Rice field robot cluster remote cooperative control system based on edge calculation

Abstract

The invention discloses a rice field robot cluster remote cooperative control system based on edge computing, and relates to the field of program control, and the system comprises a master control module which serves as a central control end, issues a control instruction to distributed edge nodes, and receives state feedback; the crop identification module is deployed at an edge node and acquires attribute data of crop types, growth stages and distribution density of each region of the target rice field through an image sensor; the environment acquisition module is used for being deployed at an edge node and acquiring rice field meteorological data and soil environment data in real time; the crop growth environment and the machine state are monitored and responded in real time, when an abnormal environment event is detected, crop identification data can be rapidly integrated to generate an influence event and evaluate an influence coefficient, and a nearby automatic machine is immediately matched and scheduled for intervention, so that the timeliness of operation is ensured, crop loss can be effectively prevented, and the working efficiency is improved. And the efficiency of agricultural production is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of program control, in particular to a rice field robot cluster remote collaborative control system based on edge computing. BACKGROUND

[0002] With the global population growth and the shortage of arable land resources, improving the yield and quality of crops through digital and intelligent means has become a national strategy and industry consensus, especially for staple foods such as rice. Precision agriculture requires grid and fine management of farmland, replacing the traditional extensive farming mode. Unmanned tractors, rice transplanter, plant protection drones, intelligent weeding robots and other automated agricultural robots can perform tasks such as seeding, fertilizing, spraying and field patrolling, but current single machines mainly work independently, with limited efficiency and collaboration. Commercialization of 5G, Internet of Things and edge computing technologies makes it possible to solve the massive and real-time data processing of agricultural sites.

[0003] Existing agricultural automation equipment often works independently, and the equipment lacks data interaction and collaboration, resulting in repeated work, resource conflicts or work blind spots. Many automated equipment performs work based on a pre-set, fixed schedule or path. The agricultural production environment is highly dynamic, and weather changes, crop emergencies and equipment failures can all cause static plans to fail. Existing systems cannot sense these changes and make real-time adjustments, have low intelligence, cannot quantify the severity of events, and are difficult to adaptively adjust strategies. Even in scenarios with multiple machines, task allocation is often pre-set or simply polled, without considering the real-time state of the machine, the spatiotemporal properties of the task itself and the execution path optimization. SUMMARY

[0004] (I) Technical problems solved In view of the above-mentioned shortcomings of the prior art, the present application provides a rice field robot cluster remote collaborative control system based on edge computing, which can effectively solve the problems of the prior art.

[0005] (II) Technical solutions To achieve the above object, the present application is realized by the following technical solutions: The present application discloses a rice field robot cluster remote collaborative control system based on edge computing, comprising: A general control module as a central control end, issuing control instructions to the distributed edge nodes and receiving state feedback; A crop recognition module deployed at the edge node, collecting attribute data of crop species, growth stage and distribution density of each area of the target rice field through an image sensor; An environment acquisition module for deployment at the edge node, for real-time acquisition of rice field meteorological data and soil environment data; The machine acquisition module is used to dynamically record the working status, location coordinates and functional attribute data of each automated machine in the cluster. The functional attributes include job type, load capacity and energy consumption level. The event generation module is used to receive real-time data from the crop identification module and the environmental acquisition module, and generate dynamic events that affect crop growth based on preset rules. The dynamic events include pest and disease warnings, irrigation needs, and fertility deficiency. The judgment module is used to calculate the impact coefficient of the dynamic event and compare it with a preset threshold. If the threshold is exceeded, a machine matching instruction is triggered. The machine matching module is used to receive instructions from the judgment module, extract the feature vector of the dynamic event submitted by the event generation module, match the corresponding automated machine based on the feature similarity, and generate the device call instruction. The progress editing module is used to generate machine configuration schemes, including execution time windows, geofences, and task sequences, by combining the current machine usage status and event impact characteristics. The machine control module is used to convert the configuration scheme into a set of machine-executable instructions to drive the target machine to perform the job task.

[0006] Furthermore, the working logic of the event generation module includes: It receives real-time meteorological data streams and soil sensor data streams from the environmental acquisition module, and receives crop image analysis results from the crop identification module; Construct a time-series-based environmental data change model to generate climate anomaly events when the fluctuation amplitude of meteorological data exceeds the standard deviation threshold. The convolutional neural network identifies crop leaf morphological features, and generates a pest and disease event when the proportion of lesion area detected is greater than a preset threshold. When the difference between the soil nitrogen, phosphorus and potassium content and the fertilizer requirement of the crop at the growth stage exceeds a preset threshold, a precision fertilization event is generated. The above event data is encapsulated into an event set including event type, severity level, geographic coordinates, timestamp, and confidence score, and then output to the judgment module.

[0007] Furthermore, the meteorological data collected by the environmental acquisition module includes light intensity, temperature, humidity, and precipitation, while the soil environmental data includes pH value and nitrogen, phosphorus, and potassium content.

[0008] Furthermore, the formula for calculating the dynamic event influence coefficient in the judgment module is as follows: ; In the formula, The overall impact coefficient of the representative event Representing the The area affected by the event Representing the The total area of ​​the target paddy fields Representing the Unit price of each crop The estimated production loss caused by the representative event. Represents the expected total yield of rice paddies. Represents the remaining effective processing time. Represents the maximum allowed response time. Represents crop sensitivity factors, Represents the correction coefficient for critical growth stages of crops. Represents spatial weighting factor, Represents economic weighting factors. The time-weighted factor increases with the critical growth period.

[0009] Furthermore, the machine matching module, based on the event type, spatial impact range, and urgency characteristics extracted from the events generated by the event generation module, and combined with the current working status, location coordinates, and functional attributes of the automated machines recorded by the machine acquisition module, queries a preset database for available automated machines that can handle the event type, cover the spatial impact range, and meet the urgency timeliness requirements. It then calculates a comprehensive matching degree based on the spatial distance between the available machine and the event location, the machine's current task queue load, and the efficiency level of the functions required to execute the event. Finally, it selects the automated machine with the highest comprehensive matching degree as the target device and generates a corresponding device call instruction.

[0010] Furthermore, the configuration scheme generation process of the progress editing module includes: The optimal movement path is planned based on the machine's current location and the topographic data of the paddy field. Based on the remaining battery power and estimated energy consumption of the machine, dynamically allocate task execution time windows; After rehearsing the task execution process and verifying the feasibility of the plan, the execution command is locked.

[0011] Furthermore, the machine control module can receive priority adjustment instructions input manually, dynamically modify the threshold parameters of the judgment module, visually display the real-time operation range map of the machine cluster, and support manually delineating key operation areas.

[0012] Furthermore, the machine control module is interconnected with a feedback module via a wireless network. The feedback module is used to monitor the machine's execution status in real time and send task progress and abnormal alarms back to the central control module.

[0013] Furthermore, the central control module is interconnected with a data repository via a wireless network. The data repository is used for distributed storage of crop attributes, machine data, event analysis results, and historical operation logs, providing a data sharing interface for each functional module.

[0014] Furthermore, the central control module is interconnected with the crop identification module, the environment acquisition module, and the machine acquisition module via a wireless network; the event generation module is interconnected with the crop identification module, the environment acquisition module, and the judgment module via a wireless network; the machine matching module is interconnected with the machine acquisition module, the judgment module, and the progress editing module via a wireless network; and the progress editing module is interconnected with the machine control module via a wireless network.

[0015] (III) Beneficial Effects Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: 1. Through edge computing architecture, data processing and decision-making functions are completed directly at distributed edge nodes in the paddy field, thereby enabling real-time monitoring and response to crop growth environment and machine status. When the environment acquisition module detects abnormal environmental events, the event generation module can quickly integrate crop identification data to generate impact events, and the judgment module evaluates the impact coefficient. If the coefficient exceeds the threshold, the machine matching module will immediately match and dispatch nearby automated machines to intervene, ensuring the timeliness of operations, effectively preventing crop losses, improving agricultural production efficiency, and enhancing the system's adaptability. It is especially suitable for large-scale paddy field clusters, and can maintain stable operation in dynamically changing environments, thereby improving overall crop yield and quality.

[0016] 2. Through the collaborative work of the machine matching module, progress editing module, and data storage module, intelligent allocation and management of automated machine resources are achieved, avoiding the problem of resource waste in traditional agriculture. The machine acquisition module records the attribute data of each machine in real time, while the progress editing module generates optimized configuration schemes based on the impact characteristics of current events, thereby ensuring that the machines only operate when necessary and in the best positions, accurately matching low-load machines and planning the minimum path, reducing unnecessary movement and energy consumption, lowering the operating costs of the machines, and improving the utilization rate of paddy field resources. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a schematic diagram of the framework of the present invention.

[0019] The numbers in the diagram represent: 1. Central control module; 2. Crop identification module; 3. Environmental acquisition module; 4. Machine acquisition module; 5. Event generation module; 6. Judgment module; 7. Machine matching module; 8. Progress editing module; 9. Machine control module; 10. Feedback module; 11. Data storage module. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] The present invention will be further described below with reference to embodiments.

[0022] This embodiment describes a remote collaborative control system for a rice paddy robot cluster based on edge computing, such as... Figure 1 As shown, it includes: The central control module 1, acting as the central control terminal, sends control commands to the distributed edge nodes and receives status feedback.

[0023] Crop identification module 2, deployed at edge nodes, collects attribute data on crop types, growth stages, and distribution densities in various areas of the target paddy field through image sensors.

[0024] Environmental data acquisition module 3 is deployed at edge nodes to collect real-time meteorological and soil environmental data of paddy fields. Meteorological data includes light intensity, temperature, humidity and precipitation, while soil environmental data includes pH value and nitrogen, phosphorus and potassium content.

[0025] Machine acquisition module 4 is used to dynamically record the working status, location coordinates and functional attribute data of each automated machine in the cluster. The functional attributes include job type, load capacity and energy consumption level. Event generation module 5 receives real-time data from the crop identification module and the environmental acquisition module, and generates dynamic events affecting crop growth based on preset rules. These dynamic events include pest and disease warnings, irrigation needs, and fertility deficiencies. The working logic of event generation module 5 includes: It receives real-time meteorological data streams and soil sensor data streams from the environmental acquisition module 3, and receives crop image analysis results from the crop identification module 2; Construct a time-series-based environmental data change model to generate climate anomaly events when the fluctuation amplitude of meteorological data exceeds the standard deviation threshold. The convolutional neural network identifies crop leaf morphological features, and generates a pest and disease event when the proportion of lesion area detected is greater than a preset threshold. When the difference between the soil nitrogen, phosphorus and potassium content and the fertilizer requirement of the crop at the growth stage exceeds a preset threshold, a precision fertilization event is generated. The above event data is encapsulated into an event set including event type, severity level, geographic coordinates, timestamp, and confidence score, and output to the judgment module 6.

[0026] Module 6 is used to calculate the impact coefficient of dynamic events and compare it with a preset threshold. If the threshold is exceeded, a machine matching instruction is triggered. The machine matching module 7 is used to receive instructions submitted by the judgment module 6, extract the feature vector of the dynamic event submitted by the event generation module 5, match the corresponding matching degree of the automated machine according to the feature similarity, and generate the device call instruction. Progress editing module 8 is used to generate machine configuration schemes, including execution time windows, geofencing, and task sequences, by combining the current machine usage status and event impact characteristics. The configuration scheme generation process includes: The optimal movement path is planned based on the machine's current location and the topographic data of the paddy field. Based on the remaining battery power and estimated energy consumption of the machine, dynamically allocate task execution time windows; After rehearsing the task execution process and verifying the feasibility of the plan, the execution command is locked.

[0027] Machine control module 9 is used to convert the configuration scheme into a set of machine-executable instructions to drive the target machine to perform the task. Machine control module 9 can receive priority adjustment instructions input by humans, dynamically modify the threshold parameters of judgment module 6, visualize the real-time operation range map of the machine cluster, and support manual delineation of key operation areas.

[0028] The machine control module 9 is connected to the feedback module 10 via a wireless network. The feedback module 10 is used to monitor the machine's execution status in real time and send back task progress and abnormal alarms to the central control module 1.

[0029] The central control module 1 is connected to the data storage repository 11 via a wireless network. The data storage repository 11 is used for distributed storage of crop attributes, machine data, event analysis results and historical operation logs, and provides a data sharing interface for various functional modules.

[0030] The main control module 1 is connected to the crop identification module 2, the environment acquisition module 3, and the machine acquisition module 4 via a wireless network. The event generation module 5 is connected to the crop identification module 2, the environment acquisition module 3, and the judgment module 6 via a wireless network. The machine matching module 7 is connected to the machine acquisition module 4, the judgment module 6, and the progress editing module 8 via a wireless network. The progress editing module 8 is connected to the machine control module 9 via a wireless network.

[0031] Compared with existing technologies, by collecting and processing crop, meteorological and soil data in real time through edge nodes and combining convolutional neural networks to quickly generate dynamic events, automated response and precise operation are achieved, reducing latency and human intervention, ensuring flexible collaborative control of robot clusters, and improving operational efficiency, energy consumption utilization and crop yield by dynamically matching machines, planning optimal paths and allocating resources. Compared with traditional centralized agricultural management systems, it is more adaptable to complex environments and improves overall sustainability.

[0032] At other levels, this embodiment provides a formula for calculating the impact coefficient of a dynamic event, specifically: ; In the formula, The overall impact coefficient of the representative event Representing the The area affected by the event Representing the The total area of ​​the target paddy fields was determined by retrieving pre-stored terrain data from data repository 11. Representing the Unit price of each crop The estimated production loss caused by the representative event. Represents the expected total yield of rice paddies. Represents the remaining effective processing time. Represents the maximum allowed response time. Representing crop susceptibility factors, determined by varietal resistance: resistant varieties = 0.1, susceptible varieties = 0.3. The correction coefficients for key growth stages of crops are: booting stage = 1.5, maturity stage = 0.8, and other stages = 1.0. The spatial weighting factor is determined by the crop type. Represents economic weighting factors. The time-weighted factor increases with the critical growth period.

[0033] In this embodiment, the machine matching module 7, based on the event type, spatial impact range, and urgency characteristics extracted from the events generated by the event generation module 5, and combined with the current working status, location coordinates, and functional attributes of the automated machines recorded by the machine acquisition module 4, queries a preset database for available automated machines that can handle the event type, cover the spatial impact range, and meet the urgency and timeliness requirements. It then calculates a comprehensive matching degree based on the spatial distance between the available machine and the event location, the machine's current task queue load, and the efficiency level of the functions required to execute the event. Finally, it selects the automated machine with the highest comprehensive matching degree as the target device and generates the corresponding device call instruction.

[0034] Compared with existing technologies, this system comprehensively analyzes characteristics such as event type, spatial impact range, and urgency, and combines them with the machine's current state, location, and functional attributes. It performs intelligent queries and matching calculations in a preset database, quantifies the overall matching degree through multiple factors such as spatial distance, task load, and execution efficiency, and selects the optimal target device to generate calling instructions. This improves the system's real-time response, resource allocation accuracy, and overall operational efficiency. Compared with existing systems that rely on traditional manual or simple rule matching, it reduces latency, lowers the risk of human error, and enhances adaptability in dynamic agricultural environments.

[0035] Working principle: This invention uses the central control module 1 as the central control terminal to issue control commands. The crop identification module 2 collects crop planting attributes of each area of ​​the current target paddy field. The environmental acquisition module 3 collects current weather data and field crop data. The machine acquisition module 4 records the attribute data of each available automated machine. The event generation module 5 obtains the data from the crop identification module 2 and the environmental acquisition module 3 to generate the current impact event. The judgment module 6 judges whether the impact coefficient of the impact event exceeds the preset threshold. If it does, the machine matching module 7 matches the corresponding associated automated equipment based on the extracted impact event features. The progress editing module 8 generates a configuration scheme based on the current usage status of the automated machine and the characteristics of the event impact. The machine control module 9 generates and applies the corresponding control instructions for the automated machine based on the scheme. The feedback module 10 provides feedback on the implementation status of the automated machine. The data storage repository 11 stores various analysis data and serves as a data repository for crops and machines.

[0036] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A remote collaborative control system for a rice field robot cluster based on edge computing, characterized in that, Comprise: A total control module as a central control terminal, issuing control instructions to the distributed edge nodes and receiving state feedback; A crop identification module deployed on the edge node, collecting attribute data of crop species, growth stage and distribution density of each region of the target rice field through image sensors; An environment acquisition module for real-time acquisition of rice field meteorological data and soil environment data deployed on the edge node; A machine acquisition module for dynamically recording the working state, position coordinates and functional attribute data of each automatic machine in the cluster, the functional attributes including job type, load capacity and energy consumption level; An event generation module for receiving real-time data from the crop identification module and the environment acquisition module, generating dynamic events affecting crop growth based on preset rules, dynamic events including pest warning, irrigation demand and fertilizer deficiency; A judgment module for calculating the influence coefficient of the dynamic event and comparing it with the preset threshold value, if it exceeds the threshold value, triggering the machine matching instruction; A machine matching module for receiving the instruction from the judgment module, extracting the feature vector of the dynamic event submitted by the event generation module, matching the automatic machine with the corresponding matching degree according to the feature similarity, and generating the device calling instruction; A progress editing module for generating a machine configuration scheme including execution time window, geographic fence and task sequence in combination with the current machine usage state and event influence characteristics; A machine control module for converting the configuration scheme into a machine executable instruction set to drive the target machine to execute the job task. 2.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The working logic of the event generation module includes: Receive real-time meteorological data stream and soil sensor data stream from the environment acquisition module, and receive crop image analysis results from the crop identification module; Build an environment data change model based on time series, generate climate anomaly events when the fluctuation amplitude of meteorological data exceeds the standard deviation threshold; Identify crop leaf shape features through convolutional neural network, generate pest events when the detection area ratio of disease spots is greater than the preset threshold; Generate precision fertilization events when the difference between soil nitrogen, phosphorus and potassium content and the required amount of fertilizer for crop growth stage exceeds the preset threshold; Encapsulate the above event data into an event set including event type, severity level, geographic coordinates, timestamp and confidence score, and output to the judgment module. 3.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The meteorological data collected by the environment acquisition module includes light intensity, temperature and humidity, and precipitation, and the soil environment data includes pH value and nitrogen, phosphorus and potassium content. 4.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The calculation formula of the dynamic event influence coefficient in the judgment module is: ； In the formula, represent the event comprehensive influence coefficient, represent the first event influence area, represent the first total area of the target rice field, represent the first crop unit price, represent the estimated yield loss caused by the event, represent the expected total yield of the rice field, represent the remaining effective processing time, represent the maximum allowed response time, represent the crop sensitivity factor, represent the crop growth critical period correction coefficient, represent the spatial weight factor, represent the economic weight factor, represent the time weight factor, which increases with the growth critical period. 5.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The machine matching module extracts the event type, spatial influence range and urgency characteristics from the events generated by the event generation module, combines the current working state, location coordinates and functional attributes of the automatic machine recorded by the machine acquisition module, and queries the available automatic machine in the preset database that can handle the event type, cover the spatial influence range, and meet the urgency requirements of the emergency degree; According to the spatial distance between the available machine and the event occurrence place, the current task queue load of the machine and the efficiency level of the required function, the comprehensive matching degree is calculated; The target device is selected as the automated machine with the highest comprehensive matching degree, and corresponding device calling instructions are generated. 6.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The configuration scheme generation process of the progress editing module includes: Planning an optimal movement path based on the current position of the machine and the rice field topography data; According to the remaining battery capacity of the machine and the work energy consumption estimation, dynamically allocate the task execution time window; Preform the task execution process, verify the feasibility of the scheme, and lock the execution instructions. 7.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The. 8.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The machine control module is connected to the feedback module through a wireless network, and the feedback module is used to monitor the machine execution state in real time and return the task progress and abnormal alarm to the general control module. 9.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The general control module is connected to the data storage through a wireless network, and the data storage is used to store crop attributes, machine data, event analysis results and historical operation logs in a distributed manner, and provides a data sharing interface for each functional module. 10.The edge computing based remote collaborative control system for rice field robot swarm according to claim 1, wherein, The general control module is connected to the crop recognition module, the environment acquisition module and the machine acquisition module through a wireless network, the event generation module is connected to the crop recognition module, the environment acquisition module and the judgment module through a wireless network, the machine matching module is connected to the machine acquisition module, the judgment module and the progress editing module through a wireless network, and the progress editing module is connected to the machine control module through a wireless network.

Images