Agricultural informationization internet of things terminal control system

Abstract

The application discloses an agricultural informatization Internet of Things terminal control system and relates to the technical field of embedded control of agricultural informatization Internet of Things, which comprises a cloud management platform, an edge Internet of Things terminal, on-site sensors and an execution mechanism. The edge Internet of Things terminal stores the control strategy issued by the cloud to form a shadow state matrix when the network connection is normal, freezes the matrix as an optimization reference value when the network connection is interrupted, and adjusts the proportional-integral-derivative parameters adaptively in combination with the change rate of on-site environmental parameters. According to the network interruption duration and the cloud-edge environmental deviation, a fusion weight is constructed to obtain transition control parameters by weighted fusion of the local optimization state parameters and the cloud control strategy parameters and drive the execution mechanism. The application avoids the problem of insufficient control accuracy of the fixed threshold during network interruption, eliminates the step disturbance caused by the hard switching of the control parameters in the networking moment, and improves the water and fertilizer application accuracy and the system operation stability.

Description

Technical Field

[0001] This invention relates to the field of embedded control technology for agricultural information IoT, specifically to an agricultural information IoT terminal control system. Background Technology

[0002] With the development of agricultural informatization, the cloud-edge collaborative agricultural IoT control architecture has been widely used in automated irrigation, fertilization, and environmental control systems for greenhouses and large-scale farmland. This architecture typically consists of three parts: a cloud management platform, edge IoT terminals, and field sensors and actuators. The edge IoT terminals are usually equipped with embedded microcontrollers and run lightweight real-time operating systems. The cloud management platform performs global optimization calculations based on large-scale meteorological data, historical planting data, and information gathered from multiple terminals, and then distributes the control strategy to each edge IoT terminal for execution. Proportional-integral-derivative (PID) control, as a mature closed-loop feedback control algorithm, is widely used in continuous quantity control scenarios such as irrigation valve opening and temperature and humidity regulation. In addition, the disturbance-free switching technology has a relatively mature mathematical foundation in traditional distributed control systems and is mainly used to solve the problem of abrupt changes in control quantities during control mode switching. Lyapunov stability theory is a mathematical tool for analyzing the convergence and stability of dynamic systems and is often used for stability proof and transient process design of control systems. The above contents are all well-known technologies in the art and provide a theoretical and hardware foundation for the implementation of this invention.

[0003] The typical working scheme of agricultural IoT terminal control systems currently deployed in the industry is as follows: When the network connection is normal, the terminal periodically receives control strategies from the cloud management platform, including target setpoints and control parameters, and drives the actuators accordingly. At the same time, it uploads the data collected by the field sensors to the cloud management platform for global optimization calculation. When the terminal detects a network connection interruption, the control system automatically switches to local degradation mode. In this mode, the terminal performs on / off control based on the factory-preset fixed upper and lower thresholds, and no longer performs parameter optimization or adaptive adjustment. When the terminal detects that the network connection has been restored, the system immediately receives and executes the control strategy recalculated by the cloud management platform based on global data during the network interruption. The strategy parameters are directly written into the terminal's execution control module, replacing the execution state in the previous local degradation mode, and the control is immediately and completely transferred to the cloud management platform.

[0004] The aforementioned existing technical solutions have the following drawbacks: First, because the terminal only executes fixed upper and lower limit threshold logic during network connection interruption, it does not consider the real-time changing trends of environmental parameters such as on-site weather and soil. Therefore, when environmental parameters change drastically during network interruption, the terminal cannot adaptively adjust the control strategy, resulting in insufficient control accuracy, delayed action of the actuator, or malfunction, which in turn leads to waste of water and fertilizer resources or crop damage. Second, existing non-disruptive handover technologies generally only perform parameter gradual change processing based on the handover time, without considering the impact of the objective deviation between the local actual environmental state and the target state in the cloud on the handover rate. This application breaks through the inertia of this single-time-dimensional handover approach and innovatively considers the duration of network interruption. A fusion weighted attenuation model is constructed by coupling two variables, cloud edge and environmental deviation, to achieve adaptive adjustment of the switching rate according to environmental deviation. However, it does not consider the objective deviation between the actual local execution state and the expected state in the cloud. Therefore, the control parameters will undergo instantaneous hard switching, causing the actuators such as irrigation solenoid valves and film rolling machine motors to receive step-change control commands, which will lead to water hammer effect in the pipeline network and mechanical impact damage to the motors, and cause a sharp decline in the accuracy of water and fertilizer application. Therefore, how to enable agricultural IoT terminals to achieve adaptive fine control during network connection interruption and achieve smooth fusion switching between local control state and cloud strategy when network connection is restored is a technical problem that urgently needs to be solved in this field. To this end, we propose an agricultural information IoT terminal control system. Summary of the Invention

[0005] The purpose of this invention is to provide an agricultural information Internet of Things (IoT) terminal control system.

[0006] To address the shortcomings of the existing technologies, the technical problem this invention aims to solve is: how to enable agricultural information IoT terminals to overcome the limitations of fixed threshold logic during network connection interruptions, achieve adaptive and fine control based on real-time changes in on-site environmental parameters, and how to avoid direct hard overwriting of local execution states by cloud-based policy parameters upon network connection restoration, thereby achieving a smooth transition and fusion between local control states and cloud-based global policies. This would eliminate the step disturbances generated during control switching, avoid mechanical oscillations of actuators and water hammer effects in pipelines, and improve the accuracy of water and fertilizer application and the stability of system operation.

[0007] To address the aforementioned technical problems, this invention provides an agricultural information IoT terminal control system, comprising a cloud management platform, an edge IoT terminal, and field sensors and actuators connected to the edge IoT terminal. The edge IoT terminal includes a network status monitoring module, a shadow state storage module, an edge autonomous optimization module, a state reconciliation and fusion module, and an execution control drive module. The shadow state storage module stores control strategy parameters issued by the cloud management platform at preset intervals when the network connection is normal, forming a shadow state matrix containing proportional-integral-derivative parameter sets, environmental target setpoints, actuator state variables, and corresponding timestamps. When the network status monitoring module detects a network connection interruption, the shadow state storage module freezes and saves the current shadow state matrix as an optimization benchmark value. The edge autonomous optimization module, during the network connection interruption, uses the frozen and saved optimization benchmark value, combined with the environmental parameters collected by the field sensors, to... The rate of change is used to adaptively and dynamically adjust the proportional-integral-derivative parameter set, and the adjusted parameter set is stored as the local optimization state parameters. When the network status monitoring module detects that the network connection has been restored, the state harmonization and fusion module records the duration of the network connection interruption, obtains the control strategy parameters recalculated and issued by the cloud management platform, calculates the deviation between the environmental parameter values ​​collected in real time by the field sensors at the moment of network restoration and the target environmental parameter values ​​calculated by the cloud management platform, and constructs a fusion weight based on the duration and the deviation. The fusion weight is calculated nonlinearly based on the product relationship between the duration and the deviation, and decreases as the duration and the deviation increase. The state harmonization and fusion module performs weighted fusion of the local optimization state parameters and the control strategy parameters issued by the cloud based on the fusion weight to obtain the transition control parameters, and outputs the transition control parameters to the execution control drive module to drive the actuator.

[0008] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0009] 1. This invention introduces an edge autonomous optimization mechanism based on the real-time change rate of field environmental parameters during network connection interruption, replacing the traditional fixed threshold control logic. This enables the terminal to adaptively adjust the proportional, integral, and derivative parameters according to the real-time changes in the field environment, improving the control accuracy during network outages and effectively alleviating the problems of water and fertilizer waste and crop damage caused by sudden environmental changes.

[0010] 2. This invention introduces a fusion weight based on the network interruption duration and the cloud-edge environment deviation when the network connection is restored. This weighted fusion of local optimization state parameters and cloud control strategy parameters gradually reduces the control parameters during network switching rather than causing abrupt changes. This eliminates the oscillation shocks caused by the instantaneous control switching, avoids water hammer effects in the pipeline and mechanical damage to the actuators, and improves the accuracy of water and fertilizer application and the stability and reliability of system operation. Attached Figure Description

[0011] Figure 1 This is a vertical flowchart illustrating the overall system architecture in an embodiment of the present invention;

[0012] Figure 2 This is a schematic diagram of the vertical process of edge autonomous optimization during network outage in an embodiment of the present invention;

[0013] Figure 3 This is a schematic diagram of the vertical process of reconciling and merging the state after network recovery in an embodiment of the present invention. Detailed Implementation

[0014] The following is in conjunction with the appendix Figure 1-3 The specific embodiments of the present invention will be further described below. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0015] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0016] System overall architecture:

[0017] The agricultural information IoT terminal control system provided in this embodiment includes a cloud management platform, edge IoT terminals, and field sensors and actuators connected to the edge IoT terminals. The cloud management platform is equipped with a global strategy calculation unit, which is used to perform global optimization calculations based on large-scale meteorological data, historical planting data, and information gathered from multiple edge IoT terminals to obtain control strategy parameters and periodically send them to each edge IoT terminal. The edge IoT terminal is equipped with an embedded microcontroller and runs a lightweight real-time operating system. Internally, it is divided into a network status monitoring module, a shadow status storage module, an edge autonomous optimization module, a status reconciliation and fusion module, and an execution control drive module. The field sensors include soil moisture sensors and meteorological sensors, which are used to collect field environmental parameters. The actuators include an irrigation solenoid valve and a film rolling machine motor, which are used to perform agricultural production operations such as irrigation and film rolling according to control commands.

[0018] The network status monitoring module is used to detect the communication network connection status between the edge IoT terminal and the cloud management platform in real time. Its detection results are output to the edge autonomous optimization module and the status reconciliation and fusion module respectively, as conditions for triggering the corresponding processing flow. The shadow status storage module provides reference data for the edge autonomous optimization module and the status reconciliation and fusion module. The execution control drive module drives the field actuator to act according to the processed control parameters. The feedback signal of the actuator is transmitted back to the acquisition unit of the field sensor to form a closed-loop control loop.

[0019] The mechanism for constructing and freezing the shadow state matrix:

[0020] When the network connection is normal, the edge IoT terminal receives the control strategy issued by the cloud management platform at a preset cycle and stores it in a structured form as a shadow state matrix. The matrix includes the currently effective proportional, integral, and differential parameter groups, environmental target setpoints, current state variables of the actuators, and corresponding timestamps. When the network status monitoring module detects a communication interruption event, the edge IoT terminal immediately freezes and saves the shadow state matrix at the current moment as the optimization benchmark value. This benchmark value is not modified in subsequent local optimization processes and is only used as a reference starting point for adaptive adjustment. Through this freezing mechanism, the edge IoT terminal can use the historical control parameters issued by the cloud as the starting basis for local optimization during network outages, providing a traceable and recoverable reference benchmark for adaptive control during network outages.

[0021] Edge-autonomous optimization state machine:

[0022] During network connectivity interruptions, edge IoT terminals do not rely on fixed thresholds, but rather on frozen optimization benchmarks, combined with the real-time rate of change of data collected by on-site meteorological and soil sensors, to adaptively and dynamically adjust the proportional-integral-derivative parameters.

[0023] Specifically, after the network status monitoring module detects a network connection interruption event, it triggers the edge autonomous optimization module to start and freezes and saves the current shadow state matrix as an optimization benchmark value. This benchmark value includes a benchmark proportional coefficient, a benchmark integral coefficient, and a benchmark differential coefficient, denoted as follows: , , Subsequently, the edge IoT terminal collects the current values ​​of environmental parameters through on-site meteorological and soil sensors according to a preset sampling period. The edge autonomous optimization module calculates the rate of change of environmental parameters based on two adjacent sampling values. The calculation method is as follows:

[0024] ;

[0025] in, express Environmental parameter values ​​collected by the time sensor This represents the environmental parameter values ​​collected in the previous sampling period. This indicates the preset sampling period. express The rate of change of environmental parameters at any given time;

[0026] The edge autonomous optimization module is based on the rate of change. The baseline proportional-integral-derivative parameter set is adaptively adjusted in the following manner:

[0027] ;

[0028] ;

[0029] ;

[0030] in, , , These are the baseline proportional coefficient, baseline integral coefficient, and baseline differential coefficient for the aforementioned frozen storage. , , These are pre-calibrated adjustment gain coefficients, used to control the adjustment range of the parameters. This is the hyperbolic tangent function, whose value range is limited to the open interval from -1 to +1. It is used to limit the rate of change of the environment and prevent abrupt changes in the parameter. , , for Real-time proportional, integral, and derivative coefficients obtained after adaptive adjustment at each moment;

[0031] The edge IoT terminal applies the adjusted proportional-integral-derivative parameter set to the local execution control drive module to drive the actions of actuators such as irrigation solenoid valves and film rolling machine motors in real time. At the same time, it writes the adjusted parameter set and corresponding timestamps into a dynamic copy in local storage, which is recorded as the local optimization state parameter. This copy is stored independently of the aforementioned frozen optimization benchmark value and does not overwrite each other. The edge IoT terminal cyclically executes the above-mentioned environmental parameter acquisition, rate of change calculation and parameter adaptive adjustment process until the network status monitoring module detects that the network connection has been restored, and then enters the state reconciliation and fusion processing flow.

[0032] State harmonic fusion algorithm:

[0033] After the network connection is restored, the edge IoT terminal is not immediately taken over by the control strategy issued by the cloud management platform. Instead, it initiates a state reconciliation and fusion process to progressively weight and fuse the local optimized state parameters with the cloud control strategy parameters.

[0034] Specifically, after the network status monitoring module detects that the network connection has been restored, it triggers the state reconciliation and fusion module to start, records the duration of the network connection interruption, and records it as the network interruption duration. The edge IoT terminal reports the local optimization status parameters generated during the network interruption and the network interruption duration to the cloud management platform. The cloud management platform recalculates the control strategy parameters based on the global data accumulated during the network interruption, obtains the latest control strategy parameters, and sends them to the edge IoT terminal.

[0035] The cloud-edge environment deviation value at the instant the edge IoT terminal computing network recovers is calculated as follows:

[0036] ;

[0037] in, These are environmental parameter values ​​collected in real time by local sensors on edge IoT terminals. These are the target environment parameter values ​​calculated by the cloud management platform based on global data. The absolute deviation between the two is used to characterize the degree of difference between the actual local state and the expected state in the cloud.

[0038] The state reconciliation and fusion module takes the network outage duration and the deviation between the cloud and edge environments as inputs to construct a fusion weight decay function:

[0039] ;

[0040] in, This indicates the running time elapsed since the start of the state reconciliation and fusion processing procedure. The duration of the network outage recorded above. The pre-calibrated decay rate constant, This is a pre-calibrated environmental deviation sensitivity coefficient, used to characterize the amplification effect of environmental deviations on the fusion rate. The cloud edge environmental deviation value obtained from the aforementioned calculation. for The fusion weight of the local optimization state parameters at each time step, and its value is within... When it equals zero, it is one, and it follows... It increases and then monotonically decreases, gradually approaching zero;

[0041] The state reconciliation and fusion module is based on The local optimization state parameters and the cloud control strategy parameters are weighted and fused to obtain... Transition control parameters actually output at any given time:

[0042] ;

[0043] in, These are the local optimization control parameters acquired during the network outage. These are the control policy parameters issued by the cloud management platform. The fusion weights obtained from the aforementioned calculations, For the calculation The transient control parameters can be a set of proportional-integral-derivative parameters or a valve opening command;

[0044] The execution control driver module will Real-time data is sent to actuators such as irrigation solenoid valves and film rolling machine motors, as... Increase As the weight of local optimization state parameters gradually decreases, the weight of cloud control strategy parameters gradually increases, and the control quantity exhibits a continuous and gradual change rather than a step-like change. The state reconciliation and fusion module continuously makes judgments. Is it less than the preset convergence threshold? When the value is less than the preset convergence threshold, the state reconciliation and fusion processing is considered complete, control is fully transferred to the cloud control strategy parameters, and the edge IoT terminal switches back to network follow mode. When the value is not less than the preset convergence threshold, the aforementioned fusion calculation process continues. After the state reconciliation and fusion process is completed, the edge IoT terminal updates the shadow state matrix to the cloud control strategy parameters and clears the local optimization state parameter cache generated by this network interruption. The system then resumes normal network operation mode.

[0045] because Using the exponential decay form, the error function can be defined. As a candidate function of Lyapunov, the function value follows The control variable is increased monotonically and gradually converges to zero, thus ensuring that the change of the control variable is continuous, bounded and convergent during the state harmonic fusion process, and that overshoot oscillation will not occur.

[0046] Alternative implementation methods:

[0047] As another embodiment of the present invention, in the edge autonomous optimization stage, a fuzzy logic control rule base can be used to replace the proportional integral differential parameter adaptive adjustment method based on the hyperbolic tangent function. That is, a fuzzy rule table is pre-constructed with the environmental change rate as input and the control parameter adjustment amount as output. During the network connection interruption, the edge IoT terminal directly looks up the table to obtain the adjusted control parameters without performing continuous function calculations. The state harmonization and fusion processing flow after the network connection is restored and the aforementioned fusion weight decay function remain unchanged, and the aforementioned non-disruptive switching technical effect can still be achieved. This embodiment is suitable for application scenarios with low requirements for computing resources.

[0048] In another embodiment of the present invention, in the state reconciliation and fusion processing flow for network connection recovery, a piecewise linear decay function can be used instead of the aforementioned exponential decay function, i.e., a fusion weight is set. ,in The pre-set fusion duration constant, The linear decay method has low computational complexity and can be implemented in low-cost terminals with limited computing power, given the fusion time that has already elapsed. Although it is not as adaptive as the exponential decay function in terms of convergence rate, it can still achieve the gradual fusion of local optimization state parameters and cloud control strategy parameters, avoiding abrupt changes in control parameters and achieving the technical effect of eliminating network disturbances in this invention.

[0049] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. An agricultural information IoT terminal control system, comprising a cloud management platform, an edge IoT terminal, and field sensors and actuators connected to the edge IoT terminal, wherein the edge IoT terminal includes a network status monitoring module and an execution control drive module, the network status monitoring module is used to detect the network connection status between the edge IoT terminal and the cloud management platform, and the execution control drive module is used to drive the actuators to operate, characterized in that, The edge IoT terminal also includes a shadow state storage module, an edge autonomous optimization module, and a state reconciliation and fusion module. The shadow state storage module stores control strategy parameters issued by the cloud management platform at preset intervals when the network connection is normal, forming a shadow state matrix containing proportional-integral-derivative (PID) parameter sets, environmental target setpoints, actuator state variables, and corresponding timestamps. When the network state monitoring module detects a network connection interruption, the shadow state storage module freezes and saves the current shadow state matrix as an optimization benchmark value. The edge autonomous optimization module, during a network connection interruption, adaptively and dynamically adjusts the PID parameter sets based on the frozen and saved optimization benchmark value, combined with the real-time change rate of environmental parameters collected by the field sensors, and stores the adjusted PID parameter sets as local optimization state parameters. When the network status monitoring module detects that the network connection has been restored, the state reconciliation and fusion module records the duration of the network connection interruption, obtains the control strategy parameters recalculated and issued by the cloud management platform, calculates the deviation between the environmental parameter values ​​collected in real time by the field sensors at the moment of network restoration and the target environmental parameter values ​​calculated by the cloud management platform, and constructs a fusion weight based on the duration and the deviation. The fusion weight is calculated by nonlinear decay based on the product relationship between the duration and the deviation, and decreases as the duration and the deviation increase. The state reconciliation and fusion module performs weighted fusion of the local optimization state parameters and the control strategy parameters issued by the cloud management platform based on the fusion weight to obtain transition control parameters, and outputs the transition control parameters to the execution control drive module to drive the actuator to move.

2. The agricultural information IoT terminal control system according to claim 1, characterized in that: The edge autonomous optimization module adaptively and dynamically adjusts the proportional-integral-differential parameter set in the following manner: ； ； ； in, , , These are the benchmark proportional coefficient, benchmark integral coefficient, and benchmark differential coefficient included in the optimization benchmark value, respectively. , , These are the pre-calibrated adjustment gain coefficients. for The rate of change of the environmental parameters at any given time. It is the hyperbolic tangent function. , , for The proportional coefficient, integral coefficient, and derivative coefficient obtained after adaptive adjustment at each moment, the rate of change Calculate as follows: ； in, for The environmental parameter values ​​collected by the on-site sensors at the specified time. These are the environmental parameter values ​​collected in the previous sampling period. This is the preset sampling period.

3. The agricultural information IoT terminal control system according to claim 1, characterized in that, The fusion weights are constructed as follows: ； in, This refers to the running time elapsed since the state reconciliation and fusion module started the fusion process. The duration of the network connection interruption. The pre-calibrated decay rate constant, For the pre-calibrated environmental deviation sensitivity coefficient, For the aforementioned deviation, for The state harmonic fusion module calculates the transition control parameters according to the following method, where the weight of the local optimization state parameters at any given time is determined in the weighted fusion: ； in, The local optimization state parameters are... These are the control policy parameters issued by the cloud management platform. for The transition control parameters at time t.

4. The agricultural information IoT terminal control system according to claim 3, characterized in that: The state harmonization and fusion module is also used to continuously determine whether the fusion weight is less than a preset convergence threshold. When the fusion weight is less than the preset convergence threshold, the control is transferred to the control strategy parameters issued by the cloud management platform, the shadow state matrix in the shadow state storage module is updated to the control strategy parameters issued by the cloud management platform, and the local optimization state parameters are cleared.

5. The agricultural information IoT terminal control system according to claim 1, characterized in that: The actuators include an irrigation solenoid valve and a film rolling machine motor, and the field sensors include a soil moisture sensor and a weather sensor.

6. The agricultural information IoT terminal control system according to claim 1, characterized in that, The deviation is calculated as follows: ； in, The environmental parameter values ​​collected in real time by the field sensors at the moment the network is restored. The target environment parameter value is calculated by the cloud management platform based on global data.

7. A method for controlling an agricultural information IoT terminal, applied to an edge IoT terminal, characterized in that, include: When the network connection is normal, the system receives control strategy parameters from the cloud management platform at a preset cycle and stores them as a shadow state matrix containing proportional-integral-derivative (PID) parameter sets, environmental target setpoints, actuator state variables, and corresponding timestamps. When a network connection interruption is detected, the shadow state matrix at the current moment is frozen and saved as an optimization benchmark value. Based on the optimization benchmark value and the real-time rate of change of environmental parameters collected by field sensors, the PID parameter sets are adaptively and dynamically adjusted. The adjusted PID parameter sets are stored as local optimization state parameters, and the actuator is driven to move according to the adjusted PID parameter sets. When the network connection is detected to be restored, the system records... The system records the duration of network connection interruption, obtains the control strategy parameters recalculated and issued by the cloud management platform, calculates the deviation between the environmental parameter values ​​collected in real time by the field sensors and the target environmental parameter values ​​calculated by the cloud management platform at the moment the network is restored, constructs a fusion weight based on the duration and the deviation, and calculates the fusion weight nonlinearly based on the product relationship between the duration and the deviation, decreasing as the duration and the deviation increase. The system then performs a weighted fusion of the local optimization state parameters and the control strategy parameters issued by the cloud management platform based on the fusion weight to obtain transition control parameters, and drives the actuator to move based on the transition control parameters.

8. The agricultural information IoT terminal control method according to claim 7, characterized in that, The adaptive dynamic adjustment of the proportional-integral-differential parameter set is performed in the following manner: ； ； ； in, , , These are the benchmark proportional coefficient, benchmark integral coefficient, and benchmark differential coefficient included in the optimization benchmark value, respectively. , , These are the pre-calibrated adjustment gain coefficients. It is the hyperbolic tangent function. , , for The proportional coefficient, integral coefficient, and derivative coefficient obtained after adaptive adjustment at each time step. Calculate as follows: ； in, for The environmental parameter values ​​collected by the on-site sensors at the specified time. These are the environmental parameter values ​​collected in the previous sampling period. This is the preset sampling period.

9. The agricultural information IoT terminal control method according to claim 8, characterized in that, The fusion weights are constructed as follows: ； in, This refers to the running time elapsed since the process of acquiring the fusion weights began. The duration of the network connection interruption. The pre-calibrated decay rate constant, For the pre-calibrated environmental deviation sensitivity coefficient, For the deviation, the transition control parameter is calculated as follows: ； in, The local optimization state parameters are... These are the control policy parameters issued by the cloud management platform. for The transition control parameters at time t.

10. The agricultural information IoT terminal control method according to claim 9, characterized in that, Also includes: The system continuously determines whether the fusion weight is less than a preset convergence threshold. When the fusion weight is less than the preset convergence threshold, the control is transferred to the control strategy parameters issued by the cloud management platform. The shadow state matrix is ​​updated to the control strategy parameters issued by the cloud management platform, and the local optimization state parameters are cleared.

Images