A gate control method and system based on neural network

Abstract

The present invention provides a gate control method and system based on a neural network, and relates to the technical field of gate operation control in agricultural irrigation areas. The present invention can learn and predict optimal gate opening instructions and gate opening time instructions according to historical data by using a long short-term memory neural network model, thereby realizing dynamic adjustment of the irrigation process and improving irrigation efficiency. In addition, on the basis of utilizing the neural network, genetic algorithm optimization is combined to dynamically adjust the gate opening and gate opening time to realize precise irrigation, thereby significantly improving water resource utilization efficiency and reducing unnecessary water consumption. At the same time, a fuzzy logic processing mechanism is introduced, so that the system can better cope with uncertainty and fuzzy factors existing in the irrigation environment. At the same time, the application of the genetic algorithm ensures that the optimal solution can be found even in a complex and changeable environment, thereby enhancing the stability and reliability of the system.

Description

Technical Field

[0001] The invention relates to the technical field of agricultural irrigation area gate operation control, and in particular to a neural network-based gate control method and system thereof. Background Art

[0002] Traditional irrigation systems usually use fixed irrigation schedules or rely on manual judgment to operate. This method is difficult to accurately respond to actual soil moisture, climate conditions and changes in crop demand, resulting in inefficient use of water resources. Sometimes, it also causes water waste due to over-irrigation or affects crop growth due to insufficient irrigation. With the development of modern agricultural technology, people pay more and more attention to improving the efficiency and accuracy of irrigation systems through intelligent means.

[0003] In recent years, artificial intelligence technology, especially machine learning algorithms such as neural networks, has been applied to various prediction and control systems. Long short-term memory networks, as a special recursive neural network, can effectively process sequence data, so they have application potential in irrigation environments where time series characteristics need to be considered. In addition, genetic algorithms, as a global search algorithm, can be used to optimize complex multi-objective problems, while fuzzy logic provides an effective method for dealing with uncertainty and fuzzy information.

[0004] Therefore, it is necessary to provide a gate control method and system based on neural network to solve the above technical problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a gate control method and system based on neural network, which can achieve the beneficial effects of significantly improving the efficiency of water resource utilization and better coping with the influence of uncertain factors brought about by environmental changes.

[0006] The present invention provides a gate control method based on a neural network, the gate control method comprising the following steps:

[0007] S1: Obtain comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes current soil moisture data, current climate condition data, current crop information data, and current gate system status data;

[0008] S2: Inputting the comprehensive irrigation environment data of each irrigation area into the pre-trained gate control neural network model to obtain the gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction;

[0009] S3: Using a genetic algorithm to adjust the obtained gate preliminary control instruction combination to obtain a gate adjustment control instruction combination;

[0010] S4: fuzzy processing is performed on the gate adjustment control instruction combination to obtain the gate preliminary fuzzy control instruction vector;

[0011] S5: Processing the preliminary fuzzy control instruction vector of the gate through a pre-selected fuzzy inference engine to obtain a fuzzy output, defuzzifying the fuzzy output through a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

[0012] Preferably, in step S2, the pre-trained gate control neural network model adopts a long short-term memory network model.

[0013] Preferably, step S3 comprises the following steps:

[0014] S301: Initializing relevant parameters of the genetic algorithm, and randomly generating a population based on preset constraints, wherein the population includes a plurality of individuals, each of which represents a set of preliminary gate control instruction combinations;

[0015] S302: obtaining the irrigation uniformity of each irrigation area, the soil moisture of each irrigation area, and the crop transpiration rate of each irrigation area as evaluation indicators, assigning corresponding weights to each evaluation indicator, and constructing a fitness function based on the evaluation indicators and their corresponding weights;

[0016] S303: using a computer simulation method to predict the irrigation uniformity, soil moisture and crop transpiration rate of each individual in the population, and calculating the fitness value of each individual in the population according to the fitness function;

[0017] S304: Based on the pre-selected selection strategy, select individuals with high fitness values ​​from the population as parents, perform crossover and mutation operations on the parents to generate new individuals, and replace old individuals in the population with the new individuals to form a new generation population;

[0018] S305: Repeat S303 and S304 until the preset maximum number of iterations is reached, select the individual with the highest fitness value from the latest generation population, and generate a gate adjustment control instruction combination.

[0019] Preferably, in step S302, the calculation formula of the fitness function is:

[0020]

[0021] in, It is a set of gate preliminary control instruction combination, For irrigation uniformity, is soil moisture, is the crop transpiration rate, , , are the weight coefficients of irrigation uniformity, soil moisture and crop transpiration rate, respectively.

[0022] Preferably, in step S304, probability parameters of crossover and mutation operations are automatically adjusted based on the fitness distribution during population evolution.

[0023] Preferably, step S4 comprises the following steps:

[0024] S401: Determine the gate opening instruction and the gate opening time instruction in the gate adjustment control instruction combination as input variables, and define a set of corresponding fuzzy sets for each input variable, and at the same time, determine a corresponding membership function for each fuzzy set;

[0025] S402: For each input variable, using the corresponding membership function, calculate its membership on the corresponding fuzzy set, and combine the membership of all input variables on their corresponding fuzzy sets into a membership matrix;

[0026] S403: Based on the membership matrix and the preset basic rule base, a preliminary fuzzy control instruction vector of the gate is obtained.

[0027] Preferably, in step S401, the fuzzy set also includes an upper extreme value set and a lower extreme value set. When the input variable exceeds a preset upper threshold, it is automatically included in the upper extreme value set; when the input variable is lower than a preset lower threshold, it is automatically included in the lower extreme value set.

[0028] Preferably, step S5 comprises the following steps:

[0029] S501: Initialize the fuzzy inference engine. After the initialization is completed, input the gate preliminary fuzzy control instruction vector into the fuzzy inference engine;

[0030] S502: Execute complex fuzzy reasoning based on a preset complex rule base and using a fuzzy reasoning engine to obtain corresponding fuzzy output;

[0031] S503: Using a preset defuzzification method, the fuzzy output obtained by complex fuzzy reasoning is converted into a final control instruction combination of the gate.

[0032] Preferably, in step S502, the complex fuzzy reasoning adopts a multi-layer reasoning structure.

[0033] The present invention also provides a gate control system based on a neural network, which is applied to a gate control method based on a neural network. The gate control system includes:

[0034] The data acquisition module is used to obtain the comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes current soil moisture data, current climate condition data, current crop information data, and current gate system status data;

[0035] A neural network prediction module is used to input the comprehensive irrigation environment data of each irrigation area into a pre-trained gate control neural network model to obtain a gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction;

[0036] A genetic algorithm optimization module is used to adjust the obtained gate preliminary control instruction combination by using a genetic algorithm to obtain a gate adjustment control instruction combination;

[0037] A fuzzy processing module is used to perform fuzzy processing on the gate adjustment control instruction combination to obtain a preliminary fuzzy control instruction vector of the gate;

[0038] The fuzzy reasoning module is used to process the preliminary fuzzy control instruction vector of the gate through a pre-selected fuzzy reasoning engine to obtain a fuzzy output, and to defuzzify the fuzzy output through a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

[0039] Compared with the related art, the gate control method and system based on neural network provided by the present invention have the following beneficial effects:

[0040] The present invention uses a long short-term memory neural network model to learn and predict the optimal gate opening instructions and opening time instructions based on historical data, thereby realizing dynamic adjustment of the irrigation process and improving irrigation efficiency. In addition, on the basis of utilizing neural networks, genetic algorithm optimization is also combined to dynamically adjust the gate opening and opening time to achieve precise irrigation, thereby significantly improving water resource utilization efficiency and reducing unnecessary water consumption. At the same time, a fuzzy logic processing mechanism is introduced to enable the system to better cope with the uncertainty and fuzzy factors existing in the irrigation environment. At the same time, the application of genetic algorithms ensures that the optimal solution can be found even in complex and changeable environments, thereby enhancing the stability and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic flow chart of a gate control method based on a neural network of the present invention;

[0042] Figure 2 The figure is a schematic diagram of the module structure of a gate control system based on a neural network of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only the parts related to the present invention, rather than all structures, are shown in the accompanying drawings. In addition, the embodiments of the present invention and the features in the embodiments may be combined with each other without conflict.

[0044] It should also be noted that, for ease of description, only the parts related to the present invention, but not all of the contents, are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0045] Embodiment 1

[0046] A gate control method based on neural network, in the specific implementation process, such as Figure 1 As shown, it shows a flow chart of a gate control method based on a neural network, including:

[0047] Step S1: Obtain comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes current soil moisture data, current climate condition data, current crop information data, and current gate system status data.

[0048] During the specific implementation process, for example, a number of sensors are evenly distributed in each irrigation area to obtain current soil moisture data, including but not limited to soil water content and soil conductivity at different depths; one or more small meteorological stations are set up in each irrigation area to obtain current climate condition data, including but not limited to temperature, humidity, and rainfall; current crop information data are obtained through satellite remote sensing and field observations by ground agricultural technicians, including but not limited to crop types, crop growth periods, and other growth indicators; each gate is equipped with a status monitoring device to obtain current gate system status data, including but not limited to gate opening and closing status, gate opening size, and gate operation time; these data are used as the original basis for the subsequent generation of preliminary gate control instruction combinations.

[0049] Step S2: Input the comprehensive irrigation environment data of each irrigation area into the pre-trained gate control neural network model to obtain the gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction.

[0050] During the specific implementation process, the acquired comprehensive irrigation environment data is normalized, and the normalized current comprehensive irrigation environment data is input into the trained neural network model in real time to generate a preliminary gate control instruction combination corresponding to each irrigation area. The preliminary gate control instruction combination includes a gate opening instruction, which indicates the recommended gate opening value, which determines the speed and flow of water entering the irrigation area. At the same time, a gate opening duration instruction indicates the recommended gate opening duration to ensure that an appropriate amount of water is delivered to the irrigation area.

[0051] Step S3: Using a genetic algorithm to adjust the obtained gate preliminary control instruction combination to obtain a gate adjustment control instruction combination.

[0052] Specifically, step S3 includes the following steps:

[0053] Step S301: Initialize the relevant parameters of the genetic algorithm, and randomly generate a population based on preset constraints, wherein the population includes a plurality of individuals, each of which represents a set of preliminary gate control instruction combinations;

[0054] Step S302: obtaining the irrigation uniformity of each irrigation area, the soil moisture of each irrigation area, and the crop transpiration rate of each irrigation area as evaluation indicators, assigning corresponding weights to each evaluation indicator, and constructing a fitness function based on the evaluation indicators and their corresponding weights;

[0055] Step S303: using a computer simulation method to predict the irrigation uniformity, soil moisture and crop transpiration rate of each individual in the population, and calculating the fitness value of each individual in the population according to the fitness function;

[0056] Step S304: Based on the pre-selected selection strategy, individuals with high fitness values ​​are selected from the population as parents, crossover and mutation operations are performed on the parents to generate new individuals, and the new individuals are used to replace the old individuals in the population to form a new generation of population;

[0057] Step S305: Repeat S303 and S304 until the preset maximum number of iterations is reached, select the individual with the highest fitness value from the latest generation population, and generate a gate adjustment control instruction combination.

[0058] Specifically, in step S302, the calculation formula of the fitness function is:

[0059]

[0060] in, It is a set of gate preliminary control instruction combination, For irrigation uniformity, is soil moisture, is the crop transpiration rate, , , are the weight coefficients of irrigation uniformity, soil moisture and crop transpiration rate, respectively.

[0061] Specifically, in step S304, the probability parameters of the crossover and mutation operations are automatically adjusted based on the fitness distribution during the population evolution process.

[0062] During the specific implementation process, first initialize the relevant parameters of the genetic algorithm. The relevant parameters include the population size, which determines the number of individuals in each generation. Usually, a suitable size is selected according to the complexity of the problem and computing resources; the crossover probability, which defines the probability of two parent individuals generating a new individual. The default value is generally set to 0.7 to 0.9; the mutation probability, which sets the probability of random changes in individual genes. It is usually small. For example, it is set to 0.01; the maximum number of iterations, which specifies the maximum number of iterations at which the genetic algorithm stops, ensures that the calculation process does not continue indefinitely, wherein the population contains multiple individuals, each of which represents a set of preliminary gate control instruction combinations. According to the requirements of the actual irrigation system, exemplary constraints such as minimum / maximum gate opening and minimum / maximum gate opening time are set to ensure that the generated solution meets the physical and operational feasibility.

[0063] Then, we collected evaluation index data, including irrigation uniformity, which measures the consistency of irrigation water distribution in the field, soil moisture, which reflects the moisture content in the soil, and crop transpiration rate, which assesses the degree of crop demand for water. Different growth stages have different ranges. According to the opinions of agricultural experts or historical data analysis, we assigned appropriate weights to each evaluation index and constructed a fitness function that comprehensively considers all evaluation indicators. The calculation formula of the fitness function is:

[0064]

[0065] The fitness calculation formula is obtained by weighted summation, where: It is a set of gate preliminary control instruction combination, For irrigation uniformity, is soil moisture, is the crop transpiration rate, , , are the weight coefficients of irrigation uniformity, soil moisture and crop transpiration rate, respectively.

[0066] Secondly, a computer is used to simulate the operation of the irrigation system. The preliminary control instructions of each individual gate are input and the corresponding irrigation effect is output. The simulation results are substituted into the pre-built fitness function to calculate the fitness value of each individual. The higher the fitness value, the better the effect of the combination of preliminary gate control instructions.

[0067] Again, illustratively, the roulette wheel selection method is selected, in which the probability of an individual being selected is proportional to its fitness. The higher the fitness, the greater the probability of the individual being selected. The uniform crossover method and the bit flipping mutation method are selected for crossover and mutation operations. The new individuals generated by crossover and mutation replace the individuals with lower fitness in the current population to form a new generation of population. In the early stages of evolution, a higher crossover probability is used to promote population diversity and accelerate the exploration of the solution space. At the same time, a low mutation probability is maintained to maintain existing good genes. As the evolutionary process develops, the crossover probability is gradually reduced, unnecessary disturbances are reduced, local search capabilities are increased, and the mutation probability is appropriately increased to avoid premature convergence to the local optimal solution.

[0068] Finally, steps S303 and S304 are executed repeatedly according to the set maximum number of iterations to continuously optimize the fitness values ​​of individuals in the population. When the preset maximum number of iterations is reached, the iteration is stopped, and the individual with the highest fitness value is selected from the final generation population, and its corresponding gate control instruction combination is used as the gate adjustment control instruction combination.

[0069] Step S4: fuzzy process the gate adjustment control instruction combination to obtain a preliminary fuzzy control instruction vector of the gate.

[0070] Specifically, step S4 includes the following steps:

[0071] Step S401: determining the gate opening instruction and the gate opening time instruction in the gate adjustment control instruction combination as input variables, and defining a set of corresponding fuzzy sets for each input variable, and at the same time, determining a corresponding membership function for each fuzzy set;

[0072] Step S402: for each input variable, using the corresponding membership function, calculate its membership on the corresponding fuzzy set, and combine the membership of all input variables on their corresponding fuzzy sets into a membership matrix;

[0073] Step S403: Based on the membership matrix and the preset basic rule base, a preliminary fuzzy control instruction vector of the gate is obtained.

[0074] Specifically, in step S401, the fuzzy set also includes an upper extreme value set and a lower extreme value set. When the input variable exceeds the preset upper threshold, it is automatically included in the upper extreme value set; when the input variable is lower than the preset lower threshold, it is automatically included in the lower extreme value set.

[0075] During the specific implementation process, first, the input variables are determined and fuzzy sets and membership functions are defined. The gate opening instruction and the gate opening time instruction in the gate adjustment control instruction combination are determined as input variables. The gate opening instruction indicates the degree of gate opening, and the gate opening time instruction indicates the length of time the gate remains in the open state. For each input variable, a set of fuzzy sets describing its possible value range is defined. Exemplarily, for the gate opening instruction, the fuzzy sets can be defined as extremely small, very small, small, medium, large, very large, and extremely large, where extremely small and extremely large are extreme value sets, and extremely large is an upper limit extreme value set. When the input variable exceeds the preset upper limit threshold, it is automatically classified as extremely large, and extremely small is a lower limit extreme value set. When the input variable is lower than the preset lower limit threshold, it is automatically classified as extremely small. Similarly, for the gate opening time instruction, the fuzzy sets are defined as extremely short, very short, short, moderate, long, very long, and extremely long. An appropriate membership function is selected for each fuzzy set. Exemplarily, the membership function types include but are not limited to triangular, trapezoidal, and Gaussian types. These functions define the degree to which the input variable belongs to a certain fuzzy set.

[0076] Then, for each input variable, the membership function is applied to all the corresponding fuzzy sets to calculate its membership to each fuzzy set. The membership reflects the strength of the relationship between the input variable and each fuzzy set. The membership of all input variables on their corresponding fuzzy sets is combined into a matrix form, called the membership matrix. Each row of the matrix represents an input variable, each column represents a fuzzy set, and the matrix elements are membership values.

[0077] Finally, according to the professional knowledge and actual needs in the field of agricultural irrigation, the basic rule base is set in advance, the fuzzy control instruction vector is preliminarily inferred, and the preliminary fuzzy control instruction vector of the gate is obtained. The preliminary fuzzy control instruction vector of the fuzzy gate here is still in fuzzy form and needs further processing before it can be applied in practice.

[0078] Step S5: Processing the preliminary fuzzy control instruction vector of the gate by a pre-selected fuzzy inference engine to obtain a fuzzy output, defuzzifying the fuzzy output by a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

[0079] Specifically, step S5 includes the following steps:

[0080] Step S501: Initialize the fuzzy inference engine. After the initialization is completed, input the gate preliminary fuzzy control instruction vector into the fuzzy inference engine;

[0081] Step S502: Execute complex fuzzy reasoning based on a preset complex rule base and using a fuzzy reasoning engine to obtain corresponding fuzzy output;

[0082] Step S503: using a preset defuzzification method to convert the fuzzy output obtained by complex fuzzy reasoning into a final control instruction combination of the gate.

[0083] Specifically, in step S502, complex fuzzy reasoning adopts a multi-layer reasoning structure.

[0084] During the specific implementation process, the fuzzy inference engine must be initialized first. This step requires the execution of complex fuzzy inference. For example, the Mamdani fuzzy inference engine is selected and the complex rule base is loaded. The complex rule base covers the control logic under various irrigation scenarios and is verified.

[0085] Then, the fuzzy output is obtained through multi-level fuzzy reasoning. Each layer independently configures different reasoning parameters and rule bases. Each level focuses on a specific reasoning task and gradually refines the control strategy. For example, through three-level fuzzy reasoning, a rough control reasoning is performed to obtain a preliminary fuzzy output, and then the preliminary fuzzy output is used as input for medium-refined fuzzy reasoning to obtain a medium fuzzy output. Then, the medium fuzzy output is used as input for fine fuzzy reasoning to obtain the final fuzzy output.

[0086] Finally, illustratively, a defuzzification method such as the centroid method is selected to defuzzify the fuzzy output to obtain a defuzzified value, and the defuzzified value is formatted into a final control instruction combination of the gate.

[0087] The working principle of a gate control method based on neural network provided by the present invention is as follows:

[0088] The present invention combines genetic algorithms with multi-level fuzzy logic reasoning, and realizes precise control of the irrigation system through intelligent multi-stage optimization processing based on neural networks. First, the genetic algorithm is used to optimize the initial control instruction combination of the gate to ensure that the water resource utilization efficiency is maximized while meeting the irrigation needs. Then, the preliminary instructions are progressively refined through a multi-layer fuzzy reasoning engine, and each layer is independently configured with different reasoning parameters and rule bases, gradually considering the influence of various factors from global to local to real-time, and generating more refined fuzzy control instructions. Finally, the defuzzification method is used to convert the fuzzy output into a specific final control instruction combination of the gate, which is directly applied to the gate control in the actual irrigation system.

[0089] Embodiment 2

[0090] The present invention also provides a neural network-based gate control system applied to a neural network-based gate control method. In a specific implementation process, Figure 2 As shown, it shows a schematic diagram of the module structure of a gate control system based on a neural network provided by the present invention, including:

[0091] The data acquisition module 100 is used to obtain the comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes current soil moisture data, current climate condition data, current crop information data, and current gate system status data;

[0092] The neural network prediction module 200 is used to input the comprehensive irrigation environment data of each irrigation area into the pre-trained gate control neural network model to obtain the gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction;

[0093] A genetic algorithm optimization module 300 is used to adjust the obtained gate preliminary control instruction combination using a genetic algorithm to obtain a gate adjustment control instruction combination;

[0094] The fuzzy processing module 400 is used to perform fuzzy processing on the gate adjustment control instruction combination to obtain the gate preliminary fuzzy control instruction vector;

[0095] The fuzzy reasoning module 500 is used to process the preliminary fuzzy control instruction vector of the gate through a pre-selected fuzzy reasoning engine to obtain a fuzzy output, and to defuzzify the fuzzy output through a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

[0096] The working principle of a gate control system based on neural network provided by the present invention is as follows:

[0097] The data acquisition module 100 acquires the comprehensive irrigation environment data of each irrigation area in real time, covering soil moisture, climate conditions, crop information and gate status; the neural network prediction module 200 generates a preliminary gate control command combination based on these data using a pre-trained neural network model; the genetic algorithm optimization module 300 further optimizes these commands to ensure that they remain effective in complex environments; the fuzzy processing module 400 converts the optimized commands into fuzzy control command vectors to cope with uncertainty; the fuzzy reasoning module 500 uses multi-level reasoning and defuzzification processing to ultimately generate specific gate control commands that are directly applied to the irrigation system.

[0098] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0099] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0100] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A gate control method based on neural network, characterized in that: The gate control method comprises the following steps: S1: Obtain comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes current soil moisture data, current climate condition data, current crop information data, and current gate system status data; S2: Inputting the comprehensive irrigation environment data of each irrigation area into the pre-trained gate control neural network model to obtain the gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction; S3: Using a genetic algorithm to adjust the obtained gate preliminary control instruction combination to obtain a gate adjustment control instruction combination; S4: fuzzy processing is performed on the gate adjustment control instruction combination to obtain the gate preliminary fuzzy control instruction vector; S5: Processing the preliminary fuzzy control instruction vector of the gate through a pre-selected fuzzy inference engine to obtain a fuzzy output, defuzzifying the fuzzy output through a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

2. A gate control method based on neural network according to claim 1, characterized in that: In step S2, the pre-trained gate control neural network model adopts a long short-term memory network model.

3. A gate control method based on neural network according to claim 2, characterized in that: Step S3 includes the following steps: S301: Initializing relevant parameters of the genetic algorithm, and randomly generating a population based on preset constraints, wherein the population includes a plurality of individuals, each of which represents a set of preliminary gate control instruction combinations; S302: obtaining the irrigation uniformity of each irrigation area, the soil moisture of each irrigation area, and the crop transpiration rate of each irrigation area as evaluation indicators, assigning corresponding weights to each evaluation indicator, and constructing a fitness function based on the evaluation indicators and their corresponding weights; S303: using a computer simulation method to predict the irrigation uniformity, soil moisture and crop transpiration rate of each individual in the population, and calculating the fitness value of each individual in the population according to the fitness function; S304: Based on the pre-selected selection strategy, select individuals with high fitness values ​​from the population as parents, perform crossover and mutation operations on the parents to generate new individuals, and replace old individuals in the population with the new individuals to form a new generation population; S305: Repeat S303 and S304 until the preset maximum number of iterations is reached, select the individual with the highest fitness value from the latest generation population, and generate a gate adjustment control instruction combination.

4. A gate control method based on neural network according to claim 3, characterized in that: In step S302, the calculation formula of the fitness function is: in, It is a set of gate preliminary control instruction combination, For irrigation uniformity, is soil moisture, is the crop transpiration rate, , , are the weight coefficients of irrigation uniformity, soil moisture and crop transpiration rate, respectively.

5. A gate control method based on neural network according to claim 4, characterized in that: In step S304, the probability parameters of the crossover and mutation operations are automatically adjusted based on the fitness distribution during the population evolution process.

6. A gate control method based on neural network according to claim 5, characterized in that: Step S4 includes the following steps: S401: Determine the gate opening instruction and the gate opening time instruction in the gate adjustment control instruction combination as input variables, and define a set of corresponding fuzzy sets for each input variable, and at the same time, determine a corresponding membership function for each fuzzy set; S402: For each input variable, using the corresponding membership function, calculate its membership on the corresponding fuzzy set, and combine the membership of all input variables on their corresponding fuzzy sets into a membership matrix; S403: Based on the membership matrix and the preset basic rule base, a preliminary fuzzy control instruction vector of the gate is obtained.

7. A gate control method based on neural network according to claim 6, characterized in that: In step S401, the fuzzy set also includes an upper extreme value set and a lower extreme value set. When the input variable exceeds the preset upper threshold, it is automatically included in the upper extreme value set; when the input variable is lower than the preset lower threshold, it is automatically included in the lower extreme value set.

8. A gate control method based on neural network according to claim 7, characterized in that: Step S5 includes the following steps: S501: Initialize the fuzzy inference engine. After the initialization is completed, input the gate preliminary fuzzy control instruction vector into the fuzzy inference engine; S502: Execute complex fuzzy reasoning based on a preset complex rule base and using a fuzzy reasoning engine to obtain corresponding fuzzy output; S503: Using a preset defuzzification method, the fuzzy output obtained by complex fuzzy reasoning is converted into a final control instruction combination of the gate.

9. A gate control method based on neural network according to claim 8, characterized in that: In step S502, the complex fuzzy reasoning adopts a multi-layer reasoning structure.

10. A gate control system based on neural network, characterized in that: A gate control method based on a neural network as claimed in any one of claims 1 to 9, wherein the gate control system comprises: The data acquisition module is used to obtain the comprehensive irrigation environment data of each irrigation area, wherein the comprehensive irrigation environment data includes the current soil moisture data, the current climate condition data, the current crop information data, and the current gate system status data; A neural network prediction module is used to input the comprehensive irrigation environment data of each irrigation area into a pre-trained gate control neural network model to obtain a gate preliminary control instruction combination corresponding to each irrigation area, wherein the gate preliminary control instruction combination includes a gate opening instruction and a gate opening time instruction; A genetic algorithm optimization module is used to adjust the obtained gate preliminary control instruction combination by using a genetic algorithm to obtain a gate adjustment control instruction combination; A fuzzy processing module is used to perform fuzzy processing on the gate adjustment control instruction combination to obtain a preliminary fuzzy control instruction vector of the gate; The fuzzy reasoning module is used to process the preliminary fuzzy control instruction vector of the gate through a pre-selected fuzzy reasoning engine to obtain a fuzzy output, and to defuzzify the fuzzy output through a pre-selected defuzzification method to obtain a final control instruction combination of the gate.

Images