CN-121970673-A - Accurate thing networking irrigation control system based on soil moisture content prediction

CN121970673ACN 121970673 ACN121970673 ACN 121970673ACN-121970673-A

Abstract

The application relates to the field of agricultural intelligent management and discloses an accurate Internet of things irrigation control system based on soil moisture content prediction, which comprises a data acquisition module, a soil moisture modeling module, an optimal irrigation decision module, an reinforcement learning optimization module and an intelligent irrigation execution module, wherein the data acquisition module is used for acquiring soil moisture; the method comprises the steps of collecting environmental data through an Internet of things sensor, establishing a soil moisture model by utilizing an Ito random differential equation, calculating transpiration evaporation loss by combining Penman-Montetith equation, solving an optimal irrigation strategy based on a random HJB equation, optimizing through deep reinforcement learning, and finally accurately controlling irrigation and dynamically adjusting the strategy by an intelligent irrigation execution module. According to the application, the Internet of things, edge calculation and deep reinforcement learning are combined, so that precise irrigation control is realized, water resource utilization is optimized, network delay is reduced, response speed is improved, closed-loop optimization is formed, crop requirements are precisely matched, irrigation adaptability and flexibility are improved, excessive or insufficient irrigation is avoided, and irrigation efficiency is improved.

Inventors

LI YAN
ZHOU JING
LI LING
SUO YANG
LI YING

Assignees

景天下生态环境科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260324

Claims (10)

1. Accurate thing networking irrigation control system based on soil moisture content prediction, its characterized in that includes: The data acquisition module is used for acquiring parameters of soil humidity, environmental temperature, rainfall and wind speed and transmitting acquired data to the soil moisture modeling module; the soil moisture modeling module is used for constructing a soil moisture dynamic model, predicting soil moisture based on the parameters acquired by the data acquisition module and providing a prediction result for the optimal irrigation decision module; The optimal irrigation decision module is used for calculating an optimal irrigation strategy based on the prediction result of the soil moisture modeling module and providing the calculation result to the reinforcement learning optimization module; the reinforcement learning optimization module is used for optimizing irrigation decisions by adopting a deep reinforcement learning method based on the calculation result of the optimal irrigation decision module and transmitting the optimized decisions to the intelligent irrigation execution module; and the intelligent irrigation execution module is used for receiving the optimization result of the reinforcement learning optimization module, controlling irrigation equipment to execute irrigation operation and feeding back the soil humidity change data after execution to the data acquisition module.
2. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the data acquisition module comprises: The soil humidity acquisition unit is used for acquiring soil humidity data of different depths, and comprises a surface layer, a middle layer and a deep layer, and the soil humidity acquisition unit comprises a soil humidity sensor and is used for measuring the moisture content of soil of each layer; an environmental monitoring unit for acquiring environmental meteorological parameters, the environmental monitoring unit comprising: the air temperature and humidity sensor is used for detecting the ambient temperature and the relative humidity; The wind speed and direction sensor is used for detecting wind speed and wind direction; the rainfall sensor is used for detecting rainfall; a solar radiation sensor for detecting solar radiation intensity; And the data transmission unit is used for transmitting the acquired data to the remote server through LoRa, wiFi or 5G wireless communication technology.
3. The precise internet of things irrigation control system based on soil moisture content prediction of claim 1, wherein the soil moisture modeling module comprises: The moisture dynamic modeling unit is used for constructing a soil moisture evolution model based on an Ito random differential equation based on the acquired soil humidity data and environmental factors; The transpiration and evaporation calculation unit is used for calculating transpiration and evaporation loss based on Penman-Monteth equation, and the transpiration and evaporation calculation unit calculates based on the temperature and humidity, wind speed and radiation data acquired by the data acquisition module; the data preprocessing unit is used for filtering, outlier rejection and interpolation correction of the soil humidity data acquired by the data acquisition module, and carrying out normalization processing on the data.
4. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the optimal irrigation decision module comprises: the target optimization calculation unit is used for constructing an optimal control target function and keeping the soil humidity within a target range by minimizing irrigation cost; and the random optimal control unit is used for calculating an optimal irrigation strategy based on a random Hamilton-Jacobi-Bellman equation and providing a calculation result to the reinforcement learning optimization module.
5. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the reinforcement learning optimization module comprises: the reinforcement learning environment modeling unit is used for defining a reinforcement learning state space, an action space and a reward function based on the calculation result of the optimal irrigation decision module; The strategy optimization unit is used for optimizing an optimal irrigation strategy by adopting a depth deterministic strategy gradient method and adjusting the strategy based on historical data and environmental feedback; and the value function approximation unit is used for approximating a value function solution of the HJB equation based on the neural network and optimizing the calculated optimal strategy based on reinforcement learning.
6. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the intelligent irrigation execution module comprises: an intelligent irrigation control unit for controlling irrigation equipment to perform irrigation based on an optimized irrigation strategy, wherein the irrigation equipment comprises a drip irrigation system, a spray irrigation system or an infiltrating irrigation system; the real-time monitoring feedback unit is used for acquiring the executed soil humidity change data, feeding back the soil humidity change data to the reinforcement learning optimization module through the data acquisition module and adjusting a subsequent decision strategy; And the edge calculation unit is used for locally processing part of calculation tasks, reducing data transmission delay and improving response speed.
7. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the soil moisture modeling module calculates using a soil moisture evolution model of the form: ; Wherein, the Is soil humidity; For irrigation input, dependent on control variables ; The transpiration and evaporation loss is affected by temperature and humidity, wind speed and radiation; for rainfall input, obey probability distribution ; Modeling the environmental randomness for a standard Wiener process; The noise intensity reflects the degree of random fluctuation.
8. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the optimal irrigation decision module calculates an optimal value function The following random HJB equation is satisfied: ; Wherein, the Partial derivatives of the value function with respect to time; For irrigation input, dependent on control variables ; The transpiration and evaporation loss is affected by temperature and humidity, wind speed and radiation; for rainfall input, obey probability distribution ; Is a function of the optimal value; as a function of irrigation cost; The soil moisture loss function is adopted; Partial derivatives that are functions of values; Second derivative of the value function; Is a coefficient related to system noise.
9. The precise internet of things irrigation control system based on soil moisture prediction of claim 1, wherein the reinforcement learning environment modeling unit of the reinforcement learning optimization module performs policy optimization using a reward function of the form: ; Wherein, the Is a prize value; Is irrigation cost; and (3) minimizing water resource consumption and keeping the soil humidity close to a target value as a soil moisture loss function.
10. The accurate internet of things irrigation control method based on soil moisture content prediction is applied to the accurate internet of things irrigation control system based on soil moisture content prediction as claimed in any one of claims 1 to 9, and is characterized by comprising the following steps: s1, collecting soil humidity, air temperature, wind speed and rainfall environment data through an Internet of things sensor; S2, establishing a soil moisture dynamic model based on an Ito random differential equation, and calculating transpiration evaporation loss based on Penman-Monteth equation; s3, calculating an optimal irrigation strategy based on a random HJB equation; s4, optimizing the calculated optimal irrigation strategy by adopting a deep reinforcement learning method; S5, controlling the irrigation equipment to perform irrigation through the intelligent irrigation execution module, and optimizing a subsequent decision strategy through real-time monitoring feedback.

Description

Accurate thing networking irrigation control system based on soil moisture content prediction Technical Field The invention relates to the field of agricultural intelligent management, in particular to an accurate Internet of things irrigation control system based on soil moisture content prediction. Background With the increase of the water resource demand of agricultural production, the traditional irrigation method is faced with the problems of water resource waste and unbalanced crop growth, the traditional irrigation system mostly depends on preset rules and simple sensor feedback, and cannot be flexibly adjusted according to the real-time change of soil humidity and different environmental factors, for example, some traditional irrigation control systems only can irrigate according to preset irrigation time without considering the specific humidity condition of soil, which often results in excessive irrigation or insufficient irrigation, wastes a large amount of water resources and cannot guarantee healthy growth of crops. In addition, many existing technologies adopt a centralized data processing method, all data are required to be transmitted to a central server for calculation and decision making, the method not only can lead to data transmission delay, but also can influence the real-time performance of irrigation decision making due to unstable network, but also is crucial in large-scale agricultural production in timely response, and if a system is difficult to quickly respond to soil humidity change in an instant, irrigation strategies can not be effectively adjusted, even crops are deficient in water or water is caused, and yield and quality are influenced. Finally, the traditional irrigation system is difficult to achieve comprehensive self-adaptive adjustment, although some systems can acquire soil humidity data through sensors and feed back the soil humidity data, the systems often lack strong data processing capacity and are difficult to intelligently optimize according to the feedback, in addition, many existing schemes cannot effectively combine deep learning and reinforcement learning technologies, so that irrigation strategies cannot be independently learned and optimized in complex environments, compared with the variability of natural environments and crop growth requirements, the traditional systems cannot usually make accurate and long-term effective decision adjustment, and therefore, the invention provides an accurate Internet of things irrigation control system based on soil moisture prediction, and the defects of the prior art are overcome. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an accurate Internet of things irrigation control system based on soil moisture content prediction, which solves the problems of response lag, inaccurate irrigation decision, water resource waste and difficulty in self-adaptive environment change of the traditional irrigation mode. In order to achieve the aim, the invention is realized through the following technical scheme that the accurate Internet of things irrigation control system based on soil moisture content prediction comprises: The data acquisition module is used for acquiring parameters of soil humidity, environmental temperature, rainfall and wind speed and transmitting acquired data to the soil moisture modeling module; the soil moisture modeling module is used for constructing a soil moisture dynamic model, predicting soil moisture based on the parameters acquired by the data acquisition module and providing a prediction result for the optimal irrigation decision module; The optimal irrigation decision module is used for calculating an optimal irrigation strategy based on the prediction result of the soil moisture modeling module and providing the calculation result to the reinforcement learning optimization module; the reinforcement learning optimization module is used for optimizing irrigation decisions by adopting a deep reinforcement learning method based on the calculation result of the optimal irrigation decision module and transmitting the optimized decisions to the intelligent irrigation execution module; and the intelligent irrigation execution module is used for receiving the optimization result of the reinforcement learning optimization module, controlling irrigation equipment to execute irrigation operation and feeding back the soil humidity change data after execution to the data acquisition module. Preferably, the data acquisition module includes: The soil humidity acquisition unit is used for acquiring soil humidity data of different depths, and comprises a surface layer, a middle layer and a deep layer, and the soil humidity acquisition unit comprises a soil humidity sensor and is used for measuring the moisture content of soil of each layer; an environmental monitoring unit for acquiring environmental meteorological parameters, the environmental monitoring unit comprising: th