Agricultural equipment information scheduling management method and system applying cloud-side collaboration

Abstract

The invention relates to the technical field of cloud edge collaboration, solves the technical problems that in the prior art, conflict detection and solution means are single, complex and changeable scenes are difficult to dynamically adapt, and real-time conflicts cannot be fully considered, and particularly relates to an agricultural equipment information scheduling management method and system applying cloud edge collaboration. The method comprises the steps that original environment data and equipment data of agricultural equipment information are collected, expected data are preset, and a standardized environment data set, a standardized equipment data set and a standardized expected data set are generated through classification preprocessing. According to the method, the equipment capability is accurately measured through comprehensive efficiency sorting, the time, energy consumption, efficiency and load can be comprehensively considered through multi-target cost aggregation, the Hungary algorithm is improved to ensure global optimal matching, the reasonability of task allocation, the resource utilization rate and the system robustness are remarkably improved, and the processing efficiency of sudden changes of the environment is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of cloud-edge collaboration technology, and in particular to a method and system for scheduling and managing agricultural equipment information using cloud-edge collaboration. Background Technology

[0002] Cloud-edge collaboration is a computing architecture that combines the central computing power of cloud computing with the real-time response capabilities of edge computing. Agricultural equipment information scheduling and management refers to the management method of task allocation, status monitoring and resource optimization for intelligent equipment in farmland. Existing technologies usually adopt a method of global planning in the cloud and local execution at the edge. The cloud is responsible for big data analysis and long-term scheduling strategy generation, while edge nodes dynamically adjust equipment tasks based on real-time environmental data. For example, lightweight algorithms can be deployed through containerization technology, combined with resource awareness and task migration mechanisms, to achieve low-latency and high-reliability farmland operation scheduling.

[0003] However, existing technologies lack overall collaborative optimization. Path planning may not fully consider real-time conflicts, and conflict detection and resolution methods are relatively simple, making it difficult to dynamically adapt to complex and ever-changing scenarios. For example, multiple drones may be spraying pesticides, and only detecting conflicts in the intersecting flight paths may ignore pesticide drift caused by factors such as wind direction in different areas, resulting in conflicts in the actual operation area, affecting crop growth, and leading to problems such as poor adaptability to changing scenarios, inefficient and simple task scheduling, and inaccurate detection. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for scheduling and managing agricultural equipment information using cloud-edge collaboration. This solves the problems of existing technologies having limited conflict detection and resolution methods, difficulty in dynamically adapting to complex and ever-changing scenarios, and inability to fully consider real-time conflicts.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: an agricultural equipment information scheduling and management method employing cloud-edge collaboration. The method comprises the following steps: collecting raw environmental data and equipment data of agricultural equipment, and pre-setting expected data; generating a standard dataset through classification preprocessing; wherein the standard dataset includes at least: a standardized environmental dataset, a standardized equipment dataset, and a standardized expected dataset. Dynamic priority processing is performed based on standard datasets to obtain dynamically prioritized tasks. Perform collaborative matching scheme processing on the standardized equipment dataset and the dynamic priority task queue to obtain a set of matching schemes; Based on the matching scheme set, a time-sequential scheduling instruction set is generated through adaptive scheduling decision processing. The system collects the real-time status and progress of agricultural equipment, and obtains a scheduling instruction set through flexible adjustments based on the time-series scheduling instruction set and the real-time status and progress. The scheduling instruction set is then sent to the control center.

[0006] Preferably, a standard dataset is generated through classification preprocessing, including: obtaining a standardized environmental dataset by alignment and normalization processing based on the original environmental data; The device data is processed by state mapping to obtain a standardized device dataset; A standardized expectation dataset is obtained by using natural language processing based on the expected data.

[0007] Preferably, dynamic priority processing based on standard datasets includes: obtaining a granular expectation task set by spatial grid decomposition based on a standardized expectation dataset; Based on the particle expectation task set and the standardized environment dataset, a particle urgency assessment set is obtained through multidimensional fuzzy evaluation. By performing weighted dynamic prioritization on the particle urgency assessment set and the particle expected task set, dynamic priority tasks are obtained.

[0008] Preferably, the urgency assessment set and the expected task set of particles are subjected to weighted dynamic prioritization processing, including: weighted scoring processing of the urgency assessment set and the expected task set of particles to obtain an initial priority task table; Based on the initial priority task table and the standardized device dataset, a dynamic priority score set is obtained through dynamic priority processing. Dynamic priority tasks are obtained by prioritizing the dynamic priority score set and the granular expected task set.

[0009] Preferably, the standardized equipment dataset and the dynamic priority task queue are processed by a collaborative matching scheme, including: extracting performance index matrix from the standardized equipment dataset through index quantification; Based on the performance index matrix, the index weight vector is obtained by processing it using the entropy weight method. The performance evaluation and scheme generation operations are performed on the indicator weight vector and performance indicator matrix to obtain a set of matching schemes.

[0010] Preferably, the process of evaluating the performance of the indicator weight vector and the performance indicator matrix and generating a scheme includes: obtaining a comprehensive performance score set by performing performance evaluation on the indicator weight vector and the performance indicator matrix; The cost matrix is ​​obtained by multi-objective aggregation based on dynamic priority tasks, comprehensive performance score set, and standardized equipment dataset; Based on the cost matrix, a set of matching schemes is obtained through optimal matching.

[0011] Preferably, the adaptive scheduling decision processing based on the matching scheme set includes: processing the matching scheme set through instruction template filling to obtain a time-free scheduling instruction set; Based on the timeless scheduling instruction set and standardized equipment dataset, a planned path instruction set is obtained through critical path processing; The planned path instruction set is dynamically scheduled and optimized to obtain a time-sequential scheduling instruction set.

[0012] Preferably, the planned path instruction set is dynamically scheduled and optimized, including: obtaining a set of conflict relationships based on the planned path instruction set through conflict cross detection; Dynamic scheduling optimization is performed on the conflict relationship set and the planned path instruction set to obtain an optimized instruction start set; Based on the optimized instruction start set and the planned path instruction set, a time-sequential scheduling instruction set is generated through scheduling instruction processing.

[0013] Preferably, the process involves flexible adjustment based on the time-series scheduling instruction set and real-time status progress, including: obtaining an execution deviation report by comparing and processing deviations based on the time-series scheduling instruction set and real-time status progress. Based on the performance deviation report, the deviation assessment and adjustment process is carried out to generate an adjustment decision set; Based on the adjustment decision set, execution deviation report, standardized equipment dataset, and dynamic priority tasks, an updated scheduling instruction set is obtained through incremental scheduling adjustments.

[0014] The technical solution also provides a system for agricultural equipment information scheduling and management applied to the above-mentioned cloud-edge collaborative method, the system comprising: The standardization module is used to collect raw environmental and equipment data from agricultural equipment, and preset expected data to generate a standard dataset through classification preprocessing; the standard dataset includes at least: a standardized environmental dataset, a standardized equipment dataset, and a standardized expected dataset; The dynamic prioritization module is used to perform dynamic prioritization based on a standard dataset and obtain dynamic priority tasks. The matching scheme module is used to perform collaborative matching scheme processing on the standardized equipment dataset and the dynamic priority task queue to obtain a set of matching schemes. The adaptive scheduling module is used to generate a time-sequential scheduling instruction set based on the matching scheme set through adaptive scheduling decision processing. The flexible adjustment module is used to collect the real-time status and progress of agricultural equipment. Based on the time-series scheduling instruction set and the real-time status and progress, it obtains a scheduling instruction set through flexible adjustment and sends the scheduling instruction set to the control center.

[0015] By employing the above technical solution, the present invention provides a method and system for scheduling and managing agricultural equipment information using cloud-edge collaboration, which has at least the following beneficial effects: 1. This invention eliminates data discrepancies through spatiotemporal alignment and normalization, providing a unified and accurate environmental status basis for subsequent analysis. It quickly and accurately parses raw equipment status data, ensuring the real-time nature and accuracy of equipment status information. By using natural language processing and key information extraction to process expected data, it improves the standardization and processability of task information. The overall process enhances data utilization efficiency and decision-making scientificity.

[0016] 2. This invention improves the precision of task processing through spatial gridding and atomic task generation. The multi-dimensional feature urgency fuzzy evaluation comprehensively considers multiple key factors, making the urgency assessment more scientific and comprehensive. It rationally allocates initial priorities and uses a dynamic priority adjustment algorithm to enhance the flexibility of task scheduling, ensuring that high-priority tasks are executed first, thus improving overall execution efficiency. In response to sudden large-scale agricultural disasters, such as large-scale insect outbreaks, it can also quickly and accurately locate the affected areas and generate granular tasks. Based on equipment deployment and the passage of time, it can flexibly adjust task priorities to ensure that equipment is deployed to the most urgent and critical areas first, minimizing disaster losses.

[0017] 3. This invention can handle emergencies in real time, such as sudden strong winds during spraying. Existing technologies may cause problems such as drones deviating from their flight paths and uneven spraying due to the lack of dynamic adjustment of equipment performance evaluation and task allocation. In the face of similar agricultural emergencies, the invention uses the entropy weight method to objectively assign weights to avoid subjective biases, comprehensively ranks performance to accurately measure equipment capabilities, aggregates multi-objective costs to comprehensively consider time, energy consumption, performance and load, and improves the Hungarian algorithm to ensure global optimal matching. This significantly improves the rationality of task allocation, resource utilization and system robustness, and greatly improves the efficiency of handling sudden environmental changes.

[0018] 4. This invention can take into account both path planning and conflict resolution, comprehensively consider the conflict at the intersection point and the conflict in the work area, and detect more accurately. It avoids the problem that when multiple drones are spraying pesticides, the pesticide drift caused by wind direction and other factors in different areas may be ignored, which may lead to actual conflicts in the work area. At the same time, through reinforcement learning and dynamic optimization, it can adapt to different scenarios and improve the real-time performance, adaptability, flexibility and accuracy of scheduling. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 The flowchart illustrates the cloud-edge collaborative agricultural equipment information scheduling and management method of this invention. Figure 2 This is a structural block diagram of the cloud-edge collaborative agricultural equipment information scheduling and management system of the present invention. Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding and implementation of how the present application uses technical means to solve technical problems and achieve technical effects.

[0021] Example 1: Because existing conflict detection and resolution methods are relatively simple and difficult to adapt dynamically to complex and ever-changing scenarios, and cannot fully consider the technical issues of real-time conflicts, please refer to... Figure 1 This embodiment provides a method and system for scheduling and managing agricultural equipment information using cloud-edge collaboration. This avoids the problem of pesticide drift caused by wind direction and other factors in different areas when multiple drones are spraying pesticides, which could lead to actual conflicts in the work area. Furthermore, through reinforcement learning and dynamic optimization, it can adapt to different scenarios, improving the real-time performance, adaptability, flexibility, and accuracy of scheduling. The method includes the following steps: S1. Collect raw environmental and equipment data of agricultural equipment, and pre-set expected data. Generate a standard dataset through classification preprocessing. The standard dataset includes at least: a standardized environmental dataset, a standardized equipment dataset, and a standardized expected dataset. In terms of existing multi-source data collection, the lack of unified standards often leads to difficulty in aligning data from different sources in the spatiotemporal dimensions, making it impossible to accurately analyze complex equipment status data. Furthermore, the status mapping is unclear, resulting in inaccurate and unreal-time equipment status information. To solve the above problems, the specific implementation steps are as follows: S11. Based on the original environmental data, a standardized environmental dataset is obtained through alignment and normalization. In this step, the original environmental data is first aligned in time and space to ensure that data collected from different sources and at different times are consistent in time and space, laying the foundation for subsequent unified processing. After completing the time and space alignment, various environmental parameters are normalized. Specifically, for each environmental parameter, its historical data is collected, the minimum and maximum values ​​in the historical data are found, the current original data value is subtracted from the historical minimum value, and the result is divided by the difference between the historical maximum and the historical minimum value. In this way, the original data is mapped to the range of 0 to 1, and finally a standardized environmental dataset is obtained. Each element in this set represents the standardized value of the corresponding environmental parameter. The original environmental data includes meteorological data such as weather data, temperature data, and humidity data, as well as environmental data such as soil moisture data.

[0022] S12. Perform state mapping processing on the device data to obtain a standardized device dataset. This step first parses the data according to the communication protocol followed by the device. Since different devices use different communication protocols, the protocol parsing process extracts the messy information from the original data according to predetermined rules, such as Modbus protocol parsing rules, CAN bus protocol parsing rules, or ASCII encoding rules, to obtain key information fragments of the device. Then, state mapping is performed. For each type of parsed information, it is transformed according to pre-set rules. For device location information, it is converted from the original positioning data format to a unified coordinate representation. For battery power, based on historical data such as the device's battery rated capacity, the original power value is converted into the remaining energy percentage. This intuitively reflects the remaining battery usage. The device type is converted into a concise code using specific coding rules. The working status is determined based on the actual operation of the device, referring to a preset status classification to clarify whether it is idle, in operation, or faulty. After protocol parsing and status mapping, the device's unique identifier, converted location coordinates, remaining energy ratio, device type code, and working status are combined to form a state vector. Numerous such state vectors constitute a standardized device dataset. Agricultural equipment includes smart tractors, drones, irrigation robots, various sensor nodes such as temperature and humidity sensors, and edge computing gateways for target agriculture. Device data includes GPS location, battery level, operating speed, and device type.

[0023] S13. Based on the expected data, a standardized expected dataset is obtained through natural language processing. The expected data includes agricultural task number, execution time, operation type, operation area, crop information, manpower requirements, material requirements, quality requirements, time arrangement, and other agricultural implementation-related data. For the input expected data, natural language processing technology is first used to perform word segmentation, part-of-speech tagging, and other operations on the text content of the expected data. The continuous text is divided into meaningful word units, and the part of speech of each word is determined to prepare for subsequent key information extraction. Then, according to the pre-set rules and models, such as regular expressions for matching text of a specific format, the models include deep learning-based named entity recognition models, such as the BiLSTM CRF model, which will not be elaborated here. Key information is extracted from the processed text. For agricultural task numbers, information is obtained from specific identifiers in the expected data or according to certain numbering rules, such as the distance of the number coordinates from the center point. For the area to be worked on, the area described in the text is converted into polygon coordinates by combining common regional description methods in historical data and relevant knowledge of geographic information systems. The task type is determined by matching the task actions explicitly mentioned in the expected data with the preset task type classification. The expected completion time is determined by recognizing the time expression in the text and converting it according to a unified time format. Finally, the extracted task ID, polygon coordinates of the area to be worked on, task type, and expected completion time are combined to form a task requirement vector. Many such task requirement vectors constitute the standardized expected dataset.

[0024] This invention eliminates data discrepancies through spatiotemporal alignment and normalization, providing a unified and accurate environmental status basis for subsequent analysis. It quickly and accurately parses raw equipment status data, ensuring the real-time nature and accuracy of equipment status information. By using natural language processing and key information extraction to process expected data, it improves the standardization and processability of task information. The overall process enhances data utilization efficiency and the scientific nature of decision-making.

[0025] S2. Dynamic priority processing is performed based on standard datasets to obtain dynamic priority tasks. Existing technologies do not decompose tasks finely enough, resulting in unreasonable resource allocation. In response to sudden large-scale agricultural disasters, such as large-scale insect outbreaks, the task decomposition may be too coarse to quickly and accurately locate the core disaster area, make it impossible to judge the urgency of each area in a timely manner, and set fixed priorities that cannot be flexibly adjusted according to equipment deployment and time changes, thus delaying disaster relief and increasing disaster losses. To solve the above problems, the specific implementation steps are as follows: S21. Based on the standardized expectation dataset, spatial grid decomposition is used to obtain the granular expectation task set. In this step, the standardized expectation dataset is used as input. This dataset contains key information such as task identifier, work area, work type, and expected completion time. First, based on the single operation capability of agricultural equipment and combined with the actual operation efficiency data of the equipment in different areas in historical operations, the work area is decomposed into a detailed geospatial grid. Specifically, the entire work area is divided into small sub-grid polygons according to the range that the equipment can effectively operate in one go. These sub-grid polygons are the basic work areas of subsequent atomic tasks. For each sub-grid polygon, its parent task identifier is determined according to its relationship with the parent task area. At the same time, each sub-grid is assigned a unique atomic task identifier. Then, the area of ​​each sub-grid polygon is calculated. This area data is determined based on the measurement method of similar areas in historical operations and the equipment operation accuracy requirements. Finally, the information such as parent task identifier, atomic task identifier, sub-grid polygon, work type, and sub-grid area are combined to form atomic tasks. All these atomic tasks are combined to form the granular expectation task set. For example, in a large smart farm, the farm manager issues an irrigation task through the system. After the previous steps, the task forms a standardized expectation dataset. If there is a 100-acre farmland that needs irrigation, based on the irrigation equipment's ability to cover 5 acres at a time, it is divided into 20 sub-grid polygons. For each sub-grid polygon, the system assigns a parent task identifier, which is the identifier of the entire irrigation task, an atomic task identifier, used to uniquely distinguish the irrigation task of each sub-grid, the specific geographical coordinates of the sub-grid polygon, the operation type such as irrigation, and the sub-grid area such as 5 acres. This information is combined to form 20 atomic tasks, which constitute the granular expectation task set.

[0026] S22. Based on the expected task set and standardized environment dataset, obtain the granular urgency assessment set through multidimensional fuzzy evaluation. In this step, for each task in the expected task set, three key features are determined by combining historical task execution and current actual needs: time to deadline, associated disaster index, and meteorological window urgency. The time to deadline feature considers the remaining time before the task is scheduled to be completed. The associated disaster index feature assesses the correlation between the current task and possible disasters based on data on the impact of past disasters on agricultural operations. The meteorological window urgency feature refers to historical meteorological data and meteorological information in the current standardized environment state set to determine the remaining time of suitable meteorological conditions for task execution. Next, weights are assigned to these three features. The magnitude of the weights is determined based on the influence of each feature on the urgency of the task in historical data. For example, if the time remaining before the deadline has had a significant impact on the urgency of the task in the past, its corresponding weight will be relatively high. Simultaneously, membership functions are constructed for each feature. The membership function for the time remaining before the deadline can be set using the maximum-minimum standardized formula. The membership function for the disaster index can use the probability density function formula of the normal distribution. The membership function for the urgency of the weather window can be obtained by dividing the remaining time by the total duration of the weather conditions. These formulas are all commonly used calculation formulas. Without going into details, based on the correspondence between each feature value and the urgency of tasks in historical data, the urgency level of different feature values ​​is determined. Through the calculation of membership functions, the evaluation value of each task on the three features is obtained. These evaluation values ​​are combined into an evaluation matrix. Finally, the weight vector and the evaluation matrix are fused. This is equivalent to performing a certain correlation operation between each weight in the weight vector and the evaluation value of the corresponding column in the evaluation matrix, and finally obtaining a scalar value. This scalar value represents the urgency of the corresponding task. The scalar values ​​of the urgency of all tasks are combined to form a granular urgency evaluation set.

[0027] S23. Perform weighted dynamic prioritization on the particle urgency assessment set and the particle expected task set to obtain dynamic priority tasks. S231. Perform weighted scoring on the particle urgency assessment set and the particle expected task set to obtain an initial priority task table. In this step, firstly, obtain the urgency of each task from the particle urgency assessment set. This urgency is based on a comprehensive assessment of multiple characteristics such as the task's deadline, associated disaster index, and meteorological window urgency. At the same time, obtain the corresponding work area information for each task from the particle expected task set. Then, determine the adjustment weights, which can be adjusted based on historical data. The sum of the two weights is fixed to a specific value, such as 1. When calculating the initial priority score of a task, multiply the task's urgency by one of the weights. This is equivalent to initially weighting the task priority based on urgency. Then, multiply the work area of ​​the task by the ratio of the largest work area among all tasks and the other weight. Finally, merge the two product results to obtain the initial priority score for each task. Sort all tasks in descending order of initial priority score to form the initial priority task table.

[0028] S232. Based on the initial priority task table and the standardized device dataset, a dynamic priority score set is obtained through dynamic priority processing. In this step, the initial priority task table records the initial priority score of each task, while the standardized device status set reflects the current resource competition situation, including information such as the number of idle devices and the total number of devices of each type. For each task in the initial priority task table, on the one hand, the time decay factor is considered. From the time the task is generated, its initial priority score will gradually decrease according to a certain rule. Here, an exponential function related to the time decay coefficient is used to simulate this change. On the other hand, the resource competition factor is considered. The priority increase value is obtained by multiplying the ratio of the number of idle devices that can execute this type of task to the total number of devices of this type by the resource competition influence factor. The resource competition influence factor can be obtained through the analytic hierarchy process to measure the impact of resource scarcity on task priority. The more scarce the resources, that is, the smaller the proportion of idle devices to the total number of devices, the more obvious the effect of this factor on the task priority. The initial priority score is added to the priority increase value to finally obtain the dynamic priority score of each task at the current moment. The dynamic priority scores of all tasks constitute the dynamic priority score set.

[0029] S233. Based on the dynamic priority score set and the granular expected task set, dynamic priority tasks are obtained by prioritizing them. In this step, each score in the dynamic priority score set is associated and matched with the corresponding atomic task in the granular expected task set to form task combinations with priority scores. Then, according to the priority scores in these task combinations, all atomic tasks are arranged in descending order, with tasks with higher priority scores being placed at the front. The final atomic task sequence obtained after this descending order arrangement is the dynamic priority task queue, which clearly shows the execution priority order of each task at the current moment.

[0030] This invention enhances the precision of task processing through spatial gridding and atomic task generation. Multi-dimensional feature urgency fuzzy evaluation comprehensively considers multiple key factors, making urgency assessment more scientific and comprehensive. A weighted scoring method balances urgency and efficiency, rationally allocating initial priorities. A dynamic priority adjustment algorithm enhances the flexibility of task scheduling, ensuring high-priority tasks are executed first, improving overall execution efficiency. In response to sudden large-scale agricultural disasters, such as widespread insect outbreaks, this invention can quickly and accurately locate affected areas and generate granular tasks, rapidly assessing the urgency of each area. It flexibly adjusts task priorities based on equipment deployment and time progression, ensuring equipment is prioritized for the most urgent and critical areas, minimizing disaster losses.

[0031] S3. Perform collaborative matching scheme processing on the standardized equipment dataset and dynamic priority task queue to obtain a matching scheme set; In the scenario of agricultural drone spraying, if a sudden strong wind is encountered during spraying, the existing technology may cause the drone to deviate from the flight path, spray unevenly, or even crash due to excessive load, resulting in resource waste and operation failure because the existing technology does not dynamically adjust the equipment performance assessment and task allocation. To solve the above problems, the specific implementation steps are as follows: S31. Based on the standardized equipment dataset, extract the performance index matrix through index quantification. In this step, the index quantification method is used to calculate the equipment performance index matrix. For each piece of equipment, its performance is considered from four dimensions. First is the energy sufficiency, which directly uses the energy value of the current state of the equipment as the index of this dimension. The more sufficient the energy, the stronger the continuous operation capability of the equipment. Second is the lightness of the current load, which is based on the number of tasks currently undertaken by the equipment or the estimated remaining operation time. Through a specific conversion method, such as the fewer the number of tasks or the operation time, the closer the index value is to a larger value, to measure the lightness of the current load of the equipment. The lighter the load, the stronger the ability of the equipment to quickly put into new tasks. Secondly, there's the historical reliability score, which is derived from a comprehensive analysis of the equipment's past operational records and performance, reflecting its stability and reliability. The more stable and reliable the historical performance, the higher the score. Finally, there's the comprehensive capability matching benchmark score, which assigns a fixed score based on the degree of matching between the equipment type and general operational requirements. For example, a plant protection machine would receive a perfect score when performing a spraying task because its type is highly matched to the task. Combining these four dimensions of indicators forms the static and dynamic performance indicator vectors for each piece of equipment. The sum of all the performance indicator vectors constitutes the performance indicator matrix.

[0032] S32. Based on the performance index matrix, the entropy weight method is used to obtain the index weight vector. In this step, the performance index matrix is ​​first normalized by dividing the data in each index column by the sum of all data in that column. Each element in the new matrix after this process reflects the relative proportion of the corresponding equipment in that index. Next, the entropy value of each index is calculated. The entropy value is used to measure the dispersion of the index. The calculation process is to first determine a constant related to the number of equipment, and then multiply this constant by the sum of the products of the relative proportion of each equipment under each index and its natural logarithm. After obtaining the entropy value, the difference coefficient is calculated. The difference coefficient is 1 minus the entropy value. The larger the difference coefficient, the greater the dispersion of the index, and the higher the discrimination of the equipment performance evaluation. Finally, the weight of each index is calculated based on the difference coefficient. The difference coefficient of each index is divided by the sum of the difference coefficients of all indices. The value obtained is the weight of the index. The weights of all indices are combined to form the index weight vector.

[0033] S33. Perform performance evaluation and scheme generation operations on the indicator weight vector and performance indicator matrix to obtain a set of matching schemes; S331. Based on the indicator weight vector and the performance indicator matrix, a comprehensive performance score set is obtained through performance evaluation. In this step, the performance indicator matrix is ​​first weighted using the indicator weight vector, multiplying the weight of each indicator by the corresponding equipment value for that indicator, thus constructing a weighted normalized decision matrix. This process is equivalent to requantifying the equipment's performance across different performance dimensions based on the importance of each indicator. Next, the ideal solution and negative ideal solution are determined from the weighted normalized decision matrix. The ideal solution is the combination of the maximum values ​​in each column, representing the optimal state the equipment can achieve across all performance indicators. The negative ideal solution is the combination of the maximum values ​​in each column. The minimum combination in a column represents the worst-case state. Then, the distance to the ideal solution and the negative ideal solution for each device is calculated separately. When calculating the distance, the difference between the device and the ideal solution or the negative ideal solution in each index is squared, summed, and then squared to measure the degree of deviation of the device from the optimal and worst-case states. Finally, the relative proximity is calculated to evaluate the overall performance of the device. The relative proximity is the distance of the device to the negative ideal solution divided by the sum of the distances of the device to the ideal solution and the negative ideal solution. Its value ranges from zero to one. The larger the value, the closer the device is to the ideal solution and the better the overall performance. The relative proximity of all devices constitutes the overall performance score set.

[0034] S332. Based on dynamic priority tasks, comprehensive performance score set and standardized equipment dataset, a cost matrix is ​​obtained through multi-objective aggregation. In this step, the estimated time for the equipment to move to the center of the task area is first calculated. According to the distance between the current position of the equipment and the center of the task area, combined with the equipment's moving speed, the distance is divided by the speed to obtain the estimated time. This time reflects the time required for the equipment to reach the task location. Then, the estimated energy consumption cost is calculated. According to the energy consumption per unit area corresponding to the equipment type, it is multiplied by the area of ​​the task area and then divided by the current remaining power of the equipment. The energy consumption cost is directly proportional to the energy consumption per unit area of ​​the equipment type and the task area, and inversely proportional to the current remaining power, reflecting the energy consumption of the equipment when performing the task. Then, a comprehensive equipment performance score is introduced. Since high-performance equipment has an advantage in performing tasks, a negative value is assigned to the comprehensive performance score to reduce the matching cost of high-performance equipment. In addition, to balance equipment load, a load balancing penalty term is set. This term is calculated by multiplying the difference between the current load of the equipment and the average load of the system by a coefficient, which can be obtained through maximum likelihood estimation. If the current load of the equipment is higher than the average load of the system, the penalty term is positive, increasing its matching cost and encouraging the assignment of tasks to equipment with lower current loads to prevent equipment overload. Finally, the above terms are weighted and summed using normalized coefficients and weights to ensure that the dimensions of each term are consistent. The matching cost of each task and each equipment combination is then obtained, and the matching costs of all tasks and equipment are combined into a cost matrix.

[0035] S333. Based on the cost matrix, the matching scheme set is obtained through optimal matching processing. The input in this step is the cost matrix that was constructed earlier. This matrix reflects the matching cost between different combinations of tasks and different devices. The core objective of the algorithm is to find a way to allocate tasks and devices so that the sum of the matching costs of all tasks and corresponding devices is minimized. To achieve this goal, a decision variable is introduced. This variable has only two values, representing whether a task is allocated to a certain device. Several key constraints exist during the matching process. First, each task must be assigned to one and only one device to ensure that all tasks can be executed. Second, the number of tasks assigned to each device cannot exceed its capacity limit. Different devices have different capacity limits due to their own performance and operation characteristics. For example, a drone can usually only execute one task at a time, while a tractor can attach multiple implements to work in parallel and can handle more tasks. Third, if a device cannot execute a certain type of task, based on the judgment of device type and task type, the combination of this type of task and device will not be considered. By multiplying the matching cost in the cost matrix with the decision variables and summing the results, the objective function is constructed. The Kuhn Munkres algorithm or its extensions are used to solve this type of allocation problem. The integer programming problem is solved, and the optimal decision variable assignment is finally obtained. Based on these assignments, the optimal matching scheme set of tasks and devices can be generated. The Kuhn Munkres algorithm is a classic algorithm commonly used to solve the maximum weight matching problem in bipartite graphs, which will not be elaborated here. For example, when irrigating a plot of land, the algorithm considers all smart irrigation devices. However, based on the device type and task type, only irrigation devices that can cover the plot are included as candidates. Considering the capacity limit of each irrigation device, if a device has been assigned too many tasks and is close to its capacity, the algorithm will prioritize assigning new tasks to other devices with lighter loads. By continuously adjusting the values ​​of decision variables, the algorithm minimizes the sum of the matching costs of all tasks and devices. Finally, the algorithm generates the optimal matching scheme set, clearly indicating which smart irrigation device is assigned to which plot of land, and which fertilization device is responsible for fertilizing the area.

[0036] This invention constructs a complete processing flow from equipment performance evaluation to task allocation, which can easily handle emergencies. It avoids subjective bias by objectively assigning weights through the entropy weight method, accurately measures equipment capabilities through comprehensive performance ranking, and comprehensively considers time, energy consumption, performance and load through multi-objective cost aggregation. The improved Hungarian algorithm ensures global optimal matching, significantly improving the rationality of task allocation, resource utilization and system robustness. Even when facing emergencies in agriculture, it greatly improves the efficiency of handling sudden environmental changes.

[0037] S4. Based on the matching scheme set, an adaptive scheduling decision is processed to generate a time-sequential scheduling instruction set. Existing technologies lack overall collaborative optimization, and path planning may not fully consider real-time conflicts. Conflict detection and resolution methods are relatively simple and difficult to dynamically adapt to complex and ever-changing scenarios. For example, multiple drones may be spraying pesticides. If only the conflict of intersecting flight paths is detected, the drift of pesticides in different areas due to factors such as wind direction may be ignored, causing conflicts in the actual operation area and affecting crop growth. To solve the above technical problems, the specific implementation steps are as follows: S41. Based on the matching scheme set, the instruction template is filled to obtain a time-free scheduling instruction set. In this step, the system pre-sets the instruction template, which includes key fields such as device identifier, task identifier, target area, operation type, and parameters. For each pair of task and device combinations in the matching scheme set, the corresponding unique device identifier and unique task identifier are extracted from the matching information. The target work area is determined in combination with the specific requirements of the task, the operation type that the device needs to perform is clarified, and the corresponding operation parameters are set according to the task characteristics and device performance. These extracted and determined information are sequentially filled into the various fields of the instruction template. Each time the filling is completed, a basic scheduling instruction is generated. Finally, all the generated instructions are summarized to form a time-free scheduling instruction set. This process does not involve the arrangement of the instruction execution order, but only focuses on converting the matching result into specific operation instructions that the device can recognize. For example, the matching scheme shows that the irrigation task of a certain farmland is assigned to smart irrigation equipment A, and the fertilization task is assigned to variable fertilizer applicator B. When generating a timeless scheduling instruction set, for the irrigation task of smart irrigation equipment A, the identifier of equipment A and the irrigation task identifier are obtained from the matching information. The target area is determined according to the geographical location of the farmland, the operation type is irrigation, and the parameters are set to irrigation duration and irrigation water volume. For the fertilization task of variable fertilizer applicator B, the identifier of equipment B and the fertilization task identifier are obtained, the target farmland area is determined, the operation type is variable fertilization, and the parameters are set to the fertilization amount for different areas according to the soil nutrient test results. After filling this information into the instruction template, basic scheduling instructions for equipment A and equipment B are generated, and finally, a timeless scheduling instruction set is formed.

[0038] S42. Based on the timeless scheduling instruction set and standardized equipment dataset, a planned path instruction set is obtained through critical path processing. In this step, the current location of the equipment is first determined as the starting point, and the entry point of the task target area is determined as the ending point. Then, the A-pathfinding algorithm is used for path planning. When searching for a path, this algorithm considers two key factors to evaluate the merits of each potential path node. On the one hand, based on the resistance of different terrains to equipment movement recorded in historical data, the actual cost of moving from the starting point to the current node is calculated. On the other hand, the Euclidean distance heuristic method is used to estimate the distance from the current node to the target point, providing directional guidance for path search. Combining these two factors, each node is comprehensively evaluated, prioritizing the exploration of nodes where the sum of the actual movement cost and the estimated distance to the target point is smaller. Through continuous iterative search, such as iterating 100 times, the optimal path from the starting point to the target point is gradually found. Finally, a sequence containing a series of path points is generated for each instruction. All instructions with planned paths are summarized to form the planned path instruction set.

[0039] S43. Perform dynamic scheduling optimization on the planned path instruction set to obtain a time-sequential scheduling instruction set; S431. Based on the planned path instruction set, a conflict relationship set is obtained through conflict cross detection. In this step, for each instruction, an occupation time window is determined for each point on the path according to the equipment departure time and the time required for the equipment to move to each point along the path. The start time of this time window is the equipment departure time plus the time spent to reach the point, and the end time is further determined according to the equipment's working time at the point, so as to obtain the occupation range of each instruction in the spatiotemporal dimension. Then, any two instructions in the instruction set are compared pairwise to detect whether their paths have a spatial intersection. If there is an intersection, the time for the equipment to reach the intersection is calculated according to the speed of the equipment moving along its respective path and the distance to the intersection. Then, the two times are compared. If their time difference is less than the preset safety threshold, it is determined that the two instructions have an intersection point conflict. The conflict type is recorded as intersection point conflict and the corresponding time window. At the same time, it is further checked whether the target operation areas involved in the two instructions have geographical overlap. If there is an overlap and their operation time is found to overlap according to the previously determined time window, it is determined that the two instructions also have an operation area conflict. Similarly, the conflict type is recorded as operation area conflict and the corresponding time window. Finally, all detected conflict relationships are summarized to form a conflict relationship set. For example, two smart harvesters receive cloud-edge collaborative agricultural equipment scheduling instructions. When performing conflict detection and identification based on spatiotemporal windows, for the instructions of the first harvester, the time window occupied by each point on the path is determined according to its departure time and speed along the path. Similarly, the instructions of the second harvester are processed in a similar way. When comparing these two instructions, it is found that their paths have a spatial intersection at a field ridge on the farm. Based on the distance and speed of the two harvesters to the intersection, the arrival time at the intersection is calculated. If the difference between these two time points is less than a preset safety threshold, it is determined that there is a conflict at the intersection point. The conflict type and the corresponding time window are recorded. Further inspection reveals that the target working areas of the two harvesters partially overlap, and their working times also overlap according to the time windows. Therefore, it is also determined that there is a conflict in the working area.

[0040] S432. Dynamically schedule and optimize the conflict relationship set and the planned path instruction set to obtain the optimized instruction start set. In this step, the preset start time of each instruction and the conflict set are defined together as the state to comprehensively depict the current equipment scheduling situation. For this state, the action that can be taken is to delay the start time of a certain instruction. The delay time is selected from zero and several fixed interval values. In order to evaluate the merits of each action, the system sets up a reward mechanism. The reward value consists of three parts: one part is related to the total latency time; the longer the total latency time, the greater the negative impact of this part on the reward. Another part is related to the number of remaining conflicts after taking an action; the more remaining conflicts, the lower the reward. The third part depends on the maximum time to complete all tasks; the longer the total duration, the lower the reward. These three parts are comprehensively considered according to preset weights to obtain the final reward value. After executing an action, the system enters a new state and updates the Q-value corresponding to the current state and action according to a specific formula based on the new state and the reward value. The specific formula refers to the update form of the Bellman equation, which is a commonly used update formula in Q-learning algorithms and will not be elaborated here. This formula comprehensively considers the current Q-value, the reward of the new state, and the Q-value corresponding to the optimal action in the new state, and is adjusted by the learning rate and discount factor. Through continuous iterative exploration, the system gradually learns a set of instruction start times that maximizes the long-term reward, that is, the start time that minimizes the total latency, number of conflicts, and total duration, and finally outputs the optimized instruction start set.

[0041] S433. Based on the optimized instruction start set and the planned path instruction set, a time-sequential scheduling instruction set is generated through scheduling instruction processing. In this step, detailed information of each instruction is first extracted from the planned path instruction set, including key content such as the type of task the device needs to perform and the target work area. At the same time, the start time corresponding to each instruction, which has been determined through reinforcement learning optimization, is obtained from the optimized instruction start set. Then, the detailed information of each instruction, the corresponding start time, and the pre-planned path of the instruction are bound and encapsulated. The start time is used as the time tag for instruction execution, and the path is used as the spatial guide for instruction execution. The three are integrated into a complete scheduling unit. In this way, all instructions are encapsulated and processed according to this rule, and finally a time-sequential scheduling instruction set with a clear execution sequence and path is formed. This instruction set can clearly indicate when each edge device starts a task and what path to use to execute the task. For example, a smart seeder needs to determine the area of ​​land to be sown, the type of seeds, etc., and obtain its planned sowing path. Then, it finds the optimized start time corresponding to the device from the optimized instruction start set, such as 8:00 AM. Then, it binds and encapsulates the task information of the device, the start time of 8:00 AM, and the sowing path to form a complete scheduling instruction. After performing this operation on all devices in the farm, a time-sequential scheduling instruction set is finally generated.

[0042] This invention constructs a complete and sophisticated scheduling system. From pre-plan parsing to generate initial instructions, to path planning and conflict detection, each link is closely connected and logically clear. It can efficiently handle complex task scheduling, taking into account both path planning and conflict resolution. It comprehensively considers intersection point conflicts and work area conflicts, and the detection is more accurate. It avoids the problem that when multiple drones are spraying pesticides, pesticide drift due to wind direction and other factors may be ignored, which may lead to actual conflicts in the work area. Through reinforcement learning and dynamic optimization, it can adapt to different scenarios and improve scheduling flexibility and accuracy.

[0043] S5. Collect the real-time status progress of agricultural equipment, and obtain the scheduling instruction set through flexible adjustment based on the time-sequential scheduling instruction set and the real-time status progress. Send the scheduling instruction set to the control center. Existing technologies are often not timely and comprehensive enough in status monitoring, making it difficult to detect potential problems in time, leading to the accumulation of deviations. Moreover, when making scheduling adjustments, global adjustment methods are often used, which consume a lot of time and resources and greatly interfere with the normal production process. To solve the above problems, the specific implementation steps are as follows: S51. Based on the time-sequential scheduling instruction set and real-time status progress, an execution deviation report is obtained through deviation comparison processing. In this step, the real-time feedback of equipment status and task progress information is compared in detail with the pre-defined time-sequential scheduling instruction set. Specifically, for each device and its corresponding task, the progress that the task should achieve according to the plan in the scheduling instruction is first obtained. Then, combined with the actual progress fed back by the device in real time, the progress difference is obtained by comparing the two. At the same time, the planned time and the actual time are compared to obtain the time difference. Based on these differences and the specific status information currently fed back by the device, the corresponding status is marked for each device task combination according to the established rules, such as normal execution, progress lag, fault, or task interruption. Finally, these information, including device identifier, task identifier, progress difference, time difference, and status mark, are integrated and summarized to form an execution deviation report.

[0044] S52. Based on the execution deviation report, adjust and process the deviation assessment to generate an adjustment decision set. This step first focuses on the status markers of the equipment tasks. If the status markers indicate a equipment malfunction, it means the equipment can no longer perform the task as originally planned, significantly impacting task progress. Alternatively, if the time difference exceeds a pre-set time threshold, it means the task progress has seriously deviated from the plan, potentially affecting the overall scheduling. When either of these two situations occurs, the system determines the impact is significant and makes a partial rescheduling decision, i.e., rescheduling the problematic equipment and its related tasks. If the status markers show normal operation, and all indicators such as the time difference are within acceptable limits... Within reasonable limits, this indicates that the task is performing well and requires no adjustment. The system will decide to continue execution according to the original plan. However, in extreme cases such as widespread equipment failure or severe delays in the progress of many tasks, where local rescheduling cannot effectively solve the problem, the system will determine that the impact is extremely significant and make a global rescheduling decision. This involves replanning the entire scheduling plan and ultimately forming an adjustment decision set that includes specific decisions such as continuation, local rescheduling, or global rescheduling. The entire process can be implemented using a rule-based threshold decision tree. A rule-based threshold decision tree is a commonly used tree-like logical structure model that combines preset rules and key thresholds to assist in decision-making, which will not be elaborated on here.

[0045] S53. Based on the adjustment decision set, execution deviation report, standardized equipment dataset, and dynamic priority tasks, an updated scheduling instruction set is obtained through incremental scheduling adjustment. In this step, when the system receives an adjustment decision indicating that local rescheduling is required, it will combine the execution deviation report to accurately locate the equipment and tasks with deviations. Based on the current task queue status and standardized equipment status information, the affected tasks are first re-prioritized and queued, allowing the tasks to wait for execution in a new reasonable order. Then, for the re-queued tasks and affected equipment, the most suitable equipment is re-matched to these tasks according to the established equipment and task matching rules and models. After the matching is completed, new local scheduling instructions are generated based on the new task-equipment combination, the specific requirements of the tasks, and the status of the equipment. These new instructions replace the original instructions of the affected parts, and finally, they are integrated to form an updated scheduling instruction set, ensuring that the scheduling in the local area can run reasonably and efficiently again.

[0046] This invention enables real-time monitoring and comparison of operational status, allowing for timely capture of subtle deviations in equipment and task execution. This provides accurate data for subsequent decision-making, enabling rapid and reasonable determination of adjustment directions based on deviations. It avoids blind adjustments, reduces interference with overall scheduling, improves scheduling efficiency and flexibility, and ensures efficient and stable production operation.

[0047] Example 2: Because existing conflict detection and resolution methods are relatively simple and difficult to adapt dynamically to complex and ever-changing scenarios, and cannot fully consider the technical issues of real-time conflicts, please refer to [the relevant documentation / reference]. Figure 2 The diagram shown is a structural block diagram of the agricultural equipment information scheduling and management system with cloud-edge collaboration provided in this embodiment. The system includes a standardization module, a dynamic priority module, a matching scheme module, an adaptive scheduling module, and a flexible adjustment module. The standardization module is used to collect raw environmental and equipment data from agricultural equipment, and preset expected data to generate a standard dataset through classification preprocessing. The dynamic prioritization module is used to perform dynamic prioritization based on a standard dataset and obtain dynamic priority tasks. The matching scheme module is used to perform collaborative matching scheme processing on the standardized equipment dataset and the dynamic priority task queue to obtain a set of matching schemes. The adaptive scheduling module is used to generate a time-sequential scheduling instruction set based on the matching scheme set through adaptive scheduling decision processing. The flexible adjustment module is used to collect the real-time status and progress of agricultural equipment. Based on the time-series scheduling instruction set and the real-time status and progress, it obtains a scheduling instruction set through flexible adjustment and sends the scheduling instruction set to the control center.

[0048] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code, including but not limited to disk storage, CD-ROM, optical storage, etc.

[0049] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A cloud-edge collaborative agricultural equipment information scheduling and management method, characterized in that, The method involves the following steps: collecting raw environmental and equipment data from agricultural equipment, pre-setting expected data, and generating a standard dataset through classification preprocessing; wherein the standard dataset includes at least: a standardized environmental dataset, a standardized equipment dataset, and a standardized expected dataset; Dynamic priority processing is performed based on standard datasets to obtain dynamically prioritized tasks. Perform collaborative matching scheme processing on the standardized equipment dataset and the dynamic priority task queue to obtain a set of matching schemes; Based on the matching scheme set, a time-sequential scheduling instruction set is generated through adaptive scheduling decision processing. The system collects the real-time status and progress of agricultural equipment, and obtains a scheduling instruction set through flexible adjustments based on the time-series scheduling instruction set and the real-time status and progress. The scheduling instruction set is then sent to the control center.

2. The agricultural equipment information scheduling and management method using cloud-edge collaboration as described in claim 1, characterized in that, A standard dataset is generated through classification preprocessing, including: obtaining a standardized environmental dataset by alignment and normalization processing based on the original environmental data; The device data is processed by state mapping to obtain a standardized device dataset; A standardized expectation dataset is obtained by using natural language processing based on the expected data.

3. The agricultural equipment information scheduling and management method using cloud-edge collaboration as described in claim 1, characterized in that, Dynamic prioritization based on standard datasets includes: obtaining a set of granular expectation tasks by decomposing a standardized expectation dataset into a spatial grid. Based on the particle expectation task set and the standardized environment dataset, a particle urgency assessment set is obtained through multidimensional fuzzy evaluation. By performing weighted dynamic prioritization on the particle urgency assessment set and the particle expected task set, dynamic priority tasks are obtained.

4. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 3, characterized in that, The urgency assessment set and the expected task set of particles are subjected to weighted dynamic prioritization, including: weighted scoring of the urgency assessment set and the expected task set of particles to obtain an initial priority task table; Based on the initial priority task table and the standardized device dataset, a dynamic priority score set is obtained through dynamic priority processing. Dynamic priority tasks are obtained by prioritizing the dynamic priority score set and the granular expected task set.

5. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 1, characterized in that, The standardized equipment dataset and the dynamic priority task queue are processed by a collaborative matching scheme, including: extracting performance index matrix from the standardized equipment dataset through index quantification; Based on the performance index matrix, the index weight vector is obtained by processing it using the entropy weight method. The performance evaluation and scheme generation operations are performed on the indicator weight vector and performance indicator matrix to obtain a set of matching schemes.

6. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 5, characterized in that, The process of evaluating the performance of the indicator weight vector and the performance indicator matrix and generating a solution includes: processing the performance of the indicator weight vector and the performance indicator matrix to obtain a set of comprehensive performance scores; The cost matrix is ​​obtained by multi-objective aggregation based on dynamic priority tasks, comprehensive performance score set, and standardized equipment dataset; Based on the cost matrix, a set of matching schemes is obtained through optimal matching.

7. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 1, characterized in that, Based on the matching scheme set, adaptive scheduling decision processing is performed, including: processing the matching scheme set by filling instruction templates to obtain a time-free scheduling instruction set; Based on the timeless scheduling instruction set and standardized equipment dataset, a planned path instruction set is obtained through critical path processing; The planned path instruction set is dynamically scheduled and optimized to obtain a time-sequential scheduling instruction set.

8. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 7, characterized in that, Dynamic scheduling optimization of the planned path instruction set includes: obtaining a set of conflict relationships based on the planned path instruction set through conflict cross detection; Dynamic scheduling optimization is performed on the conflict relationship set and the planned path instruction set to obtain an optimized instruction start set; Based on the optimized instruction start set and the planned path instruction set, a time-sequential scheduling instruction set is generated through scheduling instruction processing.

9. The agricultural equipment information scheduling and management method using cloud-edge collaboration according to claim 1, characterized in that, Based on the time-series scheduling instruction set and real-time status progress, flexible adjustments are made, including: based on the time-series scheduling instruction set and real-time status progress, deviation processing is performed to obtain an execution deviation report; Based on the performance deviation report, the deviation assessment and adjustment process is carried out to generate an adjustment decision set; Based on the adjustment decision set, execution deviation report, standardized equipment dataset, and dynamic priority tasks, an updated scheduling instruction set is obtained through incremental scheduling adjustments.

10. A system applied to the cloud-edge collaborative agricultural equipment information scheduling and management method described in any one of claims 1-9, characterized in that, The system includes: The standardization module is used to collect raw environmental and equipment data from agricultural equipment, and preset expected data to generate a standard dataset through classification preprocessing. The dynamic prioritization module is used to perform dynamic prioritization based on a standard dataset and obtain dynamic priority tasks. The matching scheme module is used to perform collaborative matching scheme processing on the standardized equipment dataset and the dynamic priority task queue to obtain a set of matching schemes. The adaptive scheduling module is used to generate a time-sequential scheduling instruction set based on the matching scheme set through adaptive scheduling decision processing. The flexible adjustment module is used to collect the real-time status and progress of agricultural equipment. Based on the time-series scheduling instruction set and the real-time status and progress, it obtains a scheduling instruction set through flexible adjustment and sends the scheduling instruction set to the control center.

Images