An intelligent hierarchical water-saving irrigation control system based on internet of things

Abstract

The application discloses an intelligent hierarchical water-saving irrigation control system based on Internet of Things, and is characterized in that the system comprises the following steps: integrating a temperature and humidity sensor, a pressure sensor and a wind speed sensor in an intelligent sensor node to perform multi-modal data collection and obtain multi-modal data; adopting an automatic closed-loop dual-mode control algorithm combining switch control and fuzzy control to send the multi-modal data to an edge computing gateway; sending data in the edge computing gateway to a cloud platform by using a LoRa network; performing NB-IoT and 5G dual-mode communication between the gateway and the cloud platform, and automatically enabling a local decision mode when communication is interrupted; performing hierarchical control on multi-stage actuators by a core control module in combination with an adaptive learning rate and a regularization technique; and performing dynamic hierarchical logic, that is, dividing irrigation priorities according to a plant growth cycle, real-time environmental data and a prediction model.

Description

Technical Field

[0001] This invention relates to the field of intelligent irrigation technology, and in particular to an intelligent tiered water-saving irrigation control system based on the Internet of Things. Background Technology

[0002] Water-saving irrigation refers to irrigation measures that maximize yield or benefit with minimal water consumption, that is, maximizing crop yield and output value per unit of irrigation water. Water-saving irrigation technology is an important agricultural water-saving measure, significant for solving water shortage problems and promoting sustainable agricultural development. High-efficiency water-saving irrigation intelligent control systems, also known as intelligent irrigation control systems or smart irrigation systems, are modern irrigation management systems integrating automatic control technology, sensor technology, communication technology, and computer technology. Traditional irrigation systems typically rely on traditional electricity supplies, which are inefficient, have a significant environmental impact, and result in substantial energy consumption. This not only increases agricultural production costs but may also lead to additional environmental burdens.

[0003] Therefore, it is necessary to provide an intelligent, tiered water-saving irrigation control system based on the Internet of Things to solve the above-mentioned technical problems. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an intelligent tiered water-saving irrigation control system based on the Internet of Things, characterized in that it includes:

[0007] Temperature and humidity sensors, pressure sensors, and wind speed sensors are integrated into the smart sensor node to acquire multimodal data and obtain multimodal data.

[0008] An automatic closed-loop dual-mode control algorithm combining switch control and fuzzy control is adopted to fuse the multimodal data and send it to the edge computing gateway;

[0009] Data from the edge computing gateway is sent to the cloud platform via the LoRa network; the gateway and the cloud platform communicate via NB-IoT and 5G dual-mode communication, and the local decision-making mode is automatically activated when communication is interrupted.

[0010] By combining adaptive learning rate and regularization techniques, multi-level actuators are controlled in stages through the core control module.

[0011] Dynamic tiered logic: Irrigation priorities are determined based on plant growth cycles, real-time environmental data, and prediction models.

[0012] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system described in this invention, the actuator has a built-in pressure sensor and flow meter. When a pipeline leak is detected, the valve in the corresponding area is automatically closed and the fault point is marked on the GIS map.

[0013] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system described in this invention, an LSTM neural network is introduced to predict soil moisture change trends, and a random forest algorithm is combined to dynamically update the tiered threshold.

[0014] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system described in this invention, it includes a human-computer interaction page, provides a hierarchical permission management interface, and supports data collaboration among farmers, agricultural experts, and water authorities.

[0015] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system described in this invention, a 6-level irrigation priority system is established by weighted integration of crop growth period, weather forecast, and soil moisture.

[0016] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system of the present invention, it includes a drive water pump, which is connected to a control module, and the control module is connected to a WiFi module.

[0017] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system of the present invention, it further includes an irrigation module, which includes a relay and a water pump, and automatically controls the water pump to irrigate according to the irrigation duration output by the fuzzy controller.

[0018] As a preferred embodiment of the IoT-based intelligent tiered water-saving irrigation control system of the present invention, the dual-mode control algorithm includes protection control, switching control and fuzzy control.

[0019] The beneficial effects of this invention are as follows: By combining the dual-mode control algorithm with the control module to control the water pump, and by using data collected by the sensor to perform tiered irrigation for soils under different conditions, the water-saving rate is improved. In addition, the energy consumption of the pumping station is effectively reduced by tiered and timed irrigation. At the same time, intelligent tiered irrigation based on soil conditions at different growth stages of crops effectively increases crop yield. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] in:

[0022] Figure 1 A flowchart of water demand prediction for the PSO-LSTM model of an IoT-based intelligent tiered water-saving irrigation control system, as provided in one embodiment of the present invention.

[0023] Figure 2 A functional structure flowchart of an IoT-based intelligent tiered water-saving irrigation control system provided by an embodiment of the present invention;

[0024] Figure 3 A fuzzy control system structure block diagram of an IoT-based intelligent hierarchical water-saving irrigation control system provided by an embodiment of the present invention;

[0025] Figure 4 A structural block diagram of a Mamdani-type fuzzy control system for an IoT-based intelligent hierarchical water-saving irrigation control system provided by an embodiment of the present invention;

[0026] Figure 5 A fuzzy controller flowchart of an IoT-based intelligent tiered water-saving irrigation control system provided by an embodiment of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0028] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0029] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0030] Example 1

[0031] Reference Figure 1-5According to an embodiment of the present invention, an intelligent tiered water-saving irrigation control system based on the Internet of Things is characterized by comprising:

[0032] Temperature and humidity sensors, pressure sensors, and wind speed sensors are integrated into the smart sensor node to acquire multimodal data and obtain multimodal data.

[0033] An automatic closed-loop dual-mode control algorithm combining switch control and fuzzy control is adopted to fuse the multimodal data and send it to the edge computing gateway;

[0034] Data from the edge computing gateway is sent to the cloud platform via the LoRa network; the gateway and the cloud platform communicate via NB-IoT and 5G dual-mode communication, and the local decision-making mode is automatically activated when communication is interrupted.

[0035] By combining adaptive learning rate and regularization techniques, multi-level actuators are controlled in stages through the core control module.

[0036] Dynamic tiered logic: Irrigation priorities are determined based on plant growth cycles, real-time environmental data, and prediction models.

[0037] Specifically, irrigation demand is quantified into six levels (Level 1 emergency irrigation to Level 6 delayed irrigation). Pipelines are laid in each of these six levels, with a pumping station installed in each area. Each pumping station is individually controlled, and the master switch for all pumping stations is located in the control system operating platform. A dual-mode control algorithm intelligently turns all pumping stations on and off. An LSTM neural network is introduced to predict soil moisture changes, combined with a random forest algorithm to dynamically update the tiered thresholds. The control system operating platform features an interactive human-machine interface with a hierarchical permission management system, supporting data collaboration among farmers, agricultural experts, and water authorities. The system weights and fuses data on crop growth stages, weather forecasts, and soil moisture in the six areas. Data collected by smart sensors in each area is transmitted to the control system operating platform. The sensor data is classified and analyzed, and based on the data standards set for each area, the dual-mode control algorithm calculates whether the pumps need to be turned on. The pump output can be automatically adjusted based on the crop growth in each area. The system includes a drive pump connected to a control module, which in turn is connected to a WiFi module. It also includes an irrigation module, which includes a relay and a water pump, and automatically controls the water pump to irrigate according to the irrigation duration output by the fuzzy controller.

[0038] Based on the LSTM neural network model, this application introduces the Particle Swarm Optimization (PSO) algorithm to adaptively optimize two hyperparameters of the LSTM model: the initial learning rate and the number of neurons in the LSTM layer. This enhances the matching degree between data features and network topology, and further improves the accuracy of irrigation water demand prediction by constructing a PSO-LSTM prediction model.

[0039] The PSO optimization algorithm is an intelligent algorithm inspired by the collective behavior of biological groups to solve optimization problems. This algorithm involves fewer parameters, is easy to converge and is simple to operate. The basic idea of ​​this method is to seek the optimal solution through cooperation and information sharing among individuals

[55] . Each particle has two attributes: velocity V and position X, where V represents the speed of particle movement and X represents the direction of particle movement. First, the particle swarm is randomly initialized in the solution space, and each particle is randomly assigned a position and velocity. Iterative optimization is then performed. During this process, two optimal solutions will be generated, namely the individual extreme value generated during the iteration of the individual and the global extreme value generated during the iteration of the entire particle swarm. The particle updates its position and velocity by tracking these two extreme values ​​through the defined fitness function. By performing multiple iterations to determine whether the preset target fitness value has been reached, the optimal solution can be obtained. The particle velocity update formula can be obtained from equation (1-1), and the particle position update formula can be obtained from equation (1-2).

[0040]

[0041]

[0042] In equations (1-1) and (1-2), i = 1, 2, ..., N, where N is the particle swarm size; d = 1, 2, ..., D, where D is the particle dimension; k is the current iteration number; w is the inertia weight; c1 is the individual learning factor; c2 is the swarm learning factor; and r1 and r2 are random numbers in the interval [0, 1].

[0043] This application uses the initial learning rate of the LSTM model and the number of neurons in the LSTM layer as optimization variables for particles in PSO. The PSO-LSTM model algorithm flow is as follows:

[0044] (1) Data preprocessing: In order to reduce the impact of input data scale factors on the training effect of LSTM model, the sample data needs to be normalized and the units should be unified; 80% of the experimental data is divided into training set and the remaining 20% ​​is divided into test set.

[0045] (2) Determine the initial learning rate of the LSTM model and the number of neurons in the LSTM layer as hyperparameters for PSO optimization, and initialize the particle swarm parameters, including the population size, learning weights, initial velocity and position of particles, and maximum number of iterations.

[0046] (3) Randomly initialize the initial learning rate and the number of hidden layer neurons of the LSTM model within a given range, construct the LSTM model with the parameters corresponding to each particle, train the model using the training set, and verify the training results of the model using the test set; the fitness function of each particle is the root mean square error between the true value and the predicted value in the test set, and the fitness value of each particle is calculated in this way.

[0047] (4) Determine the local optimal particle position gbest and the global optimal particle position pbest by comparing the particle fitness values; and update the particle velocity and position according to the PSO algorithm.

[0048] (5) Determine the termination condition of the optimization iteration. The LSTM prediction model will use the optimal parameter values ​​found under the maximum number of iterations for prediction. The optimized LSTM hyperparameters are used to construct an irrigation water demand prediction model based on PSO-LSTM. The model is trained using the training set and the prediction model effect is verified using the test set to realize irrigation water demand prediction. At the same time, the prediction results are compared, analyzed and summarized according to various evaluation indicators. If the condition is not met, return to step (4). The water demand prediction process based on the PSO-LSTM model constructed in this application is as follows: Figure 1 As shown.

[0049] During the PSO parameter initialization process, relevant parameters need to be set, mainly including four parameters: the number of particles in the population, the learning factor, the maximum number of iterations, and the inertia weight. The PSO algorithm initialization parameter settings are shown in Table 1.

[0050] Table 1

[0051] Population size Maximum number of iterations Learning factor Inertia weights sizepop maxgen [c1,c2] [wini,wend] 10 10 [1.5,1.5] [0.9,0.4]

[0052] To further improve the accuracy of the orchard irrigation water demand prediction model, this application introduces the PSO algorithm without changing the LSTM model network structure, thereby optimizing the hyperparameters of the LSTM model. The LSTM and PSO-LSTM models have been trained separately and experiments have been completed on the test set. The experiments have been evaluated using four regression prediction model evaluation indicators. The comparison of water demand prediction results based on LSTM and PSO-LSTM is shown in Table 2.

[0053] Table 2

[0054]

[0055]

[0056] The water demand prediction trend based on the PSO-LSTM model is closer to the actual water demand, and the fitting degree is higher than that of the LSTM model. Compared with the LSTM model, the PSO-LSTM model has lower values ​​for the four evaluation indicators RMSE, MSE, MAE and MAPE, and the prediction error is reduced by 37.8%, 61.3%, 44.1% and 43.9% respectively. The prediction result has smaller error and higher water demand prediction accuracy, which can better provide theoretical basis and data support for scientific guidance of orchard irrigation water use and rational planning of irrigation water resources.

[0057] By analyzing the system's functions, the intelligent agricultural irrigation system is divided into three parts. First, the data acquisition system utilizes high-precision sensors to improve the accuracy and stability of data collection. The wired transmission method better meets the system's demand for large-scale data transmission, ensuring stable operation. Second, a fast-response fuzzy control system improves irrigation efficiency and management level. Finally, the management system enables human-computer interaction, allowing users to remotely monitor crop environmental data and perform corresponding controls without leaving their homes. The system's functional structure is as follows: Figure 2 As shown, the system's data acquisition, transmission, and storage functions include soil moisture and air temperature, which are crucial factors in determining whether crops need irrigation, and nitrogen, phosphorus, and potassium concentrations and pH levels, which are important factors in assessing crop nutrient levels. This provides users with scientific advice for more rational watering and fertilization. The system connects the sensors to a microcomputer via a 485-type sensor and a 485-to-USB converter. Through code writing and database design, the system achieves data acquisition, transmission, and storage.

[0058] Data monitoring function: The microcomputer sends a data acquisition command to the sensor every second, and the sensor begins to collect data in real time on soil moisture, air temperature and humidity, nitrogen, phosphorus and potassium, and pH. The collected data is transmitted back to the microcomputer via the Modbus-RTU communication protocol. The microcomputer then uses the HTTP communication protocol to send the collected data to the server, which then writes it into its database. The front end sends GET requests to the server to retrieve the data and uses ECharts to create graphs, allowing users to clearly view various data points related to the on-site environment.

[0059] Control Functions: The system's control functions are divided into intelligent control and remote control. Intelligent control calculates the irrigation duration using a fuzzy control algorithm and automatically controls the water pump for irrigation. Remote control allows users to directly click on the water pump switch in a webpage, sending an HTTP request to the microcomputer main control system. The microcomputer then controls the high and low levels of the GPIO pins according to the received instructions to control the water pump.

[0060] Fuzzy control is an effective control strategy that simulates human experience and language systems. The fuzzy control method uses computers to simulate the human language system, achieving the effect of human control experience, but it is not a brain-like controller; it is a nonlinear control method. It can describe things that cannot be accurately expressed by mathematical models, reflecting the fuzziness of language systems, such as the amount of rain, the temperature of the weather, a person's weight, height, and age. Soil moisture, as the controlled object of this system, has the characteristics of large inertia and lag. Fuzzy control can handle fuzzy uncertainty and fuzzy rules in control problems, exhibiting strong robustness and adaptability. Therefore, applying fuzzy control to intelligent agricultural irrigation systems can effectively solve the problem of uncertain soil moisture changes, improving irrigation efficiency and water resource utilization. The structural block diagram of the fuzzy control system is shown below. Figure 3 As shown, it includes inputs, outputs, an actuator, a controlled object, and a measuring device. In this application, the inputs are soil moisture error and air temperature error, the output is irrigation duration, the actuator is a water pump, the controlled object is soil moisture, and the measuring devices are a soil moisture sensor and an air temperature sensor.

[0061] In fuzzy control, a fuzzy set refers to the entire set of objects possessing attributes described by a certain fuzzy concept. In general, it's a set where elements belong to it to some degree; this type of set is called a "fuzzy set," and the degree to which each element belongs to this set is called its "membership degree." When the membership degree approaches 0, it means the element's association with the fuzzy set is weak; conversely, when the membership degree approaches 1, it indicates a very close relationship between the element and the set. The universe of discourse is the object being discussed, such as plant age or growth temperature, and the elements of the universe of discourse are the values ​​of the object being discussed. When representing a fuzzy set, we first define the universe of discourse as X = {x1, x2, ..., x...}. n The membership degree of each element in} to the fuzzy set A is μ. A (x1),μ A (x2),μ A (x n Using Zadeh notation, the fuzzy set A can be represented as:

[0062]

[0063] in, This indicates the correspondence between the membership degrees of elements in the universe of discourse to the fuzzy set, with "+" indicating summarization.

[0064] In fuzzy control theory, membership functions describe the relationship between control input and output variables. They transform the actual input data into a numerical value between 0 and 1, representing the degree to which the input value belongs to a certain fuzzy set. Its definition is: for a universe of discourse X, if there exists a real-valued function μ... A , so that μ A If X→[0,1], then A is a fuzzy set on the universe of discourse X, and the characteristic function μ A (x) is called the membership function, which describes the position of element X in μ. A (x) belongs to set A to a certain degree. The membership function μ A (x) Different values ​​correspond to different degrees of membership of X to set A, when 0 < μ A When (x) < 1, it means that X is in μ A (x) belongs to A to a certain extent; when μ A When (x) = 0, it means that X does not belong to A at all; when μ A When X = 1, it means that X completely belongs to A;

[0065] (1) Mathematical expression of the membership function of a triangle:

[0066]

[0067] Where x is the universe of discourse of the membership function of the triangle; a and c are the two endpoints of the base of the triangle; and b is the vertex of the triangle.

[0068] (2) Mathematical expression of Gaussian membership function:

[0069]

[0070] Where x is the universe of discourse of the Gaussian membership function; c is the center point of the curve; and σ is the width of the curve.

[0071] (3) Mathematical expression of trapezoidal membership function:

[0072]

[0073] Where x is the universe of discourse of the trapezoidal membership function; a and d are the two lower endpoints of the trapezoid; and b and c are the two upper endpoints of the trapezoid.

[0074] (4) Bell-shaped membership function:

[0075] Similar to a Gaussian membership function, but its tail can be wider or narrower. It has two main parameters: mean and standard deviation.

[0076]

[0077] Where x is the universe of discourse of the bell-shaped membership function; a is the initial inflection point of the curve; b is the highest inflection point of the curve; and c is the center point of the curve.

[0078] This application employs a Mamdani-type fuzzy inference system with dual inputs and a single output for fuzzy decision-making. Its key characteristic lies in the form of its fuzzy rules, defined as follows: Given input variable x and output variable y, there exists a rule base consisting of several if-then statements. The output of each if statement is a set of multiple linguistic variables corresponding to different states of the input variable. The design of the fuzzy controller is crucial to the entire control system. One-dimensional fuzzy controllers have only one input, resulting in low accuracy and performance. Three-dimensional fuzzy controllers have excessively high accuracy and overly complex control rules. Two-dimensional controllers fully meet the system requirements; therefore, this system selects a two-dimensional fuzzy controller. In intelligent irrigation systems, the control signals obtained from fuzzy inference can be defuzzified to obtain the actual control quantity. This paper uses the centroid method for defuzzification, calculating the center of the area enclosed by the fuzzy set membership function curve and the horizontal coordinate, and selecting the horizontal coordinate x corresponding to that point. cen As a representative value of this fuzzy set, it is shown in equation (2-6):

[0079]

[0080] Based on the determined fuzzy sets, fuzzy rules, and membership functions, a fuzzy control rule response is established after fuzzy inference and defuzzification.

[0081] Based on the design of the fuzzy controller, the workflow of the fuzzy controller is summarized as follows: After system initialization, the sensors begin collecting data, transmitting the collected soil moisture and air temperature to the microcomputer via a 485 to USB converter. The fuzzy controller then determines whether irrigation is needed. If irrigation is required, it outputs the specific irrigation duration and triggers a relay to control the water pump for irrigation. A new round of data collection then begins, repeating the above steps. If irrigation is not needed, a new round of data collection begins directly, and the above steps are repeated. The fuzzy controller workflow is as follows: Figure 5 As shown.

[0082] In summary, this invention enables intelligent, tiered irrigation of planting areas in different regions, which reduces the energy consumption of pumping stations, minimizes water waste, and effectively increases crop yield.

[0083] Furthermore, in order to provide a concise description of exemplary embodiments, not all features of actual embodiments (i.e., those features that are not relevant to the currently considered best mode for carrying out the invention, or those features that are not relevant to implementing the invention) may be omitted.

[0084] It should be understood that numerous specific implementation decisions can be made during the development of any practical implementation, such as in any engineering or design project. Such development efforts may be complex and time-consuming, but for those skilled in the art who benefit from this disclosure, the development effort will be a routine work of design, manufacturing, and production without requiring much experimentation.

[0085] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A smart, tiered water-saving irrigation control system based on the Internet of Things, characterized in that, include: Temperature and humidity sensors, pressure sensors, and wind speed sensors are integrated into the smart sensor node to acquire multimodal data and obtain multimodal data. An automatic closed-loop dual-mode control algorithm combining switch control and fuzzy control is adopted to fuse the multimodal data and send it to the edge computing gateway; Data from the edge computing gateway is sent to the cloud platform via the LoRa network; the gateway and the cloud platform communicate via NB-IoT and 5G dual-mode communication, and the local decision-making mode is automatically activated when communication is interrupted. By combining adaptive learning rate and regularization techniques, multi-level actuators are controlled in stages through the core control module. Dynamic tiered logic: Irrigation priorities are determined based on plant growth cycles, real-time environmental data, and predictive models; The actuator has a built-in pressure sensor and flow meter. When a pipeline leak is detected, it automatically closes the valves in the corresponding area and marks the fault point on a GIS map. An LSTM neural network is introduced to predict soil moisture change trends, and a random forest algorithm is combined to dynamically update the graded thresholds. This includes a human-computer interaction page, a hierarchical permission management interface, and support for data collaboration among farmers, agricultural experts, and water authorities. A multi-level irrigation priority system is established by weighting and integrating factors such as crop growth period, weather forecast, and soil moisture.

2. The IoT-based intelligent tiered water-saving irrigation control system according to claim 1, characterized in that, It includes a drive water pump, which is connected to a control module, and the control module is connected to a WiFi module.

3. The IoT-based intelligent tiered water-saving irrigation control system according to claim 1, characterized in that, It also includes an irrigation module, which includes a relay and a water pump, and automatically controls the water pump to irrigate according to the irrigation duration output by the fuzzy controller.

4. The IoT-based intelligent tiered water-saving irrigation control system according to claim 1, characterized in that, The dual-mode control algorithm includes protection control, switching control, and fuzzy control.

Images