system for six constant environmental parameters based on the Internet of Things, the system comprising:

[0007] The environmental parameter monitoring unit is configured to collect the current values ​​of environmental parameters in real time and generate time series data;

[0008] Actuator, configured to physically intervene in the environment;

[0009] An edge control node is electrically connected to the environmental parameter monitoring unit and the actuator. The edge control node is configured to: determine a state tolerance band containing the target state; and dynamically calculate a time vector representing the trend of environmental parameter changes based on time series data. The time vector includes the first and second derivatives of the environmental parameter changes.

[0010] The time vector control gate makes decisions on the control commands of the actuator. The time vector control gate is configured as follows: when the current value of the environmental parameter is outside the state tolerance zone, and the time vector indicates that the environmental parameter is spontaneously regressing into the state tolerance zone at a trend higher than the regression threshold, the actuator is prevented from initiating a control action aimed at correcting the deviation; when the current value of the environmental parameter is within the state tolerance zone, and the time vector indicates that the environmental parameter is rushing towards the boundary of the state tolerance zone at a trend higher than the breakthrough threshold, the actuator is initiated to perform a fine-tuning control action to reduce the momentum of the environmental parameter change.

[0011] The cloud server is configured to receive historical data uploaded by the edge control nodes, and based on the historical data, optimize the width of the state tolerance band, the regression threshold, and the breakthrough threshold, and then send the optimized parameters to the edge control nodes.

[0012] Preferably, the edge control node is further configured to: when the rate of change indicated by the time vector exceeds the abrupt change threshold, not immediately execute the judgment result of the time vector control gate; command the airflow circulation device in the actuator to perform a disturbance, the duration of which is 1 to 5 seconds; determine whether the initial time vector originates from a transient local disturbance or a steady-state environment evolution based on the subsequent changes of the time vector after the disturbance; and decide whether to execute the initial judgment result of the time vector control gate based on the judgment result to filter out transient local disturbances.

[0013] Preferably, the edge control node is further configured to perform a self-correction function for control performance. The self-correction function for control performance includes: recording the controlled evolution vector generated during the execution of control actions by the actuator, and the natural evolution vector generated during the period when the actuator does not perform control actions; generating a control performance signature that can represent the current actual performance of the actuator by comparing the controlled evolution vector and the natural evolution vector; and automatically adjusting the intensity or duration of subsequent fine-tuning control actions of the edge control node based on the dynamically updated control performance signature to achieve the stability of system operation performance.

[0014] Preferably, the intensity or duration of the fine-tuning control action is proportional to the trend intensity indicated by the time vector, and the intensity or duration of the fine-tuning control action is five percent to twenty percent of the actuator's maximum power.

[0015] Preferably, the edge control node is configured to determine the specific triggering conditions of the control command based on a specific combination of the rate and acceleration of the time vector. The conditions are: when both the rate and acceleration indicate that the environmental parameters are accelerating back towards the center of the state tolerance zone, a blocking control action is triggered; when both the rate and acceleration indicate that the environmental parameters are accelerating toward the boundary of the state tolerance zone, a fine-tuning control action is triggered.

[0016] Preferably, the system also includes a central coordinator, which is configured not to participate in real-time control decisions, but to collect energy consumption data and environmental status data obtained by each edge control node, and adjust the global energy-saving strategy based on the data.

[0017] Preferably, the central coordinator is configured to reflect the global energy-saving strategy as adjustment suggestions for parameter thresholds in each edge control node, and distribute the suggestions to the edge control nodes via a cloud server. The adjustment suggestions are updated quarterly to weekly.

[0018] Preferably, the edge control node is further configured such that: when the speed v and acceleration a of the time vector satisfy v·a>0, the environmental parameters are judged to have a positive trend; when the speed v and acceleration a of the time vector satisfy v·a<0, the environmental parameters are judged to have a negative trend.

[0019] Preferably, the environmental parameter monitoring unit includes a temperature sensor, a humidity sensor, and an air quality sensor, and may optionally include a differential pressure sensor or a cleanliness sensor; the six constant environmental parameters include temperature, humidity, air quality, fresh air volume, airflow organization, and noise, and may further include environmental differential pressure or cleanliness level.

[0020] Preferably, the IoT architecture includes multiple edge control nodes that interact with a cloud server via a wireless communication network. The edge control nodes are also configured to share data with other local IoT devices via the wireless communication network to achieve coordinated environmental sensing and control. The wireless communication network includes a Wi-Fi network or a LoRa network.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. By introducing the double derivative analysis of the time vector, the system can sense the intrinsic evolution momentum of environmental parameters and actively judge the natural regression trend when the parameters deviate from the tolerance zone. This deep coupling mechanism with the inertial characteristics of the physical system enables the control system to shift from passive correction to trend prediction, effectively avoiding the energy redundancy consumption caused by the conflict between traditional control strategies and natural thermodynamic laws. The synergistic optimization of the spontaneous regression path of environmental parameters and the actuator start-stop strategy builds a deep fit between the control logic and the dynamic characteristics of the physical world.

[0023] 2. By combining a micro-perturbation questioning mechanism with trend credibility verification, the system proactively applies standardized airflow disturbances when it detects a dramatic trend. It uses the response characteristics of subsequent time vectors to distinguish between local transient disturbances and global steady-state evolution. This proactive sensing strategy based on physical excitation avoids the limitations of traditional filtering algorithms in surface-level processing of noise signals. It performs mechanism-level identification of the essential characteristics of environmental disturbances at the control decision level, significantly improving control stability under complex operating conditions. At the same time, through dynamic comparison of natural evolution vectors and controlled evolution vectors, the system continuously constructs feature signatures representing the actual performance of actuators during operation. This inverse modeling mechanism, which infers equipment status based on control results, achieves real-time coupling and matching of actuator performance degradation and control strategies. The dynamic self-correction process of control parameters forms an endogenous balance between equipment aging and system optimization, ensuring the stability of control accuracy throughout the entire life cycle.

[0024] 3. The cloud server optimizes the state tolerance band parameters using historical data, while edge nodes execute lightweight decisions based on real-time time vectors. This functional decoupling between the two forms a spatiotemporal dual-scale optimization of the control strategy. This architecture design retains the rapid response characteristics of the edge while enabling the dynamic evolution of the parameter set through long-term learning in the cloud. This allows the system to adapt to changes in environmental patterns and the collaborative relationships of device groups. Meanwhile, the central coordinator analyzes the energy consumption characteristics and state evolution patterns of each edge node to establish a mapping relationship between control thresholds and energy-saving strategies. This global optimization strategy based on swarm intelligence guides each node to form the optimal parameter configuration for energy consumption without interfering with real-time control, achieving cross-scale collaboration between micro-control behavior and macro-energy efficiency goals. Attached Figure Description

[0025] Figure 1 This is a comparison diagram of the temperature control strategies of the six constant environment systems of this invention;

[0026] Figure 2 This is a functional structure and data flow diagram of the six constant environment system of the present invention;

[0027] Figure 3 This is a timing diagram of the signal interaction of the six constant environment control system based on time vector control logic of the present invention.

[0028] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0029] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0030] This invention proposes an IoT-based real-time monitoring and AI-powered intelligent control system for six constant environmental parameters. The system comprises an environmental parameter monitoring unit, actuators, edge control nodes, a cloud server, and an optional central coordinator. It constructs an environmental trend recognition model using time vectors, and combines a state tolerance band mechanism, disturbance challenge feedback logic, and a control performance self-correction algorithm to dynamically control six constant indoor environmental parameters (temperature, humidity, air quality, fresh air volume, airflow organization, and noise). The system aims to balance environmental comfort, system responsiveness, and energy efficiency. The core of this invention lies in the construction of a real-time perception and intelligent intervention mechanism for environmental parameter changes. Its basic implementation architecture is as follows:

[0031] First, the environmental parameter monitoring unit is responsible for collecting the current values ​​of six constant parameters in real time and forming time series data. This unit includes temperature sensors, humidity sensors, air quality sensors, and multi-mode sensors related to fresh air, airflow, and noise, deployed in key locations in the environment to ensure comprehensive and representative data coverage. The collected data is received in real time by the edge control node, which, as the core computer processing unit of this information system, executes its built-in artificial intelligence software module to calculate the first derivative (rate) and second derivative (acceleration) of the time series, obtaining the time vector of the environmental parameters. The time vector not only reveals the dynamic trend direction and rate of change of the current parameters but also predicts the possible path of subsequent system evolution, thus providing a quantitative basis for control decisions. Based on time vector analysis, the edge node... A time vector control gate is provided as an intelligent decision-making module to judge the behavior of the actuator. When the current value of a parameter falls outside the set state tolerance zone, but its time vector shows that it is accelerating back into the tolerance zone at a rate higher than the regression threshold, the system determines that the deviation is a self-correcting transient behavior, and the actuator remains silent to avoid unnecessary energy consumption. Conversely, if the parameter is within the tolerance zone, but the time vector indicates that its trend is accelerating towards the tolerance zone boundary, a fine-tuning control action is triggered to reduce the system momentum through appropriate intervention and ensure that the parameter does not exceed the safe range. The state tolerance zone here is a set of upper and lower boundaries set around the target value. Its width is continuously optimized by the cloud server based on historical environmental data and control response records. In the early stage of system operation, the tolerance zone can be set to a fixed value, and then gradually evolves into a dynamic window with upper and lower limit adjustment mechanisms.

[0032] For situations involving drastic changes, this computer system also integrates a disturbance interrogation software module. When the rate of change of the time vector exceeds a set threshold for drastic change, the control gate will pause automatic decision-making and instead drive the airflow circulation device in the actuator to apply a standardized disturbance for 1 to 5 seconds. After the disturbance, the system continues to monitor the trend of the time vector change. If it recovers rapidly or decays in the opposite direction, it indicates that the original drastic change was a transient disturbance; conversely, if the trend continues to strengthen, it is determined to be a steady-state evolution. This process constitutes a mechanism-level discrimination of the source of abnormal signals in the system. Adaptive closed-loop calibration of control performance is another key mechanism. The system records the evolution trajectory of environmental parameters during the actuator intervention period and the non-intervention period, respectively, which constitute the controlled evolution vector and the natural evolution vector. The comparison of the two forms a set reflecting the actual response effect of the actuator. The control performance signature is used to evaluate the substantial impact of the executed action on the system state. If the system determines that the current actuator response intensity is lower than expected, it automatically increases the intensity of the fine-tuning control or extends its duration to achieve dynamic adaptation to actuator aging or environmental response inertia. The adjustment range of this fine-tuning intensity is generally 5% to 20% of the actuator's rated power, and the ratio is dynamically set according to the trend intensity shown by the time vector. Furthermore, the edge control node has a logical basis for judging the direction of the environmental trend, that is, judging the current trend by the consistency of the direction of the rate and acceleration. When the signs of the two are the same, it means that the parameter is accelerating in a unified direction, the trend is consistent and may continue. If the signs of the rate and acceleration are opposite, it indicates that the current trend may decay or reverse, providing a reference for the system to judge whether the current trend can spontaneously regress.

[0033] At the system architecture level, this invention embodies a typical cloud-edge-device information system integrated architecture, aiming to provide a complete set of intelligent environmental control services. Edge nodes connect to the cloud server via wireless communication networks (such as Wi-Fi or LoRa) to achieve bidirectional data exchange. The main responsibilities of the cloud include: tolerable bandwidth, periodic optimization of regression and threshold exceedance, long-term learning and generalization of parameter evolution models, and version management and update distribution of control strategies. If the system is configured with a central coordinator, its main function is to collect energy consumption data and environmental state evolution information from multiple edge nodes, perform horizontal comparisons and trend modeling, and propose... Energy-saving strategy recommendations for different nodes are provided, such as relaxing control targets in certain areas and optimizing strategy update frequency. These recommendations will be synchronized to edge control nodes periodically via cloud servers in the form of parameter adjustment schemes. The update frequency can be set from quarterly to weekly, depending on system stability and seasonal variation frequency. Overall, the system architecture of this invention achieves accurate perception and response to dynamic environments through a triple mechanism of trend perception, perturbation verification, and performance self-correction. It is significantly superior to traditional error-driven reaction control systems, effectively reducing energy consumption while maintaining six constant environmental conditions, and has good scalability and engineering feasibility.

[0034] In this system architecture design, although the central coordinator has the capability to suggest cross-node strategies, it does not directly interfere with the real-time control decisions of edge nodes. All control commands are still generated autonomously by the edge nodes based on local sensing data and dynamic trend judgment mechanisms. The energy-saving optimization schemes provided by the coordinator serve only as a reference for edge node strategy adjustments; the specific execution rhythm and control intensity are still determined independently by the edge nodes to ensure the response flexibility and control autonomy of the distributed architecture. Simultaneously, in the trend identification and feedback mechanism, the system combines the parameter change rate and magnitude to summarize trend characteristics. It does not rely on absolute models or unified algorithms, but rather forms a relative strength judgment standard based on the dynamic sensing results under different scenarios to guide appropriate adjustments to the control strategy. Regarding actuator response analysis, the system focuses on recording the relative differences between parameter change paths before and after control, accumulating feature descriptions through historical comparisons to characterize the trend of actuator response capability changes. Subsequent control actions are then appropriately modified based on these features, achieving continuous maintenance of the system's adaptive control capability throughout the operating cycle. These are all extended implementation methods known to those skilled in the art.

[0035] Example 1: This example specifically illustrates the complete operation process and core mechanism implementation of a real-time monitoring and AI intelligent control system for six constant environmental parameters based on the Internet of Things in a typical building scenario. It focuses on the trigger criterion setting logic of the time vector control gate, the construction principle of the state tolerance band, the response judgment strategy of the micro-disturbance inquiry mechanism, the quantitative adjustment method in the adaptive correction process of control efficiency, and the setting basis for the actuator fine-tuning intensity. In this example, the system is deployed in an office space with a building area of ​​approximately 800 square meters. The control objective is to maintain six environmental parameters—indoor temperature, humidity, air quality, fresh air volume, airflow organization, and noise—stable within the human comfort range, and to achieve a coordinated balance between rapid response and energy consumption optimization when facing external climate fluctuations and frequent changes in human activity. Firstly, regarding environmental parameter acquisition, the system uses widely used multimodal sensor components to collect the six parameters. The placement of each sensor node is engineered to ensure the representativeness and timeliness of the collected data. For example, temperature and humidity sensors are deployed at various locations... In the upper and middle spaces of the functional zones, the perception of fresh air volume and airflow organization is achieved by differential pressure and wind speed sensing modules deployed at the fresh air terminals and return air duct openings. Noise levels are collected in real time by sound pressure sensing modules deployed around typical noisy workstations. The data collected by the above sensors is synchronized to the edge control node in real time for processing. The data processing logic embedded in the node performs first and second derivative calculations on the time series data of each parameter at a fixed period (e.g., every 30 seconds) to represent the rate of change and acceleration of the parameter, respectively, and then constructs the current time vector of the parameter. This time vector not only characterizes the dynamic change trend of the parameter, but also provides a quantitative basis for subsequent control decisions.

[0036] The edge control node is equipped with a time vector control gate, serving as the core module for intelligent decision-making. This gate determines whether to execute control actions. The system sets a set of symmetrical upper and lower limits for each environmental parameter, forming a state tolerance band. This tolerance band is established around the target value; for example, the initial temperature tolerance band is set to ±1 degree Celsius, humidity to ±5%, and air quality tolerance can be set with upper and lower limits of 10% based on PM2.5 concentration fluctuations. The tolerance band initially has a fixed width, which is subsequently dynamically optimized and adjusted by the cloud server based on historical operating data. The update cycle can be set to once a week. When the current value of a certain parameter... When a value is outside the tolerance range, the system determines whether it is a spontaneous regression behavior based on its time vector. Specifically, if the sign of the parameter's rate and acceleration are consistent, and their absolute values ​​both exceed a preset regression threshold (e.g., the rate of the temperature parameter is not less than 0.05 degrees Celsius per minute, and the acceleration is not less than the square of 0.01 degrees Celsius per minute), the system determines that it has a self-regressive trend and controls the gate to prevent the actuator from operating, thus avoiding unnecessary energy consumption. Conversely, if the current value of the parameter is within the tolerance range, but its time vector indicates that it is accelerating towards the tolerance range boundary, and has already... If the threshold is exceeded, for example, if the rate reaches 0.1 degrees Celsius per minute and the acceleration is 0.03 degrees Celsius per minute squared in the same direction, a fine-tuning control action is triggered to reduce the momentum of parameter changes and ensure system stability. During system operation, the time vector is calculated based on the magnitude and rate of change of environmental parameters at continuous time nodes. It is judged by the average rate and acceleration trend of incremental changes at multiple points within a sliding window, thereby revealing the current trend and future trend of the parameters. The system identifies the strength of the trend based on the consistency of the rate and direction of change, and forms a measure of trend momentum to reflect its ability to continuously affect the system state. The judgment of control effectiveness is carried out by analyzing the difference between the environmental response paths generated by the system before and after the intervention of the control action, evaluating the actual performance of the control action on parameter correction, and gradually accumulating to form characteristic response trajectories to characterize the true state of the actuator's current control capability. When judging trends and response behaviors, the system not only considers whether the parameter position deviates from the tolerance range, but also integrates its change direction, momentum accumulation, and whether it has a natural recovery trend, to construct a dynamic judgment mechanism that can adapt to changing operating conditions. When a drastic change in signal occurs, the system guides the environmental parameters to release response trends through short-term disturbance operations. By comparing the trend evolution before and after the disturbance, it determines whether the signal originates from a sudden disturbance or a steady-state evolution, thus ensuring the targetedness and effectiveness of the control response.

[0037] The actuators include devices such as damper drive units, fresh air handling unit frequency converters, and local humidification or dehumidification units. The intensity or duration of their fine-tuning actions is dynamically adjusted according to the trend intensity. Specifically, the system controls the actuator's action intensity between 5% and 20% of its maximum power. The intensity value is determined by a linear mapping based on the magnitude of the trend momentum. For example, when the trend momentum is a unit value of 1.0, the action intensity is 10%. If the trend momentum doubles, the action intensity is adjusted upwards accordingly, but not exceeding 20%. When the system detects a drastic change in a certain parameter's trend momentum within a short period, such as a sudden increase in acceleration exceeding three standard deviations of normal fluctuation, the control gate suspends its original decision and instead drives the air circulation device in the actuator to perform a standardized disturbance operation. This disturbance lasts for 2 seconds. After the disturbance ends, the system re-acquires the time vector for comparison. If the trend rapidly weakens or reverses... If the current trend continues to strengthen, it is determined to be a transient disturbance, and the control action is ignored. If the trend continues to strengthen, it is confirmed to be a steady-state evolution trend, and the control logic is restored. During the long-term operation of the system, the edge control node continuously records the impact results of various control operations. The system records the parameter change trajectory before and after each actuator action as a controlled evolution vector, and at the same time, it records the natural evolution trajectory during the period when the actuator does not act as a natural evolution vector. The two are compared to form a control performance signature, which is used to measure the actual control effect of the current actuator. If the system finds that the controlled response effect has decreased significantly through comparison, it is judged that the equipment may have performance degradation, and the intensity or duration of subsequent fine-tuning control is adjusted accordingly. For example, when the signature indicates that the temperature control effect only reaches half of the expected value, the system automatically increases the intensity of the corresponding control action from 10% to 15% to achieve response compensation under equipment aging conditions.

[0038] In this embodiment, the system uses a LoRa wireless communication network to achieve bidirectional data interaction between edge control nodes and a cloud server. The cloud server is mainly responsible for the periodic optimization of state tolerance band width, regression and threshold exceedance, as well as the long-term learning of parameter evolution models and control strategy version management. If the system is configured with a central coordinator, its role is to collect energy consumption data and environmental state evolution information from all edge control nodes, and propose differentiated energy-saving strategy suggestions through horizontal comparison and trend modeling, such as adjusting the control target accuracy of specific areas and optimizing the parameter update frequency. These suggestions are distributed to each node via the cloud server, and the suggestion update cycle is set according to the system's operational stability and the degree of seasonal variation, which can be once a week or once a quarter.

[0039] Example 2: After deploying the IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters proposed in this invention in a building, the technical team found that the environmental parameters were affected by the diurnal temperature difference and the frequency of human activities, exhibiting periodic shifts. In particular, the temperature and humidity parameters fluctuated significantly. Traditional PID control strategies had response lag and overshoot problems under such dynamic conditions, resulting in frequent actuator start-stop, decreased energy utilization efficiency, and unstable system operation. To verify the control effect and energy consumption performance of the trend discrimination control mechanism based on time vector proposed in this invention in a real office scenario, this experiment was organized. The experiment aimed to clarify the response accuracy, control frequency, and energy consumption optimization capability of the intelligent control strategy under typical disturbance conditions, providing direct engineering verification support for the industrial application of the system.

[0040] The test site was an open-plan office area on the fourth floor of a commercial office building in Central China, with a total area of ​​approximately 870 square meters. The area was well-ventilated and had a north-south orientation. To enhance the representativeness of the test results, the following typical load conditions were set: during working hours (10:00-16:00), the personnel density in the office area was maintained at approximately 80%, and during other times, it was controlled to be below 20%. The target parameters included temperature (set to 24℃), humidity (set to 55%), and air quality (represented by PM2.5 concentration, with a target not exceeding 75 μg / m³). 3 The environmental parameters are sampled every 30 seconds. The acquisition system consists of multimodal sensors for temperature, humidity, air quality, etc., deployed in six key areas. The edge control node has a built-in time vector calculation module that generates the rate (i.e., first derivative) and acceleration (i.e., second derivative) of each parameter in real time to form a time vector. The control actuators include the frequency converter module of the air supply system, the air outlet damper adjustment unit, the ceiling-mounted humidifier, and the chiller unit. Data is transmitted back to the edge control node via the LoRa communication module for real-time processing and control strategy execution. Three sets of comparative operating conditions are set up for the experiment, as shown in Table 1.

[0041] Table 1. Comparison of three sets of working conditions in the test.

[0042]

[0043] Each group of experiments ran for three days, covering a typical climate fluctuation period of continuous working days. The main observation indicators included the root mean square deviation of temperature fluctuation (°C2), the number of actuator start-stop cycles, the average daily energy consumption (kWh), and the duration of temperature deviation from the tolerance zone (minutes). Typical test results for each group are shown in Table 2.

[0044] Table 2. Comparison of typical test results for each group of experiments.

[0045]

[0046] The results show that after introducing the time vector trend discrimination mechanism, Group B significantly reduced the system control frequency and daily energy consumption, and the parameter fluctuation control became more stable. After further adopting the disturbance questioning mechanism, Group C achieved more precise control response during the typical high-temperature disturbance period (14:00-16:00), triggering fine-tuning control only 4 times and without overshoot, further improving energy efficiency. The system selected three key time points to record the controlled and natural evolution rates to quantify the actual control effect of the actuator. The results are shown in Table 3.

[0047] Table 3. Signature data table of control effectiveness at key time points.

[0048]

[0049] The above data shows that the system can dynamically adjust subsequent control intensity based on the actual response, possessing good closed-loop adjustment capability. A stable coupling is formed between control actions and environmental responses, effectively adapting to equipment performance fluctuations. This continuous comparative experiment verified the engineering applicability of the time-vector-based trend discrimination control strategy in actual building environments. This strategy can effectively identify spontaneous regression trends, avoid unnecessary control behaviors, and reduce energy consumption. Simultaneously, the disturbance challenge mechanism enhances the system's ability to identify and suppress transient disturbances, improving the overall robustness of the system. The control performance signature mechanism provides a quantitative basis for equipment state identification and adaptive adjustment, ensuring that the system maintains stable control effects even when actuator performance degrades.

[0050] Example 3: This example combines Figures 1 to 3 This document describes the implementation of an IoT-based real-time monitoring and AI-powered intelligent control system for six constant environmental parameters. Figure 1 As shown, this figure compares the temperature response of the IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters under three different control strategies. The horizontal axis represents time (hours), covering the entire day from 8:00 to 20:00, and the vertical axis represents temperature (°C), with the target temperature marked as 24°C, the upper limit of the tolerance band as 25°C, and the lower limit of the tolerance band as 23°C. The three temperature lines clearly define the target parameters and tolerance band range of the system operation. Group A corresponds to traditional PID control, and is marked in the figure as having significant overshoot. Its temperature curve shows a significant overshoot exceeding the upper limit of the tolerance band around 14:00. Group B uses time vector control, and its temperature curve shows significantly reduced fluctuations compared to the traditional PID control of Group A, with overshoot effectively suppressed and control accuracy significantly improved. Group C adds a disturbance challenge mechanism to Group B, and its temperature response is more stable during the disturbance period (14:00-16:00), with the highest control accuracy. The specific performance indicators of each group are compared in Table 2.

[0051] like Figure 2As shown, the environmental parameter monitoring unit is equipped with temperature, humidity, air quality, and noise sensors, responsible for acquiring real-time data streams. The acquired data is sent to the edge control node, where time vector calculations are performed and a state tolerance band is set. The edge control node transmits the processing results to the time vector control gate, which includes trend prediction and micro-disturbance interrogation functions, further analyzing trend momentum to determine whether to trigger a control command. After the control signal is input to the actuator system, fine-tuning control (range 5%–20%) and the operation of the airflow circulation device are executed. The actuator system's actions are fed back to the edge control node, and... The data is synchronized to the cloud server; the cloud server then performs parameter optimization based on historical data analysis. The optimization results are returned to the edge control nodes through the parameter optimization path, forming a dynamic closed-loop adjustment mechanism. The figure also shows the action paths of three key analysis mechanisms in the system through dashed lines: the double derivative trend analysis acts on the time vector control gate to improve the accuracy of trend prediction; the transient disturbance identification mechanism connects the time vector control gate and the actuator system to identify the true nature of drastic trends; and the control performance self-correction mechanism relies on the actual response results of the actuator system to adjust the control strategy, ensuring the response accuracy and equipment adaptability of the system in long-term operation.

[0052] like Figure 3 As shown, firstly, the environmental parameter monitoring unit acquires six constant parameters—indoor temperature, humidity, air quality, fresh air volume, airflow organization, and noise—by collecting real-time environmental parameter values. This data is then sent to the edge control node, which generates time-series data, constructing a dataset showing the continuous changes in environmental parameter values ​​over time. Subsequently, it dynamically calculates time vectors (velocity and acceleration), representing the first and second derivatives respectively, to characterize the parameter change trends and momentum. The edge control node operates the time vector control gate to make decisions, determining whether to trigger control operations. Depending on the situation, it enters a conditional branch (alt): if the current value is outside the tolerance band and the trend is spontaneously regressing, then blocking control is executed. The system initiates control actions to avoid unnecessary intervention. If the current value is within the tolerance band and the trend is approaching the boundary, a fine-tuning control action (momentum reduction) is executed, sending control commands to the actuators to adjust the working status of equipment such as dampers, fans, and fresh air systems. The actuators, based on the control commands, physically intervene to affect environmental parameters and provide feedback to the controlled environment. The edge control nodes upload historical data to the cloud server for trend pattern analysis and system response behavior. The cloud server optimizes the tolerance band width and regresses the threshold, dynamically adjusting decision parameters based on accumulated data to make the control logic more precise and energy-efficient. Finally, the optimized parameters are sent to the edge control nodes, achieving closed-loop self-evolution and parameter update synchronization of the system.

[0053] Example 4: In a large library complex located in a subtropical climate zone in southern China, the IoT-based real-time monitoring and intelligent control system for six constant environmental parameters of this invention was deployed. The system aims to precisely adjust six environmental parameters: temperature, humidity, air quality, fresh air volume, airflow organization, and noise. This ensures that the enclosed learning space maintains comfort and air quality while achieving dynamic energy conservation. This example selects a library as a typical and complex application scenario. Its building structure is diverse, and the density of people varies significantly over time. Especially around noon and evening, the concentrated entry and exit of readers causes significant fluctuations in environmental parameters, providing a sufficient adaptive verification platform for the key mechanisms of this system. In this scenario, the environmental parameter monitoring unit consists of a temperature and humidity composite sensor, an air quality sensor, a fresh air flux sensing device, an airflow organization map reconstruction component, and an environmental noise acquisition module. The deployment of various sensors is planned in layers based on parameter sensitivity and regional functional differences. For example, the temperature and humidity sensor is deployed in the upper and middle layer boundary area of ​​the reading area and the equipment mezzanine, the fresh air velocity sensor is deployed at the end of the supply and return air ducts, and the air quality sensor covers the entrance passage and densely booked areas, thereby ensuring the representativeness of the collected data and the spatial accuracy of the response. The system uniformly sets the sampling period of various environmental parameters to thirty seconds. The collected data is aggregated to the edge control node via the wireless communication module for unified processing. Inside the edge control node, the rate and acceleration of each parameter are calculated sequentially using the sliding window method. The former reflects the first-order directional trend of environmental parameter changes, and the latter reveals the rate of change of trend momentum, that is, the second-order dynamic characteristics. In order to reduce the interference of short-period noise, the system first performs three-point moving average smoothing on each sensor data before entering the derivative calculation stage. The resulting time vector constitutes the core quantitative basis of the system's intelligent discrimination mechanism.

[0054] During the initial system operation phase, the state tolerance zone is set with target values ​​and fluctuation ranges based on the library's air conditioning design standards. For example, the target temperature is set at 24 degrees Celsius, and the tolerance zone width is initially set to ±1 degree Celsius; the target humidity is set at 55%, and the tolerance zone range is set to fluctuate by 5%; air quality is represented by fine particulate matter concentration, with a limit of 75 micrograms per cubic meter, and the initial tolerance zone fluctuation width is set at 10%. The initial values ​​of the above parameters can be dynamically optimized by the cloud server in conjunction with the library's historical operating data, energy usage records, and external meteorological conditions, and updated weekly and synchronized to the edge control node via an encrypted protocol. The edge control node is equipped with a time vector control gate, which serves as the logical center for actuator behavior decisions, used to determine the behavior based on the real-time calculated time vector and the current parameter values. In this embodiment, regarding the determination of whether to trigger control behavior based on the positional relationship of the tolerance zone, if the current temperature value exceeds the upper limit of the tolerance zone, and both the rate and acceleration of the temperature are negative, and their values ​​exceed the regression judgment thresholds set by the system (i.e., the absolute value of the rate is greater than 0.04 degrees Celsius per minute, and the absolute value of the acceleration is greater than the square of 0.01 degrees Celsius per minute), the system determines that the current temperature rise trend has entered the natural decline stage, and there is no need to activate the cold source actuator. Conversely, if the current temperature is within the tolerance zone but its rate and acceleration are both positive, and the trend momentum reaches the set breakthrough judgment threshold (the rate exceeds 0.06 degrees Celsius per minute, and the acceleration exceeds the square of 0.02 degrees Celsius per minute), the system immediately triggers a fine-tuning control action to reduce the opening of the air supply valve, thereby weakening the momentum of the temperature rise trend and preventing the parameters from exceeding the set boundaries.

[0055] To address the rapid parameter fluctuations commonly seen during peak lunch hours in the library, the system is also equipped with a disturbance challenge mechanism: when the acceleration value exceeds three standard deviations of the historical mean within two consecutive sampling periods, the control gate will pause the original control path and drive the air circulation device to perform a standardized disturbance operation. In this embodiment, the disturbance time is set to two seconds, and the disturbance wind speed is 40% of the actuator's maximum airflow. After the disturbance process ends, the system recalculates and judges the time vector after the disturbance: if the acceleration direction is reversed and the amplitude drops significantly, such that the decrease exceeds 50% of the value before the disturbance, it is considered a transient disturbance, and the current control action will not be executed; otherwise, if the trend is maintained or intensified, execution will resume. The original fine-tuning action; the actuator adjustment intensity is proportionally matched with the trend momentum. The system linearly maps the trend momentum index calculated from the rate and acceleration to the actuator output power control range, which is 5% to 20% of the maximum power. This mapping relationship is established by fitting the dual objective parameters of minimum energy consumption and shortest recovery time during the first three days of system operation. For example, in a typical heating stage, if the system determines that the trend momentum is 0.07 units, it controls the cold source actuator to output at 14% of the rated power, with an action time of no more than five minutes. If the trend momentum decreases by 60% in subsequent sampling periods, the system will actively end the action and record the control performance. After each control action, the system records the trend of environmental parameter changes caused by the control and compares it with the natural trend of changes during non-control periods to generate a control performance signature. This signature consists of the difference between the speed and acceleration of the actuator during the action period and the natural evolution trend of the previous stage. In this scenario, the system calculates the historical signature average for each type of control action using a minimum time window of three days. If the control performance signature values ​​of two consecutive actions of the same type are both lower than 70% of the historical average, the system automatically increases the output power by 5% and extends the duration by one second in the next round of action to offset the potential impact of actuator performance degradation or environmental response inertia changes. At the same time, to achieve consistent coordination between global energy saving and control strategies, the system sets up a central coordinator, which does not directly participate in real-time control decisions. Instead, it collects energy consumption data, control frequency, and parameter fluctuation data uploaded by each edge control node weekly, forms an energy consumption profile model based on horizontal comparison, and proposes strategy optimization suggestions. For example, in densely populated areas, the humidity tolerance zone can be appropriately relaxed to reduce the dehumidification load, and during nighttime closing hours, air quality can be allowed to fluctuate within a more tolerant range.

[0056] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A real-time monitoring and AI-powered intelligent control system for six constant environmental parameters based on the Internet of Things, characterized in that, The system includes: The environmental parameter monitoring unit is configured to collect the current values ​​of environmental parameters in real time and generate time series data; Actuator, configured to physically intervene in the environment; An edge control node is electrically connected to the environmental parameter monitoring unit and the actuator. The edge control node is configured to: determine a state tolerance band containing the target state; and dynamically calculate a time vector representing the trend of environmental parameter changes based on time series data. The time vector includes the first and second derivatives of the environmental parameter changes. The time vector control gate makes decisions on the control commands of the actuator. The time vector control gate is configured as follows: when the current value of the environmental parameter is outside the state tolerance zone, and the time vector indicates that the environmental parameter is spontaneously regressing into the state tolerance zone at a trend higher than the regression threshold, the actuator is prevented from starting a control action aimed at correcting the deviation; when the current value of the environmental parameter is within the state tolerance zone, and the time vector indicates that the environmental parameter is rushing towards the boundary of the state tolerance zone at a trend higher than the breakthrough threshold, the actuator is started to perform a fine-tuning control action. The cloud server is configured to receive historical data uploaded by the edge control nodes, and based on the historical data, optimize the width of the state tolerance band, the regression threshold, and the breakthrough threshold, and then send the optimized parameters to the edge control nodes.

2. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The edge control node is further configured as follows: when the second derivative of the environmental parameter change indicated by the time vector exceeds the drastic change threshold, the judgment result of the time vector control gate is not executed immediately; the airflow circulation device in the actuator is commanded to perform a disturbance, the duration of which is 1 to 5 seconds; based on the subsequent change of the time vector after the disturbance, it is determined whether the initial time vector originates from a transient local disturbance or a steady-state environmental evolution; based on the judgment result, it is decided whether to execute the initial judgment result of the time vector control gate.

3. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The edge control node is further configured to perform a self-correction function for control performance. The self-correction function for control performance includes: recording the controlled evolution vector generated during the execution of control actions by the actuator, and the natural evolution vector generated during the period when the actuator does not perform control actions; generating a control performance signature that can represent the current actual performance of the actuator by comparing the controlled evolution vector and the natural evolution vector; and automatically adjusting the intensity or duration of subsequent fine-tuning control actions of the edge control node based on the dynamically updated control performance signature.

4. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The intensity of the fine-tuning control action is proportional to the trend intensity indicated by the time vector, and the intensity of the fine-tuning control action is five percent to twenty percent of the actuator's maximum power.

5. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The edge control node is configured to determine the specific triggering conditions of the control command based on a specific combination of the rate of change and acceleration of the environmental parameters. The conditions are as follows: when both the rate and acceleration indicate that the environmental parameters are accelerating back towards the center of the state tolerance zone, a blocking control action is triggered; when both the rate and acceleration indicate that the environmental parameters are accelerating toward the boundary of the state tolerance zone, a fine-tuning control action is triggered.

6. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The system also includes a central coordinator, which is configured not to participate in real-time control decisions, but to collect energy consumption data and environmental status data obtained by each edge control node, and adjust the global energy-saving strategy based on the data.

7. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 6, characterized in that, The central coordinator is configured to reflect the global energy-saving strategy as adjustment suggestions for parameter thresholds in each edge control node, and distribute these suggestions to the edge control nodes via the cloud server. The adjustment suggestions are updated quarterly to weekly.

8. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The edge control node is further configured to: control the rate of change of environmental parameters. and acceleration satisfy When the environmental parameters show a positive trend, it is determined that the rate of change of the environmental parameters is positive. and acceleration satisfy When environmental parameters show a negative trend, it is determined that the parameters are negative.

9. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The environmental parameter monitoring unit includes a temperature sensor, a humidity sensor, and an air quality sensor, and further includes a differential pressure sensor or a cleanliness sensor; the six constant environmental parameters include temperature, humidity, air quality, fresh air volume, airflow organization, and noise, and further include environmental differential pressure or cleanliness level.

10. The IoT-based real-time monitoring and AI intelligent control system for six constant environmental parameters as described in claim 1, characterized in that, The IoT architecture includes multiple edge control nodes that interact with a cloud server via a wireless communication network. The edge control nodes are also configured to share data with other local IoT devices via the wireless communication network. The wireless communication network includes a Wi-Fi network or a LoRa network.

Images