Outdoor awning automatic control method and system

Abstract

The present application relates to the technical field of building automation, in particular to an outdoor louver canopy automatic control method and system, comprising the following steps: based on historical and real-time environmental data, using an autoregressive integrated moving average model, analyzing the trend and seasonal variation of illumination, temperature and wind speed data, using a random forest algorithm to process the nonlinear relationship of the data, constructing a model to predict environmental changes, analyzing and predicting the future trend of illumination and temperature parameters, and generating environmental prediction data.In the present application, the autoregressive integrated moving average and random forest algorithm accurately predict environmental changes, the gradient boosting machine algorithm reduces prediction error, the long short-term memory network timely identifies structural damage and functional failure, enhances reliability and safety, the fuzzy logic control handles input uncertainty, improves control accuracy, the PID algorithm adjusts the control amount according to real-time data, realizes accurate adjustment, real-time data stream analysis monitors the state, and ensures effective execution of the strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of building automation, and in particular to an automatic control method and system for an outdoor louver canopy. Background Art

[0002] The field of building automation involves using sensors, control algorithms, and actuators to automatically adjust building environmental parameters to adapt to changes in the external environment, thereby improving the comfort and energy efficiency of living or occupied spaces. By integrating information technology, network communications, and architectural design, building automation technology enables intelligent response and management of indoor and outdoor environments, creating healthier, safer, and more energy-efficient living and working environments for users.

[0003] The automatic control method for outdoor blinds aims to automatically adjust the angle and position of the blinds based on external environmental conditions, thereby regulating outdoor light, temperature, and ventilation, enhancing user comfort, and improving energy efficiency. Automatically controlling the blinds effectively blocks excessive sunlight, reducing the temperature difference between indoors and outdoors, while also utilizing natural light and ventilation to reduce the building's energy consumption. This method aims to optimize the use of outdoor spaces through intelligent means, improving the living and user experience while promoting environmental sustainability.

[0004] Traditional outdoor awning control methods rely on preset static rules or simple sensor feedback, lack the ability to carefully analyze and predict environmental changes, and find it difficult to accurately adjust the awning to cope with rapidly changing external conditions. This static or overly simplified control strategy affects the efficiency of energy utilization and reduces the user's comfort experience. It is usually slow to respond in terms of abnormal state detection and processing, and cannot effectively prevent or respond promptly to structural damage and functional failures. It has low reliability and safety, increases maintenance costs and risks, and does not handle the uncertainty and ambiguity of input parameters in a detailed manner, resulting in inaccurate control commands, affecting the effect and accuracy of the adjustment. There are limitations in the dynamic optimization of the control strategy, making it difficult to achieve rapid adaptation to complex environmental conditions. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an automatic control method and system for an outdoor louver awning.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an outdoor blind awning automatic control method, comprising the following steps:

[0007] S1: Based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze trends and seasonal changes in light, temperature, and wind speed data. The random forest algorithm is used to process the nonlinear relationships in the data and build a model to predict environmental changes. This model analyzes and predicts future trends in light and temperature parameters and generates environmental prediction data.

[0008] S2: Based on the environmental prediction data, a gradient boosting algorithm is used to optimize the prediction error through iterative decision trees. The prediction data is analyzed to determine the opening and closing angles and states of the blinds and generate an intelligent adjustment strategy.

[0009] S3: Based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependencies in the time series data, the abnormal state of the awning's structural damage and functional failure is identified. By continuously monitoring and analyzing the awning's position, angle, and movement speed parameters, abnormal patterns are identified and abnormal state detection results are generated.

[0010] S4: Based on the intelligent adjustment strategy and the abnormal state detection result, a fuzzy logic control algorithm is adopted to process the uncertainty and fuzziness of the input parameters through the fuzzy logic controller, analyze the fuzzy range of the light intensity and temperature, output a control command, adjust the control parameters, and generate parameter adjustment information;

[0011] S5: Based on the parameter adjustment information, the louver adjustment operation is performed through the PID control algorithm, and the control amount is adjusted according to the deviation value, and the angle adjustment and opening and closing state of the louver are automatically controlled according to the environmental changes, and the louver adjustment result is generated;

[0012] S6: Based on the adjustment results of the louver, using real-time data stream analysis technology, monitor the status of the louver, analyze the louver's response speed to environmental changes and the adjustment effect, evaluate the execution effect of the adjustment strategy, identify the accuracy of the angle adjustment and the timeliness of the response, and generate a performance evaluation result;

[0013] S7: Based on the performance evaluation results, apply the Q learning algorithm, set a reward mechanism, analyze the execution effect of the control strategy, evaluate the value of the state-action pair, and dynamically adjust the awning angle adjustment and opening and closing state control strategy by updating the value function, and optimize the light and temperature response strategies to generate an optimized control strategy.

[0014] As a further scheme of the present application, the environmental prediction data comprises light intensity prediction value, temperature change trend and wind speed grade prediction, the intelligent adjustment strategy comprises louver angle adjustment instruction, opening and closing state control strategy and time arrangement adjusted according to light and temperature change, the abnormal state detection result comprises structure damage alarm, function failure identification information and safety hidden danger prompt, the parameter adjustment information comprises light intensity control range, temperature adaptability adjustment parameter and wind speed response sensitivity, the louver adjustment result comprises adjusted angle and implemented opening and closing state, the performance evaluation result comprises response time of adjustment strategy, energy efficiency ratio improvement point and user comfort index, and the optimization control strategy comprises control strategy parameter set, response time setting and energy efficiency ratio target.

[0015] As a further scheme of the present application, based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze the trend and seasonal variation of light, temperature and wind speed data, a random forest algorithm is used to process the nonlinear relationship of data, a model for predicting environmental changes is constructed, the future change trend of light and temperature parameters is analyzed and predicted, and the steps of generating environmental prediction data are as follows:

[0016] S101: Based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze time series data, the autoregressive term, difference order and moving average term of the model are determined, the trend and seasonal variation of data are captured, the dynamic process of environmental parameters changing with time is identified and simulated, and a time series analysis model is generated.

[0017] S102: Based on the time series analysis model, a random forest algorithm is used to process the nonlinear relationship of data by constructing multiple decision trees, the prediction results of the decision trees are summarized to optimize and improve the accuracy and generalization ability of the prediction model, and a nonlinear prediction model is generated.

[0018] S103: Based on the nonlinear prediction model, the influence degree of environmental parameters on future conditions is analyzed, the interaction of multiple variables is identified, the prediction ability of the model for environmental changes is optimized, and an environmental data prediction model is generated.

[0019] S104: Based on the environmental data prediction model, a cross-validation method is used to evaluate the performance of the model on unknown data, the data set is divided into multiple subsets, the model is trained and verified, the generalization ability of the model is optimized, and through grid search, the model parameter combination is optimized by traversing the given parameter grid, and light and temperature parameters are predicted, and environmental prediction data is generated.

[0020] As a further solution of the present invention, based on the environmental prediction data, a gradient boosting algorithm is used to optimize the prediction error through an iterative decision tree, and the opening and closing angles and opening and closing states of the blinds are determined by analyzing the prediction data. The specific steps of generating an intelligent adjustment strategy are as follows:

[0021] S201: Based on the environmental prediction data, a gradient boosting algorithm is used to analyze the predicted light intensity, temperature change, and wind speed parameters, and a decision tree is constructed to evaluate the importance of differentiated features, thereby selecting split points to optimize the model prediction capability, optimize the accuracy of environmental change prediction, and generate basic analysis data.

[0022] S202: Based on the basic analysis data, by comparing and analyzing energy efficiency and comfort indicators at different angles and states, using a simulation algorithm, predicting the effect of the adjustment plan, planning the opening and closing angles and states of the louver, and generating a basic adjustment plan;

[0023] S203: Based on the basic adjustment scheme, the parameters of the adjustment scheme are optimized by using a simulated annealing algorithm. By referring to the physical limitations and safety requirements of the louver canopy, the feasibility and safety of the adjustment strategy in execution are optimized to generate an optimized adjustment strategy.

[0024] S204: Based on the optimization and adjustment strategy, a genetic algorithm is used to simulate the process of natural selection, and the strategy parameters are integrated and optimized through population initialization, selection, crossover and mutation operations to generate an intelligent adjustment strategy.

[0025] As a further solution of the present invention, based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependency relationship in the time series data, the abnormal state of the awning structural damage and functional failure is identified. By continuously monitoring and analyzing the position, angle and movement speed parameters of the awning, abnormal patterns are identified, and the steps of generating abnormal state detection results are specifically as follows:

[0026] S301: Based on the louver awning operation data, a long short-term memory network is used to analyze time series data. By building a network model, setting the number of hidden layers and neurons, the network is trained to capture time series features. A cross-validation method is used to optimize the model's generalization ability and generate time series data dependency analysis results.

[0027] S302: Based on the time series data dependency analysis results, a decision tree algorithm is used to construct a tree structure model to judge and classify features, identify structural damage and functional failures, analyze the health status of the louver, and generate structural damage and functional failure identification results;

[0028] S303: Based on the structural damage and functional fault identification results, a support vector machine algorithm is applied to perform abnormal pattern analysis, and a spatial hyperplane is constructed to distinguish normal and abnormal states, thereby generating a pattern analysis result.

[0029] S304: Based on the pattern analysis results, a random forest algorithm is used to construct multiple decision trees, and the number of trees and feature selection are adjusted to optimize the accuracy of abnormal state judgment and generate abnormal state detection results.

[0030] As a further solution of the present invention, based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is adopted, and the uncertainty and fuzziness of the input parameters are processed by the fuzzy logic controller, the fuzzy intervals of light intensity and temperature are analyzed, and control commands are output to adjust the control parameters. The steps of generating parameter adjustment information are specifically as follows:

[0031] S401: Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is used to perform fuzzy processing on parameters, construct a fuzzy set and a fuzzy rule base, process the uncertainty of light intensity and temperature factors, and generate a fuzzy logic processing result;

[0032] S402: Based on the fuzzy logic processing results, a neural network is used to perform fuzzy interval analysis and optimization. The neural network receives fuzzy logic data through an input layer, extracts features from the data through a hidden layer, and provides analysis results at an output layer. The neural network also uses a back-propagation algorithm to adjust weights and biases, optimize fuzzy interval recognition and analysis, and generate fuzzy interval analysis results.

[0033] S403: Based on the fuzzy interval analysis results, a genetic algorithm is used to simulate natural selection and genetic mechanisms, set appropriate population size, crossover and mutation rates, and select an appropriate fitness function to capture the optimal solution, optimize the output of the fuzzy logic controller, and generate a control command output result;

[0034] S404: Based on the control command output result, apply the particle swarm optimization algorithm, search for control parameter configuration by simulating foraging behavior, adjust the number of particles and the number of iterations in the algorithm, adjust the speed and position, update the formula, optimize the control parameters, and generate parameter adjustment information.

[0035] As a further solution of the present invention, based on the parameter adjustment information, a PID control algorithm is used to perform an adjustment operation of the blind, and the control amount is adjusted according to the deviation value, and the angle adjustment and opening and closing state of the blind are automatically controlled according to environmental changes. The steps of generating the blind adjustment result are specifically as follows:

[0036] S501: Based on the parameter adjustment information, a PID control algorithm is used to calculate the proportional, integral, and differential coefficients of the PID controller to match the dynamic response characteristics of the louver awning. The PID parameters are adjusted by simulating a control process to adjust the controller to respond to the operating state of the louver awning and changes in the external environment, thereby generating PID control parameters.

[0037] S502: Based on the PID control parameters, a PID controller is used to calculate a deviation value according to changes in light intensity and temperature environment, dynamically adjust the angle and opening and closing state of the louver, optimize light and temperature conditions, and generate dynamic adjustment information;

[0038] S503: Based on the dynamic adjustment information, a real-time data feedback mechanism is used to monitor and adjust the PID control effect. By collecting the operation data of the louver and analyzing the deviation data during the control process, the PID parameters are adjusted to reduce the control error and generate optimized PID control information.

[0039] S504: Based on the optimized PID control information, a fuzzy adaptive control strategy is used to dynamically adjust the PID parameters, and a gradient descent method is used to generate optimization to match operation and environmental requirements, thereby generating a blind adjustment result.

[0040] As a further embodiment of the present invention, based on the adjustment results of the louver awning, the louver awning status is monitored by real-time data stream analysis technology, the louver awning's response speed to environmental changes and adjustment effect are analyzed, the execution effect of the adjustment strategy is evaluated, and the accuracy of the angle adjustment and the timeliness of the response are identified. The steps of generating the performance evaluation results are specifically as follows:

[0041] S601: Based on the adjustment result of the louver awning, using real-time data stream analysis technology to monitor the operation of the louver awning, by capturing, processing and analyzing real-time data, using time series analysis and pattern recognition technology to analyze the louver awning's response speed to environmental changes and adjustment effect, and generate real-time monitoring data;

[0042] S602: Based on the real-time monitoring data, an autoregressive moving average model is used to analyze the adjustment effect. By constructing a mathematical model between the awning state and environmental factors, the response effect of the awning adjustment strategy to environmental changes is evaluated, and an analysis result of the awning adjustment effect is generated.

[0043] S603: Based on the analysis results of the blind adjustment effect, an isolation forest algorithm is applied to detect abnormal conditions. By randomly selecting features and cut values, data points that deviate from the normal pattern are identified, potential problems are identified, and a blind abnormal condition detection result is generated.

[0044] S604: Based on the abnormal state detection result of the blind canopy, the execution effect of the adjustment strategy is evaluated. A weighted scoring method is used to quantitatively evaluate the angle adjustment accuracy and response speed indicators, analyze the effect of the adjustment strategy, and generate a performance evaluation result.

[0045] As a further solution of the present invention, based on the performance evaluation results, a Q-learning algorithm is applied to analyze the execution effect of the control strategy by setting a reward mechanism, evaluate the value of the state-action pair, and dynamically adjust the louver angle adjustment and opening and closing state control strategy by updating the value function, and optimize the light and temperature response strategy. The steps for generating the optimized control strategy are as follows:

[0046] S701: Based on the performance evaluation results, an initial model is established using a Q-learning algorithm. The state space is partitioned using a decision tree algorithm. The available strategies under differentiated states are identified through the decision tree branch structure, thereby generating a decision tree-assisted Q-learning model.

[0047] S702: Based on the decision tree-assisted Q-learning model, applying the Monte Carlo method, performing action selection and strategy evaluation, evaluating differentiated strategy benefits through random sample generation, and generating strategy value information;

[0048] S703: Based on the policy value information, use a neural network to approximate the value function, fit the value function through the deep structure and nonlinear activation function of the neural network, and generate a neural network optimized value function;

[0049] S704: Based on the value function optimized by the neural network and combined with the simulated annealing algorithm, the control strategy is optimized. By using random search and gradual cooling strategies, the angle adjustment, opening and closing states, and light and temperature response strategies of the louver are adjusted to generate an optimized control strategy.

[0050] An outdoor louver awning automatic control system, the outdoor louver awning automatic control system is used to execute the outdoor louver awning automatic control method described above, the system includes a prediction modeling module, an environmental analysis module, a strategy generation module, a health status monitoring module, a regulation execution module, and a performance evaluation module;

[0051] The environmental analysis module of the predictive modeling module uses an autoregressive integrated moving average model based on historical and real-time environmental data to analyze time series data and generate a time series analysis model by determining the autoregressive term, difference order and moving average term of the model;

[0052] The environmental analysis module is based on a time series analysis model and uses a random forest algorithm to construct multiple decision trees to process nonlinear data relationships. It also uses a cross-validation method and grid search to optimize model parameters, predict light and temperature parameters, and generate environmental prediction data.

[0053] The strategy generation module uses the gradient boosting algorithm based on environmental prediction data to analyze light intensity, temperature changes and wind speed parameters, optimizes the adjustment scheme parameters through the simulated annealing algorithm, plans the opening and closing angles and opening and closing states of the louver, and generates a basic adjustment scheme;

[0054] The health status monitoring module uses the long short-term memory network to identify structural damage and functional failures based on the louver operation data, applies the support vector machine algorithm to perform abnormal pattern analysis, and uses the random forest algorithm to optimize abnormal state judgment to generate abnormal state detection results;

[0055] The adjustment execution module uses fuzzy logic control algorithm and neural network based on the basic adjustment scheme and abnormal state detection results to analyze and optimize the fuzzy interval, and uses particle swarm optimization algorithm and PID control algorithm to adjust the control parameters, match the dynamic response characteristics of the louver, and generate PID control parameters and dynamic adjustment information;

[0056] The performance evaluation module is based on PID control parameters and dynamic adjustment information, uses fuzzy adaptive control strategy and gradient descent method to optimize parameters, uses real-time data flow analysis technology and isolation forest algorithm to evaluate the effect of adjustment strategy, combines Q learning algorithm and neural network to optimize the strategy value and generate an optimized control strategy.

[0057] Compared with the prior art, the advantages and positive effects of the present invention are:

[0058] In the present invention, by adopting the autoregressive integral moving average model and the random forest algorithm, it is possible to more accurately analyze and predict environmental changes and achieve more effective awning adjustment. The gradient boosting machine algorithm is combined to further optimize the prediction error and improve the adjustment strategy of the opening and closing angle and opening and closing state of the awning. The application of the long short-term memory network enables the abnormal state of the awning structural damage and functional failure to be identified in a timely and accurate manner, thereby improving reliability and safety. The use of the fuzzy logic control algorithm can better handle the uncertainty and ambiguity of the input parameters and improve the accuracy of the control command. Through the PID control algorithm, the control quantity is adjusted according to real-time data, making the adjustment of the awning more accurate and timely. Real-time data stream analysis technology is used to monitor the state of the awning to ensure the effective execution of the adjustment strategy. The application of the Q learning algorithm achieves rapid adaptation to complex environmental conditions through dynamic adjustment strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0060] Figure 2 This is a detailed flow chart of S1 of the present invention;

[0061] Figure 3 This is a detailed flow chart of S2 of the present invention;

[0062] Figure 4 This is a detailed flow chart of S3 of the present invention;

[0063] Figure 5 This is a detailed flow chart of S4 of the present invention;

[0064] Figure 6 This is a detailed flow chart of S5 of the present invention;

[0065] Figure 7 This is a detailed flow chart of S6 of the present invention;

[0066] Figure 8 This is a detailed flow chart of S7 of the present invention;

[0067] Figure 9 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0069] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0070] Example 1

[0071] See also Figure 1 The present invention provides a technical solution: an automatic control method for an outdoor louver awning, comprising the following steps:

[0072] S1: Based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze trends and seasonal changes in light, temperature, and wind speed data. The random forest algorithm is used to process the nonlinear relationships in the data and build a model to predict environmental changes. This model analyzes and predicts future trends in light and temperature parameters and generates environmental prediction data.

[0073] S2: Based on the environmental prediction data, the gradient boosting algorithm is used to optimize the prediction error through iterative decision trees. By analyzing the prediction data, the opening and closing angles and opening and closing states of the louver are determined, generating an intelligent adjustment strategy.

[0074] S3: Based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependencies in the time series data, the abnormal state of the awning's structural damage and functional failure is identified. By continuously monitoring and analyzing the awning's position, angle, and movement speed parameters, abnormal patterns are identified and abnormal state detection results are generated.

[0075] S4: Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is adopted. The fuzzy logic controller processes the uncertainty and fuzziness of the input parameters, analyzes the fuzzy range of light intensity and temperature, outputs control commands, adjusts control parameters, and generates parameter adjustment information;

[0076] S5: Based on the parameter adjustment information, the louver adjustment operation is performed through the PID control algorithm, and the control amount is adjusted according to the deviation value. The angle adjustment and opening and closing state of the louver are automatically controlled according to the environmental changes, and the louver adjustment result is generated;

[0077] S6: Based on the louver adjustment results, real-time data stream analysis technology is used to monitor the louver's status, analyze its response speed to environmental changes and its adjustment effect, evaluate the effectiveness of the adjustment strategy, identify the accuracy of angle adjustment and the timeliness of response, and generate performance evaluation results.

[0078] S7: Based on the performance evaluation results, apply the Q-learning algorithm, set a reward mechanism, analyze the execution effect of the control strategy, evaluate the value of the state-action pair, and dynamically adjust the awning angle adjustment and opening and closing state control strategy by updating the value function. Optimize the light and temperature response strategies to generate an optimized control strategy.

[0079] Environmental prediction data includes light intensity prediction values, temperature change trends, and wind speed level predictions; intelligent adjustment strategies include louver angle adjustment instructions, opening and closing status control strategies, and adjustment schedules based on light and temperature changes; abnormal state detection results include structural damage alarms, functional failure identification information, and safety hazard prompts; parameter adjustment information includes light intensity control range, temperature adaptability adjustment parameters, and wind speed response sensitivity; louver adjustment results include the adjusted angle and the implemented opening and closing status; performance evaluation results include the response time of the adjustment strategy, energy efficiency improvement points, and user comfort indicators; and the optimization control strategy includes the control strategy parameter set, response time setting, and energy efficiency target.

[0080] In step S1, a comprehensive analysis of environmental data is performed using an autoregressive integrated moving average model and a random forest algorithm. The autoregressive integrated moving average model performs time series analysis on environmental data such as light, temperature, and wind speed. By determining the autoregressive component, the number of differencing steps, and the moving average component of the data, it captures trends and seasonal variations in the data. The random forest algorithm analyzes the output of the autoregressive integrated moving average model to address nonlinear relationships in the data. The random forest algorithm constructs multiple decision trees and aggregates the results of each tree to enhance the model's ability to predict future environmental changes. This combined approach of using the autoregressive integrated moving average model to analyze time series characteristics and the random forest algorithm to address nonlinear relationships effectively improves prediction accuracy and provides accurate environmental prediction data for subsequent intelligent adjustment strategies.

[0081] In step S2, the gradient boosting algorithm is used to refine the intelligent adjustment strategy based on the environmental prediction data. The gradient boosting algorithm iteratively constructs a decision tree, focusing on reducing the error of the previous prediction in each iteration to optimize the overall prediction model. The algorithm evaluates the prediction results of different decision trees and selects the decision tree that minimizes prediction error for inclusion in the model. The gradient boosting algorithm not only accurately predicts future light and temperature trends but also develops adjustment plans for the awning's opening and closing angles and state based on the predicted data, improving the accuracy and adaptability of the intelligent adjustment strategy.

[0082] In step S3, the LSTM network model is used to analyze the awning's operational data. The LSTM model is specifically designed to process and predict long-term dependencies. In this step, the model accurately identifies abnormal conditions such as structural damage and functional failures by learning the dependencies in time series data such as the awning's position, angle, and speed. The LSTM model's continuous monitoring and real-time analysis capabilities enable it to promptly detect and flag any abnormal patterns in the awning's operation, providing strong technical support for awning maintenance and fault prevention.

[0083] In step S4, the control strategy of the louver is refined through the fuzzy logic control algorithm. This algorithm fully considers the uncertainty and fuzziness of environmental parameters such as light intensity and temperature, takes environmental prediction data and abnormal state detection results as input, and converts continuous environmental parameters into fuzzy values ​​by defining a series of fuzzy rules and membership functions. The fuzzy logic controller performs fuzzy reasoning based on these fuzzy values ​​and pre-set control rules to obtain fuzzy control commands. These fuzzy commands are converted into specific control operations such as the angle adjustment and opening and closing status of the louver through the defuzzification process, thereby achieving precise control of the louver, enabling the control strategy to respond more flexibly to complex and changing environmental conditions, and improving control adaptability and response accuracy.

[0084] In step S5, based on the parameter adjustment information, the PID control algorithm is used to directly perform precise adjustment operations on the blinds. According to the control parameters obtained from the fuzzy logic control algorithm, the blinds are dynamically adjusted to adapt to environmental changes. By calculating the deviation between the real-time environmental data and the expected target, the PID algorithm adjusts its proportional, integral, and differential parameters to optimize the angle and opening and closing state of the blinds. The adjustment process ensures that the blinds can maintain the best state under different environmental conditions and realizes efficient and dynamic automatic adjustment.

[0085] In step S6, real-time data stream analysis technology is used to comprehensively evaluate the execution effect of the adjustment strategy. By continuously monitoring the adjusted status and environmental response of the louver, this technology conducts real-time analysis of the louver's response speed, adjustment effect and adaptability, and evaluates the accuracy of the louver's angle adjustment and the timeliness of its response to environmental changes, ensuring that the adjustment strategy can effectively improve energy efficiency and user comfort. The implementation provides a scientific basis for the continuous optimization of the adjustment strategy and ensures that the louver's operation effect is optimal.

[0086] In step S7, based on the performance evaluation results, the Q-learning algorithm is applied to optimize the control strategy. Q-learning analyzes the effects of actions under various states and assigns a value, or Q-value, to each state-action pair. Based on the performance evaluation results, the Q-value is updated through a reward mechanism, continuously optimizing the state-action pair selection strategy. This process learns how to select the optimal action based on the current state, dynamically adjusting the awning's angle and opening / closing control strategy, as well as the light and temperature response strategies, making the overall control more intelligent and efficient. The application of the Q-learning algorithm ensures that the control strategy can be adaptively optimized to cope with complex and changing environmental conditions, improving overall performance and user experience.

[0087] See also Figure 2, based on historical and real-time environmental data, using an autoregressive integrated moving average model, analyzing the trend and seasonal variation of light, temperature, wind speed data, using a random forest algorithm to process the nonlinear relationship of the data, constructing a model to predict environmental changes, analyzing and predicting the future trend of light, temperature parameters, and generating environmental prediction data steps are as follows:

[0088] S101: Based on historical and real-time environmental data, using an autoregressive integrated moving average model, time series data analysis is performed, the autoregressive term, difference order and moving average term of the model are determined to capture the trend and seasonal variation of the data, identify and simulate the dynamic process of environmental parameters changing over time, and generate a time series analysis model;

[0089] S102: Based on the time series analysis model, using a random forest algorithm, a plurality of decision trees are constructed to process the nonlinear relationship of the data, the prediction results of the decision trees are summarized to optimize and improve the accuracy and generalization ability of the prediction model, and a nonlinear prediction model is generated;

[0090] S103: Based on the nonlinear prediction model, the influence degree of environmental parameters on future conditions is analyzed, the interaction of multiple variables is identified, the prediction ability of the model for environmental changes is optimized, and an environmental data prediction model is generated;

[0091] S104: Based on the environmental data prediction model, using cross-validation method, the performance of the model on unknown data is evaluated, the data set is divided into multiple subsets, the model is trained and verified, the generalization ability of the model is optimized, and through grid search, the parameter combination of the model is optimized, and the light and temperature parameters are predicted, and the environmental prediction data is generated.

[0092] In S101, the historical and real-time environmental data are analyzed by the autoregressive integrated moving average model, the trend and seasonal characteristics of the data are identified, and the model order is provided. The autoregressive term, difference order and moving average term of the model are determined, and the autoregressive integrated moving average model can capture the internal law of time series data. The autoregressive term captures the regression relationship of the data, the difference order is used to stabilize the non-stationary time series data, and the moving average term is used to smooth the time series. The selection of model parameters is based on statistical criteria such as Akaike information criterion or Bayesian information criterion, and the optimal parameter combination is determined through iterative optimization. After completing the model training, the model is used to simulate and predict the change of environmental parameters such as light intensity and temperature over time, and a time series analysis model is generated, which can accurately reflect the dynamic change of environmental parameters over time and provide scientific basis for subsequent decision-making.

[0093] In sub-step S102, based on the time series analysis model, a random forest algorithm is used to process the nonlinear relationships in the data and improve the accuracy and generalization ability of the prediction model. The random forest algorithm optimizes the prediction ability of the entire model by constructing multiple decision trees and aggregating the prediction results of the decision trees. When constructing the decision trees, subsets of the data and feature sets are randomly selected to increase the diversity of the model and reduce the risk of overfitting. The summary results of the decision trees are voted or averaged to improve the accuracy of the prediction. The generated nonlinear prediction model can effectively process the nonlinear characteristics in time series data and improve the accuracy of predictions of future changes in environmental conditions.

[0094] In substep S103, based on the nonlinear prediction model, the model's ability to predict environmental changes is optimized through in-depth analysis of the impact of environmental parameters on future conditions and the interactions between multiple variables. This involves a detailed analysis of the relationships between the model's input variables, identifying key variables and potential interactions to improve the model's explanatory power and predictive accuracy. By adjusting the model structure and parameters, the model is optimized to better reflect complex environmental changes. The resulting environmental data prediction model can accurately predict future changes in environmental parameters such as light intensity and temperature, providing important support for the development of environmental management and control strategies.

[0095] In sub-step S104, a cross-validation method is used to evaluate the performance of the environmental data prediction model on unknown data. By dividing the dataset into multiple subsets, alternating between using one subset as training data and another as validation data, the model is trained and validated multiple times to assess the model's stability and generalization capabilities. A grid search method is then applied to traverse a given parameter grid to find the optimal combination of model parameters and improve model performance. Through these methods, model parameters are optimized and accurate predictions are made for environmental parameters such as light and temperature. The generated environmental prediction data can not only reflect future environmental change trends but also guide decision-making in actual operations, improving the scientific nature and effectiveness of environmental regulation strategies.

[0096] See also Figure 3 As a further solution of the present invention, based on environmental prediction data, a gradient boosting algorithm is used to optimize the prediction error through an iterative decision tree. The opening and closing angles and opening and closing states of the blinds are determined by analyzing the prediction data. The specific steps for generating an intelligent adjustment strategy are as follows:

[0097] S201: Based on the environmental prediction data, the gradient boosting algorithm is used to analyze the predicted light intensity, temperature change, and wind speed parameters. By constructing a decision tree and evaluating the importance of differentiated features, split points are selected to optimize the model's prediction ability, optimize the accuracy of environmental change predictions, and generate basic analysis data.

[0098] S202: Based on the basic analysis data, by comparing and analyzing the energy efficiency and comfort indicators at different angles and states, using simulation algorithms, predict the effect of the adjustment plan, plan the opening and closing angles and states of the louver, and generate a basic adjustment plan;

[0099] S203: Based on the basic adjustment scheme, the parameters of the adjustment scheme are optimized using a simulated annealing algorithm. By referring to the physical limitations and safety requirements of the louver canopy, the feasibility and safety of the adjustment strategy in execution are optimized to generate an optimized adjustment strategy.

[0100] S204: Based on the optimization adjustment strategy, a genetic algorithm is used to simulate the process of natural selection. Through population initialization, selection, crossover and mutation operations, the strategy parameters are integrated and optimized to generate an intelligent adjustment strategy.

[0101] In sub-step S201, the gradient boosting machine algorithm is used to analyze environmental prediction data, involving light intensity, temperature changes and wind speed parameters. The gradient boosting machine constructs a decision tree model through iteration. Each iteration optimizes the prediction error of the previous one. The algorithm initializes a basic prediction model and then gradually adds decision trees to improve the model. Each decision tree learns the prediction of the residual of the previous model. In this way, the model gradually reduces the prediction error after each iteration. The algorithm evaluates the impact of each feature on the prediction in detail. By calculating the importance of the features, the optimal splitting point is selected to split the data, thereby enhancing the accuracy of the model's prediction of environmental changes. The basic analysis data generated includes a quantitative assessment of the degree of influence of each environmental parameter, providing a detailed foundation for subsequent analysis.

[0102] In sub-step S202, based on the basic analysis data, a simulation algorithm is used to compare and analyze energy efficiency and comfort indicators at different angles and states. During this process, the simulation algorithm meticulously simulates the responses of various adjustment options to environmental factors such as light, temperature, and wind speed, predicting their impact on energy efficiency and comfort. The algorithm meticulously considers the changes in various environmental parameters, using a complex calculation process to predict the specific impact of different opening and closing angles and states on the indoor environment. The algorithm then generates a basic adjustment plan that clearly identifies the optimal blind adjustment strategy, including both the opening and closing angle and the state, to optimize the indoor environmental quality.

[0103] In sub-step S203, a simulated annealing algorithm is used to carefully optimize the parameters of the basic adjustment scheme. This algorithm simulates the annealing process in a physical process, performing a global search for the optimal solution. The algorithm attempts to adjust parameters, compares the performance of the new solution with the current solution, and accepts or rejects the new solution based on a certain probability, thereby avoiding being trapped in a local optimal solution. The algorithm also considers the physical limitations and safety requirements of the louver canopy to ensure the feasibility and safety of the adjustment strategy. The resulting optimized adjustment strategy comprehensively considers multiple factors to ensure efficient and safe implementation.

[0104] In sub-step S204, the optimization adjustment strategy is synthesized and optimized through a genetic algorithm. The genetic algorithm simulates the process of natural selection and finds the adjustment strategy parameters that best suit the current environmental conditions through population initialization, selection, crossover, and mutation operations. Starting from a set of initial solutions, the fitness of each solution is evaluated and the best performing solution is selected as the parent generation for crossover and mutation to generate a new generation of solutions. After multiple generations of iterations, the algorithm gradually optimizes the strategy parameters to generate an intelligent adjustment strategy with high adaptability and efficiency. The intelligent adjustment strategy can dynamically adapt to environmental changes and automatically adjust the state of the blinds to optimize energy efficiency and indoor comfort.

[0105] See also Figure 4 As a further solution of the present invention, based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependency relationships in the time series data, the abnormal state of the awning structural damage and functional failure is identified. By continuously monitoring and analyzing the position, angle, and movement speed parameters of the awning, abnormal patterns are identified and abnormal state detection results are generated. The specific steps are as follows:

[0106] S301: Based on the louver awning operation data, a long short-term memory network is used to analyze time series data. By building a network model, setting the number of hidden layers and neurons, the network is trained to capture time series features. A cross-validation method is used to optimize the model's generalization ability and generate time series data dependency analysis results.

[0107] S302: Based on the results of the time series data dependency analysis, a decision tree algorithm is used to construct a tree structure model to judge and classify features, identify structural damage and functional failures, analyze the health status of the louver, and generate structural damage and functional failure identification results;

[0108] S303: Based on the structural damage and functional fault identification results, the support vector machine algorithm is applied to perform abnormal pattern analysis. By constructing a spatial hyperplane, normal and abnormal states are distinguished and pattern analysis results are generated;

[0109] S304: Based on the pattern analysis results, a random forest algorithm is used to construct multiple decision trees and adjust the number of trees and feature selection to optimize the accuracy of abnormal state judgment and generate abnormal state detection results.

[0110] In sub-step S301, a long-short-term memory (LSTM) network is used to perform time series analysis on the awning's operating data. The core of this step is to construct a network model, setting an appropriate number of hidden layers and neurons to capture long-term dependencies in the time series. LSTM networks are particularly suitable for processing time series data because they can learn long-term dependencies in sequence data, effectively avoiding the gradient vanishing problem of traditional neural networks in processing long-series data. The model structure is initialized, including an input layer, several hidden layers, and an output layer, with each hidden layer containing multiple LSTM network units. Through the backpropagation algorithm, network weights and biases are adjusted, the loss function is optimized, and cross-validation is used to evaluate the model's generalization ability. This allows the model parameters to be adjusted to prevent overfitting. The model is able to predict future states based on historical operating data. The generated time series data dependency analysis results reveal the inherent laws of data change over time, providing a foundation for subsequent health status analysis.

[0111] In sub-step S302, a decision tree algorithm is used to analyze the dependency relationships of time series data and perform feature identification and classification to identify structural damage and functional failures. The decision tree constructs a tree-like structure model, with nodes representing feature judgment criteria, branches representing decision outcomes, and terminal nodes representing classification results. During model training, the algorithm selects optimal features and their split points to maximize classification accuracy. Feature selection is based on information gain or Gini impurity reduction. The decision tree is recursively constructed until the stopping condition is met. The resulting structural damage and functional failure identification results provide a detailed basis for subsequent abnormal pattern analysis, revealing health issues with the awning and facilitating timely maintenance or adjustment measures.

[0112] In sub-step S303, the support vector machine algorithm is applied to perform abnormal pattern analysis. The normal state and abnormal state are separated by constructing the optimal segmentation hyperplane. The support vector machine achieves accurate classification of different categories by finding the hyperplane that can maximize the boundary between classes in the feature space. When processing the structural damage and functional failure identification results, the support vector machine uses the kernel function to map the data to a high-dimensional space to solve the nonlinear classification problem. The algorithm optimization goal is to maximize the margin. The parameters of the hyperplane are determined by solving the convex optimization problem. The generated pattern analysis results clearly distinguish between normal and abnormal states, providing a basis for formulating maintenance strategies and preventive measures.

[0113] In sub-step S304, the accuracy of abnormal state detection is optimized through the random forest algorithm. Random forest improves prediction accuracy and stability by constructing multiple decision trees and synthesizing the decision results. Each decision tree is constructed using a randomly selected feature subset. This method increases model diversity and reduces the risk of overfitting. Random forest integrates the prediction results of all decision trees through a voting mechanism and determines the final classification result by majority. Adjusting the number of trees and feature selection can optimize model performance. The generated abnormal state detection results are highly accurate and reliable, providing strong support for the maintenance and abnormality handling of the blinds and ensuring the stable operation of the system.

[0114] See also Figure 5 As a further solution of the present invention, based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is adopted. The fuzzy logic controller processes the uncertainty and fuzziness of the input parameters, analyzes the fuzzy range of light intensity and temperature, outputs control commands, and adjusts the control parameters. The steps of generating parameter adjustment information are specifically as follows:

[0115] S401: Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is used to perform fuzzy processing on parameters, construct a fuzzy set and a fuzzy rule base, process the uncertainty of light intensity and temperature factors, and generate a fuzzy logic processing result;

[0116] S402: Based on the fuzzy logic processing results, a neural network is used to perform fuzzy interval analysis and optimization. The neural network receives fuzzy logic data through the input layer, extracts features from the data through the hidden layer, and outputs analysis results. The weights and biases are adjusted through the back propagation algorithm to optimize the recognition and analysis of fuzzy intervals and generate fuzzy interval analysis results.

[0117] S403: Based on the fuzzy interval analysis results, a genetic algorithm is used to simulate natural selection and genetic mechanisms, set appropriate population size, crossover and mutation rates, and select an appropriate fitness function to capture the optimal solution, optimize the output of the fuzzy logic controller, and generate a control command output result;

[0118] S404: Based on the control command output result, the particle swarm optimization algorithm is applied to search for control parameter configuration by simulating foraging behavior, adjusting the number of particles and the number of iterations in the algorithm, adjusting the speed and position, updating the formula, optimizing the control parameters, and generating parameter adjustment information.

[0119] In sub-step S401, a fuzzy logic control algorithm is used to process data based on intelligent regulation strategies and abnormal state detection results, aiming to address the uncertainty of environmental factors such as light intensity and temperature. The fuzzy logic control algorithm fuzzifies input parameters such as light intensity and temperature, converting precise quantitative data into values ​​in fuzzy sets, such as "high," "medium," and "low." A fuzzy rule base is constructed. Rules are based on expert knowledge and practical experience and are used to describe the relationship between input and output variables. For example, a simple rule could be: "If light intensity is high and temperature is low, reduce the opening and closing angle of the awning." By evaluating all relevant fuzzy rules, the fuzzy logic controller processes the rules and defuzzifies the fuzzy output results to generate precise control commands. Even in the face of data uncertainty and ambiguity, it can still make effective regulation decisions. The generated fuzzy logic processing results provide a flexible and adaptable control strategy for the regulation system, effectively improving the method's adaptability to complex environmental changes.

[0120] In sub-step S402, fuzzy intervals are analyzed and optimized through a neural network to process the fuzzy logic data obtained from the fuzzy logic control algorithm. The neural network conducts in-depth analysis of fuzzy data through its multi-layer structure, in which the input layer receives the fuzzy logic processing results, the hidden layer performs feature extraction and pattern recognition on the data, and the output layer provides analysis results based on the learned data features. Through the backpropagation algorithm, the neural network adjusts the weights and biases between each layer to minimize the difference between the output results and the expected results, thereby optimizing the recognition and analysis capabilities of fuzzy intervals, enhancing the method's ability to process fuzzy data, and improving the accuracy and efficiency of decision-making. The generated fuzzy interval analysis results enable the system to more accurately understand and process uncertainty information, providing an important foundation for the optimization of subsequent control commands.

[0121] In sub-step S403, a genetic algorithm is used to optimize the fuzzy interval analysis results, simulating natural selection and genetic mechanisms to find the optimal control strategy. The genetic algorithm defines a fitness function to evaluate the performance of each individual, selecting the best performing individuals for crossover and mutation operations to produce a new generation of individuals. Through repeated iterations, the optimal solution is gradually approached. The selection of appropriate population size, crossover rate, and mutation rate has a key impact on the algorithm's performance. The output of the genetic algorithm optimization is fine-tuned fuzzy logic controller parameters. The generated control command output more accurately matches environmental conditions and system status, improving the adaptability and efficiency of the control strategy.

[0122] In sub-step S404, a particle swarm optimization algorithm is applied to search for the optimal control parameter configuration. The control strategy is optimized by simulating the foraging behavior of a flock of birds. The algorithm initializes a group of particles, each of which represents a solution, that is, a specific set of control parameters. By calculating the fitness value of each particle and adjusting its speed and position based on the particle's own optimal position and the optimal position of the group, the particle swarm algorithm can search for the optimal solution in the solution space. The generated parameter adjustment information provides the optimal parameter configuration for the control system, ensuring that the system maintains the highest performance and efficiency under changing environmental conditions, significantly improving the system's regulation effect and stability.

[0123] See also Figure 6 As a further solution of the present invention, based on the parameter adjustment information, the louver adjustment operation is performed through the PID control algorithm, and the control amount is adjusted according to the deviation value. The angle adjustment and opening and closing state of the louver are automatically controlled according to environmental changes. The steps of generating the louver adjustment result are specifically as follows:

[0124] S501: Based on the parameter adjustment information, a PID control algorithm is used to calculate the proportional, integral, and differential coefficients of the PID controller to match the dynamic response characteristics of the louver awning. The PID parameters are adjusted by simulating the control process to adjust the controller to respond to the operating status of the louver awning and changes in the external environment, thereby generating PID control parameters.

[0125] S502: Based on the PID control parameters, the PID controller is used to calculate the deviation value according to the changes in light intensity and temperature environment, and the angle and opening and closing state of the louver are dynamically adjusted to optimize the light and temperature conditions and generate dynamic adjustment information;

[0126] S503: Based on the dynamic adjustment information, a real-time data feedback mechanism is used to monitor and adjust the PID control effect. By collecting the operation data of the louver and analyzing the deviation data during the control process, the PID parameters are adjusted to reduce the control error and generate optimized PID control information.

[0127] S504: Based on the optimized PID control information, a fuzzy adaptive control strategy is used to dynamically adjust the PID parameters, and a gradient descent method is used to generate optimization to match the operation and environmental requirements, thereby generating a blind adjustment result.

[0128] In substep S501, the PID control algorithm calculates the proportional, integral, and differential coefficients of the PID controller based on the parameter adjustment information to match the dynamic response characteristics of the awning. The PID control algorithm adjusts these three parameters to optimize the controller's response to changes in the awning's operating status and the external environment. The proportional term directly responds to current deviations, the integral term eliminates long-term deviations, and the differential term predicts future deviations, thereby achieving fast and smooth regulation. During the simulation control process, the optimal PID parameters are determined by repeatedly adjusting the coefficients using classic regulation methods such as the Ziegler-Nichols method. This process involves multiple simulation experiments, with the parameters adjusted based on the system response after each experiment until predetermined performance indicators, such as minimal overshoot and fast settling time, are achieved. The resulting PID control parameters directly influence the awning's regulation, ensuring that it maintains optimal light and temperature conditions under various external conditions.

[0129] In sub-step S502, a PID controller and PID control parameters are used to calculate a deviation value according to changes in light intensity and temperature environment, and dynamically adjust the angle and opening and closing state of the louver. The PID controller receives real-time data from environmental sensors, such as light intensity and temperature, and compares it with the set target value to generate a deviation value. The controller dynamically adjusts the control command according to the deviation value to optimize the light and temperature conditions. The dynamic adjustment information includes the specific adjustment plan and execution instructions. The information is directly transmitted to the actuator, such as the motor control system, to adjust the angle and opening and closing state of the louver in real time, ensuring the comfort and energy efficiency of the indoor environment, and achieving the goal of energy conservation and emission reduction through precise control.

[0130] In sub-step S503, a real-time data feedback mechanism is used to monitor and adjust the PID control effect. This method collects operational data from the louver, including angle adjustment, opening and closing status, and environmental changes such as light intensity and temperature. It analyzes deviations during the control process and, based on real-time feedback, adjusts the PID parameters to reduce control errors and improve control accuracy. The optimized PID control information, which includes adjusted PID parameters that are more suitable for the current operating environment and device response characteristics, enables the method to adapt to environmental changes, continuously optimize the control strategy, and improve overall control efficiency and effectiveness.

[0131] In sub-step S504, a fuzzy adaptive control strategy is used to dynamically adjust the PID parameters and optimize them through gradient descent to better match operational and environmental requirements. The fuzzy adaptive control strategy dynamically adjusts the PID parameters by evaluating the effectiveness of PID control and environmental changes, improving control flexibility and adaptability. The gradient descent method is used to find the optimal PID parameter configuration. The influence of the parameters on the control effect is calculated and the parameters are adjusted to reduce errors. The process is iterated until the optimal parameter settings are found. The resulting louver adjustment results include the optimized angle and open / closed state, ensuring that the louver can provide optimal lighting and temperature regulation in various environments. This adaptive adjustment mechanism enables the method to continuously optimize its performance to cope with complex and changing environmental conditions.

[0132] See also Figure 7 As a further solution of the present invention, based on the adjustment results of the louver, real-time data stream analysis technology is used to monitor the status of the louver, analyze the louver's response speed to environmental changes and the adjustment effect, evaluate the execution effect of the adjustment strategy, identify the accuracy of angle adjustment and the timeliness of response, and generate performance evaluation results. The specific steps are:

[0133] S601: Based on the louver awning adjustment results, use real-time data stream analysis technology to monitor the louver awning operation. Through real-time data capture, processing and analysis, using time series analysis and pattern recognition technology, analyze the louver awning's response speed to environmental changes and adjustment effect, and generate real-time monitoring data;

[0134] S602: Based on the real-time monitoring data, the autoregressive moving average model is used to analyze the adjustment effect. By constructing a mathematical model between the awning state and environmental factors, the response effect of the awning adjustment strategy to environmental changes is evaluated, and the awning adjustment effect analysis results are generated;

[0135] S603: Based on the analysis results of the blind adjustment effect, the isolation forest algorithm is applied to detect abnormal conditions. By randomly selecting features and cut values, data points that deviate from the normal pattern are identified, potential problems are identified, and abnormal condition detection results of the blind are generated.

[0136] S604: Based on the abnormal state detection result of the blinds, the execution effect of the adjustment strategy is evaluated. A weighted scoring method is used to quantitatively evaluate the angle adjustment accuracy and response speed indicators, analyze the effect of the adjustment strategy, and generate a performance evaluation result.

[0137] In S601 substep, the awning operation is monitored using real-time data stream analysis techniques, including real-time data capture, processing and analysis, focusing on time series analysis and pattern recognition techniques. Real-time data stream includes awning operation state data such as angle adjustment, opening and closing state, and related environmental parameters such as light intensity and temperature. Through data collection, continuous information capture, time series database construction, time series analysis techniques such as autoregressive model are used to analyze the dependency between data points and identify awning response patterns to environmental changes. Pattern recognition techniques are used to extract features from data streams, identify operating patterns and abnormal patterns, and generate real-time monitoring data that details awning dynamic response characteristics and adjustment effects, providing a basis for analysis and optimization.

[0138] In S602 substep, the autoregressive moving average model is used to analyze the adjustment effect of the awning in depth. By building a mathematical model, the awning state is linked to environmental factors such as light intensity and temperature, and the effectiveness of the awning adjustment strategy in response to environmental changes is evaluated. By analyzing real-time monitoring data, the model evaluates the effectiveness of awning adjustment actions in improving environmental conditions. The autoregressive moving average model identifies the autoregressive and moving average parts of the data to predict future state changes and provide scientific basis for adjusting the adjustment strategy. The generated awning adjustment effect analysis results reveal the advantages and limitations of the adjustment strategy and point out possible improvement directions.

[0139] In S603 substep, the Isolation Forest algorithm is applied to detect abnormal states in awning operation. The Isolation Forest algorithm quickly isolates data points by randomly selecting features and cutting values, effectively identifying data points that deviate from normal patterns. It is particularly suitable for handling high-dimensional data and can identify anomalies with little or no labeled data. The algorithm evaluates the normal operating range of awning adjustment state and environmental parameters, and identifies data points that significantly deviate from the range as anomalies. The generated awning abnormal state detection results help identify potential equipment failures or deficiencies in the adjustment strategy, providing a basis for timely maintenance and adjustment.

[0140] In S604 substep, based on the awning abnormal state detection results, the weighted scoring method is used to quantitatively evaluate the execution effect of the adjustment strategy. By setting different weights, the angle adjustment accuracy, response speed and other key performance indicators are comprehensively evaluated. Weighted scoring method allows different indicators to be weighted differently according to their importance in overall performance, ensuring that the evaluation results can fully reflect the actual effect of the adjustment strategy. By collecting and analyzing relevant data, the score of each indicator is calculated and weighted, and the generated performance evaluation results detail the effect of the awning adjustment strategy, including its advantages and areas for improvement in actual operation. The results provide important information for developing more effective adjustment strategies and improving system performance.

[0141] See also Figure 8 As a further solution of the present invention, based on the performance evaluation results, a Q-learning algorithm is applied to analyze the execution effect of the control strategy by setting a reward mechanism, evaluate the value of the state-action pair, and dynamically adjust the louver angle adjustment and opening and closing state control strategy by updating the value function, and optimize the light and temperature response strategy. The steps for generating the optimized control strategy are as follows:

[0142] S701: Based on the performance evaluation results, the Q-learning algorithm is used to establish an initial model. Combined with the decision tree algorithm, the state space is partitioned. Through the decision tree branch structure, the strategies that can be adopted under differentiated states are identified, and a decision tree-assisted Q-learning model is generated.

[0143] S702: Based on the decision tree-assisted Q-learning model, the Monte Carlo method is applied to perform action selection and strategy evaluation. By generating random samples, the differentiated strategy benefits are evaluated and strategy value information is generated.

[0144] S703: Based on the policy value information, a neural network is used to approximate the value function. The value function is fitted through the deep structure and nonlinear activation function of the neural network to generate a value function optimized by the neural network.

[0145] S704: Based on the value function optimized by the neural network and combined with the simulated annealing algorithm, the control strategy is optimized. Through random search and gradual cooling strategies, the angle adjustment, opening and closing state, and light and temperature response strategies of the louver are adjusted to generate an optimized control strategy.

[0146] In substep S701, the Q-learning algorithm and decision tree algorithm are used together to establish an initial model. This aims to effectively partition the state space of the awning control system and identify feasible strategies for differentiated states. As a model-free reinforcement learning method, the Q-learning algorithm can learn optimal strategies through interaction with the environment. The establishment of the initial model relies on the decision tree algorithm. By splitting the data features into a tree structure, the complex state space is divided into more manageable and understandable subspaces. Each node represents a decision point, and each branch represents a possible action. Combined with the Q-learning algorithm, an action-value function is defined for each split state to evaluate the long-term reward of taking a specific action in a given state. Through continuous exploration and exploitation, the Q-learning algorithm updates the Q value to learn the optimal strategy. The resulting decision tree-assisted Q-learning model significantly improves the efficiency of the learning process and the quality of the strategy, providing precise decision support for the dynamic adjustment of the awning.

[0147] In sub-step S702, the Monte Carlo method is applied to action selection and strategy evaluation based on the decision tree-assisted Q-learning model. The Monte Carlo method simulates the benefits of the strategy by generating a large number of random samples, and then evaluates the effectiveness of the differentiated strategy. Different state-action paths are simulated by random sampling, and the return of each path is calculated to estimate the value of the state-action pair. It does not rely on a specific model and is suitable for processing high-dimensional and continuous state space problems. Through a large number of simulation experiments, the Monte Carlo method optimizes the value estimation of state-action pairs in the Q-learning process. The generated strategy value information provides a reliable basis for strategy optimization, ensuring the efficiency of the decision-making process and the effectiveness of the strategy.

[0148] In substep S703, a neural network is used to approximate the value function based on the policy value information. Neural networks, through their deep structure and nonlinear activation functions, can fit complex functional relationships and thus approximate the value function. The neural network's input layer receives state features, the hidden layer extracts features through multiple layers of nonlinear transformations, and the output layer provides a value estimate for each action. By adjusting network parameters through backpropagation and gradient descent, the neural network gradually learns the value function that represents the optimal policy. The resulting neural network-optimized value function provides accurate value estimates for optimizing the control strategy, ensuring the scientific nature of the decision-making process and the accuracy of the control strategy.

[0149] In sub-step S704, the control strategy is optimized in conjunction with a simulated annealing algorithm. The simulated annealing algorithm simulates the annealing process in the physical process and uses a random search and gradual cooling strategy to find the global optimal solution. The algorithm starts at a higher temperature and randomly selects states and actions for exploration. By comparing the values ​​of new and old solutions and accepting the new solution based on a certain probability, the temperature is gradually lowered, and the search range of the solution space is reduced until the optimal or near-optimal solution is found. Combined with the value function optimized by the neural network, the simulated annealing algorithm adjusts the angle adjustment, opening and closing state, and light and temperature response strategies of the louver. The generated optimized control strategy increases the stability and reliability of the strategy while ensuring the adjustment effect, providing an efficient adaptive adjustment solution for the louver system.

[0150] See also Figure 9 An outdoor blind awning automatic control system is used to execute the above-mentioned outdoor blind awning automatic control method. The system includes a prediction modeling module, an environmental analysis module, a strategy generation module, a health status monitoring module, an adjustment execution module, and a performance evaluation module.

[0151] The environmental analysis module of the predictive modeling module uses the autoregressive integral moving average model to analyze time series data based on historical and real-time environmental data. By determining the autoregressive term, difference order and moving average term of the model, a time series analysis model is generated;

[0152] The environmental analysis module is based on a time series analysis model and uses a random forest algorithm to construct multiple decision trees to process nonlinear data relationships. It also uses cross-validation and grid search to optimize model parameters, predict light and temperature parameters, and generate environmental prediction data.

[0153] The strategy generation module uses a gradient boosting algorithm based on environmental prediction data to analyze light intensity, temperature changes, and wind speed parameters. It then optimizes the adjustment plan parameters using a simulated annealing algorithm, plans the opening and closing angles and states of the louver, and generates a basic adjustment plan.

[0154] The health status monitoring module uses the long-short-term memory network to identify structural damage and functional failures based on the louver operation data. It also applies the support vector machine algorithm to perform abnormal pattern analysis and uses the random forest algorithm to optimize abnormal state judgment and generate abnormal state detection results.

[0155] Based on the basic adjustment scheme and abnormal state detection results, the adjustment execution module uses fuzzy logic control algorithm and neural network to analyze and optimize the fuzzy interval, and uses particle swarm optimization algorithm and PID control algorithm to adjust the control parameters, match the dynamic response characteristics of the louver, and generate PID control parameters and dynamic adjustment information;

[0156] The performance evaluation module is based on PID control parameters and dynamic adjustment information, and uses fuzzy adaptive control strategy and gradient descent method to optimize parameters. It uses real-time data stream analysis technology and isolation forest algorithm to evaluate the effect of the adjustment strategy. It combines Q learning algorithm and neural network to optimize the strategy value and generate an optimized control strategy.

[0157] The autoregressive integral moving average model of the predictive modeling module is used to analyze historical and real-time environmental data, generate a time series analysis model, and effectively capture the trends and seasonal changes of environmental parameters such as light intensity and temperature. This not only improves the accuracy of environmental change predictions, but also provides scientific data support for subsequent control strategies. The random forest algorithm of the environmental analysis module optimizes the nonlinear relationship of the data by constructing multiple decision trees. Combined with cross-validation and grid search to optimize model parameters, it improves the accuracy of environmental parameter prediction and the generalization ability of the model. Accurate environmental prediction provides a reliable data basis for the automatic adjustment of the awning, enabling the awning to maintain the best state under different environmental conditions, improving the comfort of indoor light and temperature, and also contributing to energy conservation and emission reduction. The strategy generation module uses the gradient boosting machine algorithm and simulated annealing algorithm, which not only takes into account multiple environmental parameters such as light intensity and temperature changes, but also optimizes the parameters of the adjustment scheme so that the opening and closing angles and opening and closing states of the awning can accurately match environmental requirements. This intelligent basic adjustment solution significantly improves the efficiency and comfort of outdoor awnings, while reducing wear and tear caused by frequent adjustments and extending their service life. The health monitoring module, combining long-short-term memory networks and support vector machine algorithms, effectively identifies structural damage and functional failures. Using anomaly detection optimized by a random forest algorithm, it promptly detects and warns of faults, ensuring stable system operation and user safety. Real-time health monitoring and anomaly detection significantly reduce maintenance costs and the risk of unplanned downtime. The adjustment execution module precisely controls the dynamic response characteristics of the awning using fuzzy logic control algorithms and neural networks. It adjusts control parameters using particle swarm optimization and PID control algorithms, enabling fine-grained adjustment of the awning. This precise control not only improves the speed and accuracy of adjustment but also enables the awning to more flexibly adapt to environmental changes, enhancing user comfort and satisfaction. The performance evaluation module continuously optimizes the control strategy using a fuzzy adaptive control strategy and gradient descent. It then combines real-time data stream analysis with the isolation forest algorithm to comprehensively evaluate the effectiveness of the adjustment strategy, ensuring long-term stable operation and sustained performance improvement. This continuous performance optimization and evaluation mechanism ensures the system's adaptability to future environmental changes and technological advancements, maintaining its advanced nature and competitiveness.

[0158] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A method for automatically controlling an outdoor louver awning, characterized in that: The following steps are involved: Based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze trends and seasonal changes in light, temperature, and wind speed data. The random forest algorithm is used to process the nonlinear relationship of the data and build a model to predict environmental changes. This model analyzes and predicts future trends in light and temperature parameters and generates environmental prediction data. Based on the environmental prediction data, a gradient boosting algorithm is used to optimize the prediction error through iterative decision trees, and the opening and closing angles and opening and closing states of the louver are determined by analyzing the prediction data to generate an intelligent adjustment strategy; Based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependency relationships in the time series data, the abnormal state of the awning's structural damage and functional failure is identified. By continuously monitoring and analyzing the awning's position, angle, and movement speed parameters, abnormal patterns are identified and abnormal state detection results are generated. Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is adopted to process the uncertainty and fuzziness of input parameters through a fuzzy logic controller, analyze the fuzzy intervals of light intensity and temperature, output control commands, adjust control parameters, and generate parameter adjustment information; Based on the parameter adjustment information, the louver adjustment operation is performed through the PID control algorithm, and the control amount is adjusted according to the deviation value, and the angle adjustment and opening and closing state of the louver are automatically controlled according to environmental changes to generate the louver adjustment result; Based on the louver adjustment results, real-time data stream analysis technology is used to monitor the status of the louver, analyze the louver's response speed to environmental changes and the adjustment effect, evaluate the execution effect of the adjustment strategy, identify the accuracy of angle adjustment and the timeliness of response, and generate performance evaluation results; Based on the performance evaluation results, the Q-learning algorithm is applied to analyze the execution effect of the control strategy by setting a reward mechanism, evaluate the value of the state-action pair, and dynamically adjust the awning angle adjustment and opening and closing state control strategies by updating the value function. The light and temperature response strategies are also optimized to generate an optimized control strategy.

2. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: The environmental prediction data includes light intensity prediction values, temperature change trends, and wind speed level predictions; the intelligent adjustment strategy includes louver angle adjustment instructions, opening and closing state control strategies, and time schedules for adjustments based on light and temperature changes; the abnormal state detection results include structural damage alarms, functional failure identification information, and safety hazard prompts; the parameter adjustment information includes light intensity control range, temperature adaptability adjustment parameters, and wind speed response sensitivity; the louver adjustment results include the adjusted angle and implemented opening and closing state; the performance evaluation results include the response time of the adjustment strategy, energy efficiency improvement points, and user comfort indicators; and the optimization control strategy includes a control strategy parameter set, response time settings, and energy efficiency targets.

3. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on historical and real-time environmental data, an autoregressive integrated moving average model is used to analyze the trends and seasonal changes in light, temperature, and wind speed data. The random forest algorithm is used to process the nonlinear relationship of the data and build a model to predict environmental changes. The model analyzes and predicts the changing trends of future light and temperature parameters. The specific steps for generating environmental prediction data are as follows: Based on historical and real-time environmental data, the autoregressive integrated moving average model is used to analyze time series data. By determining the autoregressive term, difference order, and moving average term of the model, the trend and seasonal changes of the data are captured, the dynamic process of environmental parameters changing over time is identified and simulated, and a time series analysis model is generated. Based on the time series analysis model, a random forest algorithm is used to construct multiple decision trees to process the nonlinear relationship of the data. By summarizing the prediction results of the decision trees, the accuracy and generalization ability of the prediction model are optimized and improved to generate a nonlinear prediction model; Based on the nonlinear prediction model, by analyzing the impact of environmental parameters on future conditions, identifying the interaction of multiple variables, optimizing the model's ability to predict environmental changes, and generating an environmental data prediction model; Based on the environmental data prediction model, a cross-validation method is used to evaluate the performance of the model on unknown data. By dividing the data set into multiple subsets, the model is trained and verified, and the generalization ability of the model is optimized. Through grid search, by traversing the given parameter grid, the model parameter combination is optimized, and illumination and temperature parameter predictions are performed to generate environmental prediction data.

4. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the environmental prediction data, the gradient boosting algorithm is combined with an iterative decision tree to optimize the prediction error. The opening and closing angles and opening and closing states of the blinds are determined by analyzing the prediction data. The specific steps for generating an intelligent adjustment strategy are as follows: Based on the environmental prediction data, a gradient boosting algorithm is used to analyze the predicted light intensity, temperature change, and wind speed parameters. By constructing a decision tree and evaluating the importance of differentiated features, split points are selected to optimize the model prediction ability, optimize the accuracy of environmental change prediction, and generate basic analysis data. Based on the basic analysis data, by comparing and analyzing the energy efficiency and comfort indicators under different angles and states, using simulation algorithms, predicting the effect of the adjustment plan, planning the opening and closing angles and states of the louver, and generating a basic adjustment plan; Based on the basic adjustment scheme, the parameters of the adjustment scheme are optimized through a simulated annealing algorithm. By referring to the physical limitations and safety requirements of the louver canopy, the feasibility and safety of the adjustment strategy in execution are optimized to generate an optimized adjustment strategy. Based on the optimization and adjustment strategy, a genetic algorithm is used to simulate the process of natural selection, and through population initialization, selection, crossover and mutation operations, strategy parameters are integrated and optimized to generate an intelligent adjustment strategy.

5. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the awning operation data, a long short-term memory network is used to analyze the awning operation data. By learning the dependency relationship in the time series data, the abnormal state of the awning's structural damage and functional failure is identified. By continuously monitoring and analyzing the position, angle and movement speed parameters of the awning, abnormal patterns are identified. The specific steps for generating abnormal state detection results are as follows: Based on the operation data of the louver tent, a long short-term memory network is used to analyze time series data. By building a network model, setting the number of hidden layers and neurons, the network is trained to capture time series characteristics. The cross-validation method is used to optimize the generalization ability of the model and generate time series data dependency analysis results. Based on the time series data dependency analysis results, a decision tree algorithm is used to construct a tree structure model to judge and classify features, identify structural damage and functional failures, analyze the health status of the louver, and generate structural damage and functional failure identification results; Based on the structural damage and functional fault identification results, a support vector machine algorithm is applied to perform abnormal pattern analysis, and a spatial hyperplane is constructed to distinguish normal and abnormal states, thereby generating pattern analysis results; Based on the pattern analysis results, the random forest algorithm is used to construct multiple decision trees and adjust the number of trees and feature selection to optimize the accuracy of abnormal state judgment and generate abnormal state detection results.

6. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is adopted. The uncertainty and fuzziness of the input parameters are processed by the fuzzy logic controller, the fuzzy intervals of light intensity and temperature are analyzed, and control commands are output to adjust the control parameters. The steps of generating parameter adjustment information are as follows: Based on the intelligent adjustment strategy and abnormal state detection results, a fuzzy logic control algorithm is used to perform fuzzy processing of parameters, construct a fuzzy set and a fuzzy rule base, process the uncertainty of light intensity and temperature factors, and generate fuzzy logic processing results; Based on the fuzzy logic processing results, a neural network is used to perform fuzzy interval analysis and optimization. The neural network receives fuzzy logic data through the input layer, extracts features from the data through the hidden layer, and gives analysis results at the output layer. The weights and biases are adjusted through the back propagation algorithm to optimize the recognition and analysis of fuzzy intervals and generate fuzzy interval analysis results. Based on the fuzzy interval analysis results, a genetic algorithm is used to simulate natural selection and genetic mechanisms, set appropriate population size, crossover and mutation rates, and select an appropriate fitness function to capture the optimal solution, optimize the output of the fuzzy logic controller, and generate a control command output result; Based on the control command output result, a particle swarm optimization algorithm is applied to search for control parameter configuration by simulating foraging behavior, adjust the number of particles and the number of iterations in the algorithm, adjust the speed and position, update the formula, optimize the control parameters, and generate parameter adjustment information.

7. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the parameter adjustment information, the louver adjustment operation is performed through the PID control algorithm, and the control amount is adjusted according to the deviation value. The angle adjustment and opening and closing state of the louver are automatically controlled according to environmental changes. The steps for generating the louver adjustment result are specifically as follows: Based on the parameter adjustment information, a PID control algorithm is used to match the dynamic response characteristics of the louver by calculating the proportional, integral, and differential coefficients of the PID controller. The PID parameters are adjusted by simulating the control process to adjust the controller to respond to the operating state of the louver and changes in the external environment, thereby generating PID control parameters. Based on the PID control parameters, a PID controller is used to calculate the deviation value according to the changes in light intensity and temperature environment, dynamically adjust the angle and opening and closing state of the louver, optimize the light and temperature conditions, and generate dynamic adjustment information; Based on the dynamic adjustment information, a real-time data feedback mechanism is used to monitor and adjust the PID control effect. By collecting the operation data of the louver canopy and analyzing the deviation data during the control process, the PID parameters are adjusted to reduce the control error and generate optimized PID control information. Based on the optimized PID control information, a fuzzy adaptive control strategy is used to dynamically adjust the PID parameters, and a gradient descent method is used to generate optimization to match operation and environmental requirements and generate a blind adjustment result.

8. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the louver adjustment results, the louver status is monitored using real-time data stream analysis technology, the louver response speed and adjustment effect to environmental changes are analyzed, the execution effect of the adjustment strategy is evaluated, and the accuracy of the angle adjustment and the timeliness of the response are identified. The specific steps for generating the performance evaluation results are as follows: Based on the adjustment results of the louver awning, real-time data stream analysis technology is used to monitor the operation of the louver awning. By capturing, processing and analyzing real-time data, time series analysis and pattern recognition technology are used to analyze the response speed and adjustment effect of the louver awning to environmental changes, and generate real-time monitoring data. Based on the real-time monitoring data, an autoregressive moving average model is used to analyze the adjustment effect. By constructing a mathematical model between the awning state and environmental factors, the response effect of the awning adjustment strategy to environmental changes is evaluated, and the analysis results of the awning adjustment effect are generated; Based on the analysis results of the blind adjustment effect, an isolation forest algorithm is applied to detect abnormal conditions, and data points that deviate from the normal pattern are identified by randomly selecting features and cutting values, thereby identifying potential problems and generating blind abnormal condition detection results. Based on the abnormal state detection results of the louver canopy, the execution effect of the adjustment strategy is evaluated, and a weighted scoring method is used to quantitatively evaluate the angle adjustment accuracy and response speed indicators, analyze the effect of the adjustment strategy, and generate a performance evaluation result.

9. The automatic control method for outdoor venetian blinds according to claim 1, characterized in that: Based on the performance evaluation results, the Q-learning algorithm is applied to analyze the execution effect of the control strategy by setting a reward mechanism, evaluating the value of the state-action pair, and dynamically adjusting the awning angle adjustment and opening and closing state control strategies by updating the value function. The light and temperature response strategies are also optimized. The specific steps for generating the optimized control strategy are as follows: Based on the performance evaluation results, the Q-learning algorithm is used to establish an initial model. The decision tree algorithm is combined to divide the state space. The strategies that can be adopted under the differentiated state are identified through the decision tree branch structure, and a decision tree-assisted Q-learning model is generated. Based on the decision tree-assisted Q-learning model, the Monte Carlo method is applied to perform action selection and strategy evaluation. By generating random samples, the differentiated strategy benefits are evaluated and strategy value information is generated. Based on the strategy value information, a neural network is used to approximate the value function, and the value function is fitted through the deep structure and nonlinear activation function of the neural network to generate a value function optimized by the neural network; Based on the value function optimized by the neural network and combined with the simulated annealing algorithm, the control strategy is optimized. Through random search and gradual cooling strategies, the angle adjustment, opening and closing states, and light and temperature response strategies of the louver are adjusted to generate an optimized control strategy.

10. An outdoor louver awning automatic control system, characterized in that: The automatic control method for an outdoor venetian blind according to any one of claims 1 to 9, wherein the system comprises a prediction modeling module, an environmental analysis module, a strategy generation module, a health status monitoring module, a regulation execution module, and a performance evaluation module; The environmental analysis module of the predictive modeling module uses an autoregressive integrated moving average model based on historical and real-time environmental data to analyze time series data and generate a time series analysis model by determining the autoregressive term, difference order and moving average term of the model; The environmental analysis module is based on a time series analysis model and adopts a random forest algorithm to construct multiple decision trees to process nonlinear data relationships. It also uses a cross-validation method and grid search to optimize model parameters, predict light and temperature parameters, and generate environmental prediction data. The strategy generation module uses the gradient boosting algorithm based on environmental prediction data to analyze light intensity, temperature changes and wind speed parameters, optimizes the adjustment scheme parameters through the simulated annealing algorithm, plans the opening and closing angles and opening and closing states of the louver, and generates a basic adjustment scheme; The health status monitoring module uses the long short-term memory network to identify structural damage and functional failures based on the louver operation data, applies the support vector machine algorithm to perform abnormal pattern analysis, and uses the random forest algorithm to optimize abnormal state judgment to generate abnormal state detection results; The adjustment execution module uses fuzzy logic control algorithm and neural network based on the basic adjustment scheme and abnormal state detection results to analyze and optimize the fuzzy interval, and uses particle swarm optimization algorithm and PID control algorithm to adjust the control parameters, match the dynamic response characteristics of the louver, and generate PID control parameters and dynamic adjustment information; The performance evaluation module is based on PID control parameters and dynamic adjustment information, uses fuzzy adaptive control strategy and gradient descent method to optimize parameters, uses real-time data flow analysis technology and isolation forest algorithm to evaluate the effect of adjustment strategy, combines Q learning algorithm and neural network to optimize the strategy value and generate an optimized control strategy.

Images