Agricultural robot intelligent scheduling method based on big data and space-time coupling

Abstract

The invention discloses an agricultural robot intelligent scheduling method based on big data and space-time coupling, and the method comprises the following steps: collecting agricultural data and agricultural robot data, and generating an agricultural state data set; generating an agricultural operation irreversible constraint set; mapping the agricultural operation irreversible constraint set into an irreversible barrier map, and generating a barrier map constraint representation; generating a candidate scheduling scheme set; mapping each candidate scheduling scheme into a scheme spatio-temporal trajectory, and generating a barrier avoidance verification result; when at least one barrier node is triggered, a shortest counter-example trajectory is generated, a minimum contradiction substructure is extracted, a non-executable certificate is generated and bound, and when any barrier node is not triggered, an executable certificate is generated and bound, and an intelligent scheduling result of the agricultural robot is formed. According to the method, by constructing the irreversible barrier map and verifying the spatial-temporal trajectory of the scheduling scheme, pre-judgment screening of the agricultural robot scheduling scheme is achieved, and the method has the advantages of being high in execution reliability and low in rework risk.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural robot intelligent scheduling, and in particular to an agricultural robot intelligent scheduling method based on big data and space-time coupling. BACKGROUND

[0002] In the field of agricultural robot scheduling, the existing technology usually matches and schedules agricultural operation tasks and agricultural robots based on operation rules, priority strategies or heuristic optimization algorithms. The related methods mainly take operation time window, operation type or robot capability as the main constraint condition, and complete the scheduling decision in combination with real-time or quasi-real-time data. Some schemes introduce meteorological information or field information as auxiliary judgment basis to improve the rationality and execution efficiency of scheduling. Such methods focus on independent judgment of a single constraint or a small number of constraints in structure, and usually need to be modified by manual or running feedback after the scheduling result is generated.

[0003] There are various strong space-time constraints in agricultural operation, such as phenology window, meteorological window, soil carrying capacity and agronomic operation sequence, which are irreversible. The above constraints are coupled and transmitted in time and space, and the existing technology generally lacks the ability to model and transmit the irreversible constraints in the time-space dimension as a whole. It is difficult to identify the essentially unexecutable scheduling scheme before scheduling is issued, which leads to the forced interruption or rework of the scheduling scheme after triggering the irreversible condition in the execution process, thereby causing operation delay, resource waste and missed operation window problems, which restricts the reliability and intelligent level of the agricultural robot scheduling system. SUMMARY

[0004] One object of the present application is to provide an agricultural robot intelligent scheduling method based on big data and space-time coupling. The present application realizes the pre-judgment and exclusion of agricultural robot scheduling scheme by constructing an irreversible barrier graph and verifying the space-time trajectory of the scheduling scheme, which has the advantages of high execution reliability and low rework risk.

[0005] According to an embodiment of the present application, an agricultural robot intelligent scheduling method based on big data and space-time coupling comprises the following steps: Collect agricultural data and agricultural robot data, and preprocess to generate an agricultural state data set; Generate an agricultural operation irreversible constraint set based on the agricultural state data set, which includes phenological irreversible constraint, meteorological irreversible constraint, soil carrying capacity irreversible constraint and agronomic sequence irreversible constraint. Configure the corresponding trigger condition and action time window for each irreversible constraint; mapping the set of irreversible constraints of agricultural operations into an irreversible barrier graph, mapping the trigger points of irreversible constraints into barrier nodes, mapping the trigger propagation relationship between irreversible constraints into barrier edges, generating a barrier graph constraint representation based on the irreversible barrier graph; generating a set of candidate scheduling schemes based on the agricultural state dataset; mapping each candidate scheduling scheme of the set of candidate scheduling schemes into a corresponding scheme spatiotemporal trajectory, matching the scheme spatiotemporal trajectory with the irreversible barrier graph constraint representation, and generating a corresponding barrier avoidance verification result; when the barrier avoidance verification result represents that the candidate scheduling scheme triggers at least one barrier node, generating a shortest counterexample trajectory and extracting a minimal contradictory substructure, generating and binding an unexecutability certificate, and excluding the corresponding candidate scheduling scheme, when no barrier node of the irreversible barrier graph is triggered, generating and binding an executability certificate, and forming an intelligent scheduling result of the agricultural robot.

[0006] Optionally, the agricultural data specifically includes field identification data, field spatial position data, operation type data, operation time window data, crop phenological stage data, weather state data, soil bearing state data, and agronomic operation sequence data, the agricultural robot data specifically includes robot identification data, spatial position information data, operation capability parameter data, execution time data, operation state data, and travel path data, and the preprocessing specifically includes data cleaning, time alignment, and spatial mapping.

[0007] Optionally, the generation of the set of irreversible constraints of agricultural operations specifically includes: extracting operation type information, field identification information, operation time information, crop phenological stage information, weather state information, soil bearing state information, and agronomic operation sequence information from the agricultural state dataset, and aggregating the information according to the same operation object to form an operation state record; determining whether the corresponding operation time information satisfies a preset allowed operation time window condition of the crop phenological stage according to the crop phenological stage information in the operation state record, and generating a phenological irreversible constraint when the operation time does not satisfy the allowed operation time window condition, and configuring a corresponding trigger condition and action time window for the phenological irreversible constraint; determining whether the current weather state satisfies a preset weather execution condition of the corresponding operation type according to the weather state information in the operation state record, and generating a weather irreversible constraint when the weather state does not satisfy the weather execution condition, and configuring a corresponding trigger condition and action time window for the weather irreversible constraint; According to the soil bearing state information in the operation state record, it is determined whether the soil bearing state meets the preset bearing condition of the corresponding operation type performed on the corresponding plot. When the soil bearing state does not meet the preset bearing condition, a soil bearing irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the soil bearing irreversible constraint; According to the agricultural operation sequence information and operation time information in the operation state record, it is determined whether the operation execution sequence violates the preset agricultural operation sequence condition. When the operation execution sequence violates the agricultural operation sequence condition, an agricultural sequence irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the agricultural sequence irreversible constraint; The phenology irreversible constraint, the meteorological irreversible constraint, the soil bearing irreversible constraint and the agricultural sequence irreversible constraint are collected to form an agricultural operation irreversible constraint set.

[0008] Optionally, the generation of the barrier graph constraint representation specifically includes: The agricultural operation irreversible constraint set is read to obtain the trigger condition and the action time window corresponding to each agricultural operation irreversible constraint, and each agricultural operation irreversible constraint is taken as a mapping object; Each agricultural operation irreversible constraint is mapped into a barrier node, and the trigger condition and the action time window corresponding to the agricultural operation irreversible constraint are associated in the barrier node to form a barrier node set; In the barrier node set, it is determined whether there is a trigger propagation relationship between the agricultural operation irreversible constraints according to the trigger sequence and the overlapping relationship between the action time windows; When it is determined that there is a trigger propagation relationship, a directed connection is established between the corresponding barrier nodes, the directed connection is determined as a barrier edge to form a barrier edge set; The barrier node set and the barrier edge set are combined to construct an irreversible barrier graph; Based on the irreversible barrier graph, the trigger condition and the action time window associated with the barrier node and the trigger propagation relationship represented by the barrier edge are uniformly arranged to generate a barrier graph constraint representation.

[0009] Optionally, the generation of the candidate scheduling scheme set specifically includes: Data related to agricultural operation and data related to agricultural robots are retrieved from the agricultural state data set; The data related to agricultural operation is subjected to state determination, the agricultural operation in the to-be-executed state is selected, and the selected agricultural operation is collected to form an operation task set; The data related to agricultural robots is subjected to state determination, the agricultural robots in the schedulable state are selected, and the selected agricultural robots are collected to form an agricultural robot set; corresponding relationship set between the work tasks and the agricultural robots is established one by one between the work task set and the agricultural robot set; corresponding relationship set is time-allocated to generate a work time allocation result corresponding to the corresponding relationship set; The work task set, the agricultural robot set, the corresponding relationship set, and the work time allocation result are integrated to form a candidate scheduling scheme; The plurality of candidate scheduling schemes formed are collected to generate a candidate scheduling scheme set.

[0010] Optionally, the generation of the barrier avoidance verification result specifically includes: Based on each candidate scheduling scheme of the candidate scheduling scheme set, a combination of each agricultural robot performing a corresponding work task within a corresponding work time allocation interval is arranged in time sequence to generate a scheme space-time track corresponding to the candidate scheduling scheme; The irreversible barrier graph constraint representation is read to obtain the trigger condition and the action time window corresponding to each barrier node contained therein; In each work time allocation interval corresponding to the scheme space-time track, for a work time allocation interval that overlaps with the action time window of a barrier node, it is judged one by one whether the corresponding scheme space-time track satisfies the trigger condition of the barrier node; When the scheme space-time track satisfies the trigger condition of at least one barrier node in any work time allocation interval, the barrier avoidance verification result corresponding to the candidate scheduling scheme is triggered; When the corresponding scheme space-time track does not satisfy the trigger condition of any barrier node in all work time allocation intervals, the barrier avoidance verification result corresponding to the candidate scheduling scheme is not triggered.

[0011] Optionally, the generation of the agricultural robot intelligent scheduling result specifically includes: The barrier avoidance verification result corresponding to each candidate scheduling scheme in the candidate scheduling scheme set is read; When the barrier avoidance verification result is triggered, the work time allocation interval of the barrier node that is first triggered is located in the scheme space-time track corresponding to the candidate scheduling scheme; The work time allocation interval that is first triggered is taken as a termination position, and the continuous work time allocation intervals related to the barrier node in the scheme space-time track are traced back to form a shortest counterexample track; In the shortest counterexample track, the work task, the agricultural robot, and the corresponding relationship involved in the triggering of the barrier node are selected, and the remaining work tasks and corresponding relationships that are not involved in the triggering of the barrier node are eliminated to extract a minimum contradictory substructure. corresponding candidate scheduling scheme, generate and bind an unexecutable certificate, and exclude the corresponding candidate scheduling scheme; When the barrier avoidance verification result is not triggered, an executable certificate corresponding to the candidate scheduling scheme is generated and bound, and an intelligent scheduling result of the agricultural robot is formed.

[0012] The beneficial effects of the present application are: The present application breaks through the technical limitations of the prior art of only judging single constraints or local rules independently by uniformly modeling agricultural data and agricultural robot data to form an agricultural state data set, and introducing an agricultural operation irreversible constraint set and an irreversible barrier graph modeling method on this basis. The agricultural operation constraints with irreversible characteristics such as phenology, meteorology, soil bearing and agronomic sequence are structured and expressed by barrier nodes with action time windows and trigger conditions and their trigger propagation relationships, so that the present application can verify the overall executability of the scheme in the time and space dimensions before the scheduling scheme is issued, effectively avoiding the problem that the scheme is forced to interrupt in the execution process due to the neglect of irreversible constraint coupling relationship.

[0013] When the scheduling scheme triggers an irreversible barrier node, the present application generates the shortest counterexample trajectory and the smallest contradictory substructure to accurately locate the key time segment, operation task and agricultural robot that cause the scheme to be unexecutable, and generates an unexecutable certificate accordingly, and automatically excludes the unexecutable scheme from the candidate scheduling scheme set. For the scheme that does not trigger any irreversible barrier node, an executable certificate is generated and the final intelligent scheduling result of the agricultural robot is formed. This technical effect makes the scheduling system not only output executable scheduling schemes, but also clearly give the structural reasons for unexecutable schemes, significantly reduces the scheduling rework rate and operation risk, improves the reliability, explainability and overall operation efficiency of the agricultural robot scheduling result, and is especially suitable for complex agricultural operation scenarios with strong space-time coupling and irreversible constraint characteristics. BRIEF DESCRIPTION OF DRAWINGS

[0014] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application together with the embodiments thereof, and explain the present application, and do not constitute a limitation of the present application. In the drawings:

[0015] Fig. 1 A flowchart of an intelligent scheduling method of an agricultural robot based on big data and space-time coupling is proposed for the present application; Fig. 2An irreversible barrier graph construction schematic diagram of an agricultural robot intelligent scheduling method based on big data and space-time coupling is provided for the present application. Fig. 3 A scheme space-time trajectory and barrier avoidance verification schematic diagram of an agricultural robot intelligent scheduling method based on big data and space-time coupling is provided for the present application. DETAILED DESCRIPTION

[0016] The present application will now be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, and only illustrate the basic structure of the present application in a schematic manner, and therefore only show the components related to the present application.

[0017] REFERENCE Figs. 1-3 An agricultural robot intelligent scheduling method based on big data and space-time coupling, comprising the following steps: Collecting agricultural data and agricultural robot data, and preprocessing to generate an agricultural state data set; Generating an agricultural operation irreversible constraint set based on the agricultural state data set, the agricultural operation irreversible constraint set including phenological irreversible constraints, meteorological irreversible constraints, soil bearing irreversible constraints and agronomic sequence irreversible constraints, and configuring corresponding trigger conditions and action time windows for each irreversible constraint; Mapping the agricultural operation irreversible constraint set to an irreversible barrier graph, mapping the trigger points of the irreversible constraints to barrier nodes, mapping the trigger propagation relationship between the irreversible constraints to barrier edges, and generating a barrier graph constraint representation based on the irreversible barrier graph; Generating a candidate scheduling scheme set based on the agricultural state data set; Mapping each candidate scheduling scheme of the candidate scheduling scheme set to a corresponding scheme space-time trajectory, matching the scheme space-time trajectory with the irreversible barrier graph constraint representation, and generating a corresponding barrier avoidance verification result; When the barrier avoidance verification result represents that the candidate scheduling scheme triggers at least one barrier node, generating a shortest counterexample trajectory and extracting a minimum contradiction substructure, generating and binding an unexecutability certificate, and excluding the corresponding candidate scheduling scheme, when no barrier node of the irreversible barrier graph is triggered, generating and binding an executability certificate, and forming an agricultural robot intelligent scheduling result.

[0018] In this embodiment, the agricultural data specifically includes plot identification data, plot spatial position data, operation type data, operation time window data, crop phenological stage data, meteorological state data, soil bearing state data and agronomic operation sequence data, the agricultural robot data specifically includes robot identification data, spatial position information data, operation capability parameter data, execution time data, operation state data and travel path data, and the preprocessing specifically includes data cleaning, time alignment and spatial mapping.

[0019] In this embodiment, the generation of the set of irreversible constraints of agricultural operations specifically includes: Extracting operation type information, plot identification information, operation time information, crop phenology stage information, weather state information, soil carrying state information, and agronomic operation sequence information from the agricultural state data set, aggregating the information according to the same operation object to form an operation state record; According to the crop phenology stage information in the operation state record, it is determined whether the operation time information meets the preset allowed operation time window condition of the phenology stage. When the operation time does not meet the allowed operation time window condition, a phenology irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the phenology irreversible constraint; The generation of the phenology irreversible constraint specifically includes: Based on the crop phenology stage information in the operation state record, the preset allowed operation time window under the corresponding phenology stage is determined. The operation time information in the operation state record is read to determine whether the operation time information falls within the allowed operation time window. When it is determined that the operation time information does not fall within the allowed operation time window, the time range in the operation time information that does not fall within the allowed operation time window is determined. Based on the time range that does not fall within the allowed operation time window, a phenology irreversible constraint is generated, and the time range is determined as the action time window of the phenology irreversible constraint. The trigger condition of “operation time information not falling within the allowed operation time window of the corresponding phenology stage” is determined for the phenology irreversible constraint, the trigger condition is associated with the action time window, and the configuration of the phenology irreversible constraint is completed. According to the weather state information in the operation state record, it is determined whether the current weather state meets the preset weather execution condition of the corresponding operation type. When the weather state does not meet the weather execution condition, a weather irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the weather irreversible constraint; The generation of the weather irreversible constraint specifically includes: Based on the weather state information in the operation state record, the weather state under the corresponding operation time information is obtained. According to the operation type information in the operation state record, the preset weather execution condition of the operation type is determined. It is determined whether the weather state meets the weather execution condition. When it is determined that the weather state does not meet the weather execution condition, the time range in which the weather state does not meet the weather execution condition is determined. Based on the time range, a weather irreversible constraint is generated, and the time range is determined as the action time window of the weather irreversible constraint. The trigger condition of “weather state not meeting the weather execution condition of the corresponding operation type” is determined for the weather irreversible constraint, the trigger condition is associated with the action time window, and the configuration of the weather irreversible constraint is completed. According to the soil bearing state information in the operation state record, it is determined whether the soil bearing state meets the preset bearing condition of the corresponding operation type performed on the corresponding plot, and when the soil bearing state does not meet the preset bearing condition, a soil bearing irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the soil bearing irreversible constraint; The generation of the soil bearing irreversible constraint specifically includes: According to the soil bearing state information in the operation state record, the soil bearing state of the corresponding plot under the corresponding operation time information is obtained, the bearing condition preset for the operation type performed on the plot is determined according to the plot identification information and the operation type information in the operation state record, it is determined whether the soil bearing state meets the preset bearing condition, when it is determined that the soil bearing state does not meet the preset bearing condition, the time range in which the soil bearing state does not meet the preset bearing condition is determined, the soil bearing irreversible constraint is generated based on the time range, the time range is determined as the action time window of the soil bearing irreversible constraint, and the trigger condition of the soil bearing irreversible constraint is determined as "the soil bearing state does not meet the preset bearing condition of the corresponding operation type performed on the corresponding plot", the trigger condition is associated with the action time window, and the configuration of the soil bearing irreversible constraint is completed. According to the agricultural operation sequence information and the operation time information in the operation state record, it is determined whether the operation execution sequence violates the preset agricultural operation sequence condition, when the operation execution sequence violates the agricultural operation sequence condition, an agricultural sequence irreversible constraint is generated, and the corresponding trigger condition and action time window are configured for the agricultural sequence irreversible constraint. The generation of the agricultural sequence irreversible constraint specifically includes: According to the agricultural operation sequence information and the operation time information in the operation state record, the agricultural operation sequence corresponding to each operation in the operation state record is determined, the agricultural operation sequence condition preset for the corresponding operation type is determined according to the operation type information in the operation state record, it is determined whether the operation execution sequence reflected in the operation state record violates the agricultural operation sequence condition according to the agricultural operation sequence and the agricultural operation sequence condition, when it is determined that the operation execution sequence violates the agricultural operation sequence condition, the operation time information corresponding to the operation violating the agricultural operation sequence condition is determined, the agricultural sequence irreversible constraint is generated based on the operation time information, the time range corresponding to the operation time information is determined as the action time window of the agricultural sequence irreversible constraint, the trigger condition of the agricultural sequence irreversible constraint is determined as "the operation execution sequence violates the preset agricultural operation sequence condition", the trigger condition is associated with the action time window, and the configuration of the agricultural sequence irreversible constraint is completed. The phenology irreversible constraint, the meteorological irreversible constraint, the soil bearing irreversible constraint and the agricultural sequence irreversible constraint are collected to form an agricultural operation irreversible constraint set.

[0020] In this embodiment, the generation of the barrier graph constraint representation specifically includes: reading the set of agricultural operation irreversible constraint, obtaining the trigger condition and the action time window corresponding to each agricultural operation irreversible constraint, and taking each agricultural operation irreversible constraint as a mapping object; mapping each agricultural operation irreversible constraint into a barrier node, associating the trigger condition and the action time window corresponding to the agricultural operation irreversible constraint in the barrier node, and forming a set of barrier nodes; In the set of barrier nodes, determining whether there is a trigger propagation relationship between the agricultural operation irreversible constraints according to the trigger sequence and the overlapping relationship between the action time windows; The determination of the trigger propagation relationship specifically includes: In the set of barrier nodes, for two barrier nodes, reading the trigger condition and the action time window of the agricultural operation irreversible constraint corresponding to each barrier node respectively; determining whether the action time windows of the two agricultural operation irreversible constraints overlap based on the action time windows; when it is determined that the action time windows of the two agricultural operation irreversible constraints overlap, determining whether there is a trigger sequence between the two agricultural operation irreversible constraints according to the corresponding trigger conditions; when it is determined that there is a trigger sequence, and the agricultural operation irreversible constraint triggered first meets the trigger condition of the agricultural operation irreversible constraint triggered later within its action time window, it is determined that there is a trigger propagation relationship between the two agricultural operation irreversible constraints; when it is determined that the action time windows of the two agricultural operation irreversible constraints do not overlap, or there is no trigger sequence, it is determined that there is no trigger propagation relationship between the two agricultural operation irreversible constraints; When it is determined that there is a trigger propagation relationship, a directed connection is established between the corresponding barrier nodes, the directed connection is determined as a barrier edge, and a set of barrier edges is formed; The generation of the set of barrier edges specifically includes: reading the set of barrier nodes and the result obtained based on the determination of the trigger propagation relationship, obtaining the barrier node pair determined to have a trigger propagation relationship; for each barrier node pair determined to have a trigger propagation relationship, determining the barrier node triggered first and the barrier node triggered later in the barrier node pair; establishing a directed connection between the barrier node triggered first and the barrier node triggered later, the direction of the directed connection being from the barrier node triggered first to the barrier node triggered later; determining the directed connection as a barrier edge, and associating the barrier edge with the corresponding barrier node pair; collecting all barrier edges generated based on the trigger propagation relationship, and forming a set of barrier edges; combining the set of barrier nodes and the set of barrier edges to construct an irreversible barrier graph; Based on the irreversible barrier graph, the trigger conditions and action time windows associated with the barrier nodes and the trigger propagation relationships represented by the barrier edges are uniformly sorted to generate a barrier graph constraint representation; The generation of the barrier graph constraint representation specifically includes: Read the irreversible barrier graph to obtain the barrier node set and the barrier edge set contained in the irreversible barrier graph; for each barrier node in the barrier node set, collect the trigger conditions and the action time windows corresponding to the agricultural operation irreversible constraints associated with the barrier node; for each barrier edge in the barrier edge set, collect the trigger propagation relationships between the barrier nodes represented by the barrier edge; uniformly sort the trigger conditions and the action time windows associated with the barrier nodes, and uniformly sort the trigger propagation relationships represented by the barrier edges; combine the sorted barrier node information and the sorted barrier edge information to generate a barrier graph constraint representation.

[0021] In this embodiment, the generation of the candidate scheduling scheme set specifically includes: Retrieve data related to agricultural operations and data related to agricultural robots from the agricultural state data set; Perform state determination on the data related to agricultural operations, select agricultural operations in an execution-ready state, and collect the selected agricultural operations to form a set of task assignments; Perform state determination on the data related to agricultural robots, select agricultural robots in a schedulable state, and collect the selected agricultural robots to form a set of agricultural robots; Establish a one-to-one correspondence between the task assignments and the agricultural robots in the set of task assignments and the set of agricultural robots, forming a set of corresponding relationships between the task assignments and the agricultural robots; Combine the task time windows corresponding to the task assignments, and perform time allocation on the task assignments in the set of corresponding relationships to generate a task time allocation result corresponding to the set of corresponding relationships; Integrate the set of task assignments, the set of agricultural robots, the set of corresponding relationships, and the task time allocation result to form a candidate scheduling scheme; Collect the multiple candidate scheduling schemes formed to generate a set of candidate scheduling schemes.

[0022] In this embodiment, the generation of the barrier avoidance verification result specifically includes: Based on each candidate scheduling scheme in the set of candidate scheduling schemes, arrange the combination of each agricultural robot performing the corresponding task assignment within the corresponding task time allocation interval in chronological order to generate a scheme space-time trajectory corresponding to the candidate scheduling scheme; Read the irreversible barrier graph constraint representation to obtain the trigger conditions and the action time windows corresponding to each barrier node; In each of the operation time allocation intervals corresponding to the scheme space-time trajectory, for each operation time allocation interval that overlaps with the action time window of the barrier node, it is determined whether the corresponding scheme space-time trajectory satisfies the trigger condition of the barrier node; The determination of the trigger condition of the barrier node specifically includes: In the operation time allocation interval, the agricultural operation state information and the agricultural robot state information reflected in the scheme space-time trajectory corresponding to the operation time allocation interval are read; the trigger condition corresponding to the agricultural operation irreversible constraint associated with the barrier node is read; the determination elements involved in the trigger condition are matched one by one based on the scheme space-time trajectory in the operation time allocation interval; when it is determined that the scheme space-time trajectory satisfies all the limited elements in the trigger condition in the operation time allocation interval, it is determined that the scheme space-time trajectory satisfies the trigger condition of the barrier node in the operation time allocation interval; when it is determined that the scheme space-time trajectory does not satisfy at least one limited element in the trigger condition in the operation time allocation interval, it is determined that the scheme space-time trajectory does not satisfy the trigger condition of the barrier node in the operation time allocation interval; When the scheme space-time trajectory satisfies the trigger condition of at least one barrier node in any operation time allocation interval, the barrier avoidance verification result corresponding to the candidate scheduling scheme is generated as triggered; The generation of the barrier avoidance verification result as triggered specifically includes: In each of the operation time allocation intervals corresponding to the scheme space-time trajectory, an operation time allocation interval that overlaps in time with the action time window of at least one barrier node is selected; for each selected operation time allocation interval, the corresponding scheme space-time trajectory in the operation time allocation interval and the trigger condition associated with the barrier node are read; in the operation time allocation interval, it is determined whether the scheme space-time trajectory satisfies the trigger condition corresponding to the barrier node; when it is determined that the scheme space-time trajectory satisfies the trigger condition corresponding to at least one barrier node in any operation time allocation interval, it is determined that the barrier avoidance verification result corresponding to the candidate scheduling scheme is triggered; When the scheme space-time trajectory does not satisfy the trigger condition corresponding to any barrier node in all operation time allocation intervals, the barrier avoidance verification result corresponding to the candidate scheduling scheme is generated as not triggered.

[0023] In this embodiment, the generation of the agricultural robot intelligent scheduling result specifically includes: The barrier avoidance verification results corresponding to each candidate scheduling scheme in the candidate scheduling scheme set are read; When the barrier avoidance verification result is triggered, the operation time allocation interval in which the barrier node is first triggered is located in the scheme space-time trajectory corresponding to the candidate scheduling scheme; The positioning of the operation time allocation interval in which the barrier node is first triggered specifically includes: In the scheme space-time track corresponding to the candidate scheduling scheme, the job time allocation intervals are sorted in time sequence according to the job time allocation intervals; for each sorted job time allocation interval, the scheme space-time track in the corresponding job time allocation interval and the barrier node action time window that has time overlap with the job time allocation interval are read in sequence; in the job time allocation interval, it is determined whether the scheme space-time track meets the trigger condition of the corresponding barrier node; when it is determined that the scheme space-time track in a job time allocation interval meets the trigger condition of at least one barrier node, the job time allocation interval is determined as the job time allocation interval of the first triggered barrier node; With the first triggered job time allocation interval as the termination position, the continuous job time allocation intervals related to the barrier node in the scheme space-time track are traced back, and the shortest counterexample track is intercepted; The generation of the shortest counterexample track specifically includes: The job time allocation interval of the first triggered barrier node is read, and the job time allocation interval is determined as the backtracking starting position of the scheme space-time track; before the backtracking starting position, the job time allocation intervals in the scheme space-time track are traced back one by one in time sequence according to the job time allocation intervals; in the backtracking process, for each backtracked job time allocation interval, it is determined whether the scheme space-time track in the job time allocation interval is related to the trigger condition of the barrier node; when it is determined that the scheme space-time track in a job time allocation interval is not related to the trigger condition of the barrier node, the job time allocation interval is determined as the backtracking termination position; the continuous job time allocation intervals between the backtracking termination position and the job time allocation interval of the first triggered barrier node are intercepted to form the shortest counterexample track; In the shortest counterexample track, the job tasks participating in the trigger of the barrier node, the agricultural robots and the corresponding relationships are selected, the remaining job tasks and the corresponding relationships that do not participate in the trigger of the barrier node are eliminated, and the minimum contradictory substructure is extracted; The generation of the minimum contradictory substructure specifically includes: read the shortest counterexample trajectory, obtain the job time allocation interval contained in the shortest counterexample trajectory and the corresponding scheme space-time trajectory; in the shortest counterexample trajectory, based on the job time allocation interval that has been determined to satisfy the barrier node trigger condition, the corresponding job task, agricultural robot and the corresponding relationship between the job task and the agricultural robot in the job time allocation interval are selected; in the shortest counterexample trajectory, for the job time allocation interval other than the job time allocation interval that has been determined, it is determined whether the job task, the agricultural robot and the corresponding relationship in the job time allocation interval satisfy the trigger condition of the barrier node; when it is determined that the job task, the agricultural robot and the corresponding relationship do not satisfy the trigger condition of the barrier node, the job task, the agricultural robot and the corresponding relationship are removed from the shortest counterexample trajectory; the job task, the agricultural robot and the corresponding relationship remaining after being removed are combined to form a minimum contradiction substructure. The shortest counterexample trajectory and the minimum contradiction substructure are associated with the corresponding candidate scheduling scheme, an unexecutable certificate is generated and bound, and the corresponding candidate scheduling scheme is excluded; The exclusion of the corresponding candidate scheduling scheme specifically includes: The unexecutable certificate corresponding to each candidate scheduling scheme in the candidate scheduling scheme set is read, and the candidate scheduling scheme associated with the unexecutable certificate is obtained; in the candidate scheduling scheme set, the candidate scheduling scheme associated with the unexecutable certificate is located; the located candidate scheduling scheme is removed from the candidate scheduling scheme set to form an updated candidate scheduling scheme set; the association relationship between the removed candidate scheduling scheme and the unexecutable certificate corresponding thereto is recorded, and the exclusion processing of the candidate scheduling scheme is completed; When the barrier avoidance verification result is not triggered, an executable certificate corresponding to the candidate scheduling scheme is generated and bound to form an intelligent scheduling result of an agricultural robot.

[0024] Embodiment 1: In order to verify the feasibility of the present application in implementation, the present application is applied to a large-scale grain planting operation scene in a northern plain area, which mainly produces wheat and corn rotation, the land distribution is continuous but the boundary is complex, the operation is intensive in spring and summer, the number of agricultural robots is large, the operation types are frequently crossed, in the actual operation process, the agricultural operation is jointly affected by many irreversible factors such as crop phenology stage, seasonal rainfall, soil bearing capacity change and agronomic operation sequence, in the past scheduling practice, the scheduling scheme is mostly generated according to the operation time window and the idle state of the robot, but in the execution stage, the soil bearing condition is often suddenly changed due to rainfall, the operation is forced to interrupt or rework due to missing the phenology window, thereby causing operation delay and resource waste, and the reliability of the scheduling result is difficult to guarantee.

[0025] In this job scenario, first, according to the method of the application, the agricultural data and agricultural robot data in the region are collected, and the agricultural state data set is generated through a unified preprocessing process. The agricultural data includes the spatial position of each plot, the operation type arrangement, the operation allowable time window, the current phenology stage of crops, the meteorological state change, the soil bearing state, and the agronomic operation sequence, etc. The agricultural robot data includes the current position of the robot, the executable operation capability, the operation state, and the travel path, etc. Through data cleaning, time alignment and space mapping, the data from different sources are mapped to the same time axis and space reference system, forming an agricultural state data set that can be used for subsequent scheduling analysis.

[0026] After the generation of the agricultural state data set, according to the method of the application, the state records corresponding to each operation object are analyzed to generate an agricultural operation irreversible constraint set. For crops in the critical growth stage, the allowable operation time range is determined according to the crop phenology stage. When the operation time in the scheduling scheme does not meet the requirements of the phenology stage, a phenology irreversible constraint is generated. For meteorological factors, when the meteorological conditions do not meet the requirements of the corresponding operation type execution, a meteorological irreversible constraint is generated. For soil conditions, when the soil bearing capacity is insufficient to support the corresponding operation, a soil bearing irreversible constraint is generated. At the same time, according to the agronomic operation sequence requirements, when the operation execution sequence violates the established agronomic sequence conditions, an agronomic sequence irreversible constraint is generated. Different types of irreversible constraints are collected to form an agricultural operation irreversible constraint set.

[0027] According to the method of the application, the agricultural operation irreversible constraint set is mapped to an irreversible barrier graph. Each irreversible constraint is mapped to a barrier node, and its corresponding trigger condition and action time window are associated as node attributes. When different irreversible constraints overlap in the action time window and have a chronological relationship in the trigger condition, a barrier edge is established between the corresponding barrier nodes to depict the trigger propagation relationship of the irreversible constraint in the time dimension. Through the above mapping process, the originally scattered agricultural operation restrictions are transformed into barrier graph constraint representation with structured propagation relationship.

[0028] In the scheduling stage, according to the method of the application, a plurality of candidate scheduling schemes are generated based on the agricultural state data set. This process considers the matching relationship between the operation tasks to be executed and the schedulable agricultural robots. Under the premise of meeting the basic operation time window requirements, the operation tasks are combined and time allocated to form a candidate scheduling scheme set. Each candidate scheduling scheme is then mapped to a scheme space-time trajectory to describe the operation arrangement and spatial movement process of the agricultural robot in each operation time allocation interval.

[0029] In the scheme verification stage, according to the method of the application, each scheme space-time trajectory is matched with the irreversible barrier graph constraint representation, by judging whether the scheme space-time trajectory meets the trigger condition of the barrier node in each operation time allocation interval, the scheme with irreversible risk can be identified before the scheduling scheme is issued, when it is determined that a candidate scheduling scheme triggers at least one barrier node in a specific time interval, the operation time allocation interval of the first triggered barrier node is further located, and the shortest counterexample trajectory is formed by backtracking, so as to lock the minimum time range leading to the unexecutable scheme, on this basis, the operation task, agricultural robot and corresponding relationship directly related to the trigger of the barrier node are extracted from the shortest counterexample trajectory, a minimum contradictory substructure is formed, and an unexecutable certificate is generated according to the minimum contradictory substructure, and the corresponding candidate scheduling scheme is excluded from the candidate scheduling scheme set.

[0030] For the candidate scheduling scheme that does not trigger any irreversible barrier node, the executable certificate is generated according to the method of the application, and the final intelligent scheduling result of the agricultural robot is formed. In actual operation process, it can be observed that the scheduling result formed by the method of the application can effectively avoid the interruption risk caused by irreversible operation condition in the execution stage, the operation continuity is significantly improved, the agricultural robot resource configuration is more stable, and the operation window is more accurate, so that the executable of the scheduling scheme is determined in advance in the complex agricultural operation scene, and the application value and feasibility of the method of the application in engineering practice are verified.

[0031] Table 1 Comparison of overall performance of different agricultural robot scheduling methods in complex operation scene

[0032] As can be seen from Table 1, in terms of scheduling scheme generation time, the method of the application has a certain degree of increase compared with the traditional rule-driven scheduling method and the multi-constraint heuristic scheduling method. This change is mainly due to the introduction of the matching verification process of the irreversible constraint set, the irreversible barrier graph and the scheme space-time trajectory in the scheduling stage of the application, so that the scheduling scheme needs to be more fully executable in the generation stage. It is worth noting that although the generation time is increased, the overall time is still within the acceptable time scale of agricultural operation scheduling, and does not have a substantial impact on the operation organization rhythm.

[0033] In terms of average execution completion rate of the scheduling scheme, the method of the present application shows a stable improvement trend compared with the comparative method. The traditional rule-driven scheduling method is prone to expose unpredictable problems in the execution stage when facing the cross action of multiple irreversible conditions, because it mainly relies on static rules and local constraints. Although the multi-constraint heuristic scheduling method introduces more constraint factors, it still does not fully consider the superposition relationship of constraints in the time dimension. The present application verifies the overall space-time trajectory of the scheme before scheduling, so that the scheduling scheme is less disturbed by irreversible conditions in the execution stage, thereby improving the execution completion rate.

[0034] From the interruption rate and rework ratio in the scheduling execution process, it can be further observed that the improvement of the execution stability of the method of the present application is not a radical jump, but a reasonable progressive improvement trend. The traditional method often has to deal with temporary interruption or rework when encountering changes in phenology or soil conditions during execution. The multi-constraint heuristic method alleviates this problem to some extent, but there are still interruptions caused by the constraint propagation relationship not being identified in advance. The present application describes the constraint trigger propagation relationship through the irreversible barrier graph, so that the scheme that is prone to cause execution failure is excluded in advance in the scheduling stage, thereby significantly reducing the probability of interruption and rework in the execution stage.

[0035] In terms of the occurrence ratio of job window loss, the method of the present application also shows continuous and stable advantages. Job window loss is usually caused by the scheduling scheme being forced to adjust and then unable to re-enter the effective time window during execution. The present application explicitly introduces the action time window in the irreversible constraint and judges in combination with the space-time trajectory of the scheme during scheme verification, so that the feasibility of the scheduling scheme in the time dimension is more fully guaranteed, thereby reducing the occurrence ratio of job window loss.

[0036] From the job continuity index and the stability of agricultural robot resource utilization, it can be seen that the method of the present application is superior to the comparative method in terms of overall execution smoothness. The results show that the present application does not simply pursue the short-term feasibility of the scheduling result, but excludes potential unexecutable schemes in advance, so that the agricultural robot maintains a relatively stable running state during the entire work cycle, reduces frequent scheduling adjustment and path reconstruction, and thereby realizes more balanced and reliable resource utilization effect.

[0037] The above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can make equivalent substitutions or changes to the technical solution and inventive concept of the present application within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for intelligent scheduling of agricultural robots based on big data and spatiotemporal coupling, characterized in that, Includes the following steps: Collect agricultural data and agricultural robot data, and preprocess them to generate an agricultural status dataset; A set of irreversible constraints for agricultural operations is generated based on an agricultural state dataset. The set of irreversible constraints for agricultural operations includes irreversible constraints for phenology, meteorology, soil carrying capacity, and agronomic sequence. A corresponding triggering condition and action time window are configured for each irreversible constraint. The set of irreversible constraints in agricultural operations is mapped to an irreversible barrier graph, the trigger points of irreversible constraints are mapped to barrier nodes, and the trigger propagation relationship between irreversible constraints is mapped to barrier edges. Based on the irreversible barrier graph, a barrier graph constraint representation is generated. A set of candidate scheduling schemes is generated based on the agricultural status dataset; Each candidate scheduling scheme in the candidate scheduling scheme set is mapped to a corresponding scheme spatiotemporal trajectory. The scheme spatiotemporal trajectory is matched with the irreversible barrier graph constraint representation to generate the corresponding barrier avoidance verification result. When the barrier avoidance verification result indicates that the candidate scheduling scheme triggers at least one barrier node, the shortest counterexample trajectory is generated and the minimum contradiction substructure is extracted. An inoperability certificate is generated and bound, and the corresponding candidate scheduling scheme is excluded. When no barrier node of the irreversible barrier graph is triggered, an executability certificate is generated and bound, thus forming the intelligent scheduling result of the agricultural robot.

2. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The agricultural data specifically includes plot identification data, plot spatial location data, operation type data, operation time window data, crop phenological stage data, meteorological status data, soil bearing capacity data, and agronomic operation sequence data. The agricultural robot data specifically includes robot identification data, spatial location information data, operation capability parameter data, execution time data, operation status data, and travel path data. The preprocessing specifically includes data cleaning, time alignment, and spatial mapping.

3. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The generation of the set of irreversible constraints for agricultural operations specifically includes: Extract operation type information, plot identification information, operation time information, crop phenological stage information, meteorological status information, soil bearing capacity information, and agronomic operation sequence information from the agricultural status dataset, and aggregate the information according to the same operation object to form an operation status record; Based on the crop phenological stage information in the operation status record, determine whether the corresponding operation time information meets the preset allowable operation time window conditions for that phenological stage. When the operation time does not meet the allowable operation time window conditions, generate irreversible phenological constraints and configure corresponding triggering conditions and action time windows for the irreversible phenological constraints. Based on the meteorological status information in the operation status record, determine whether the current meteorological status meets the preset meteorological execution conditions for the corresponding operation type. When the meteorological status does not meet the meteorological execution conditions, generate irreversible meteorological constraints and configure corresponding triggering conditions and action time windows for the irreversible meteorological constraints. Based on the soil carrying capacity information in the operation status record, determine whether the soil carrying capacity meets the preset carrying capacity conditions for performing the corresponding operation type on the corresponding plot. When the soil carrying capacity does not meet the preset carrying capacity conditions, generate irreversible soil carrying capacity constraints and configure corresponding triggering conditions and action time windows for the irreversible soil carrying capacity constraints. Based on the agronomic operation sequence information and operation time information in the operation status record, it is determined whether the operation execution sequence violates the preset agronomic operation sequence conditions. When the operation execution sequence violates the agronomic operation sequence conditions, an irreversible constraint of the agronomic sequence is generated, and corresponding trigger conditions and action time windows are configured for the irreversible constraint of the agronomic sequence. By combining irreversible constraints on phenology, meteorology, soil carrying capacity, and agronomic sequence, a set of irreversible constraints on agricultural operations is formed.

4. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The generation of the barrier graph constraint representation specifically includes: Read the set of irreversible constraints for agricultural operations, obtain the triggering conditions and action time windows corresponding to each irreversible constraint for agricultural operations, and use each irreversible constraint for agricultural operations as a mapping object. Each irreversible constraint of an agricultural operation is mapped to a barrier node, and the triggering condition and the time window of action corresponding to the irreversible constraint of the agricultural operation are associated in the barrier node to form a set of barrier nodes; In the set of barrier nodes, based on the triggering order of irreversible constraints in agricultural operations and the overlapping relationship between their action time windows, it is determined whether there is a trigger propagation relationship between irreversible constraints in agricultural operations. When a trigger propagation relationship is determined, a directed connection is established between the corresponding barrier nodes, and the directed connection is determined as a barrier edge, forming a set of barrier edges; By combining the set of barrier nodes and the set of barrier edges, an irreversible barrier graph is constructed. Based on the irreversible barrier graph, the triggering conditions and action time windows associated with the barrier nodes, as well as the trigger propagation relationships represented by the barrier edges, are uniformly organized to generate a barrier graph constraint representation.

5. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The generation of the candidate scheduling scheme set specifically includes: Retrieve data related to agricultural operations and agricultural robots from the agricultural status dataset; The status of data related to agricultural operations is determined, agricultural operations that are in the pending execution state are selected, and the selected agricultural operations are aggregated to form a set of operation tasks. The data related to agricultural robots are status-determined, agricultural robots in a schedulable state are selected, and the selected agricultural robots are aggregated to form an agricultural robot set. A correspondence between the set of tasks and the set of agricultural robots is established one by one, forming a set of correspondences between tasks and agricultural robots. By combining the time windows corresponding to the tasks, time allocation is performed on the tasks in the corresponding relationship set to generate the task time allocation results corresponding to the corresponding relationship set; The task set, agricultural robot set, corresponding relationship set, and task time allocation results are integrated to form candidate scheduling schemes; The multiple candidate scheduling schemes are aggregated to generate a candidate scheduling scheme set.

6. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The generation of the barrier circumvention verification result specifically includes: For each candidate scheduling scheme in the candidate scheduling scheme set, the combinations of agricultural robots performing corresponding tasks within the corresponding work time allocation interval are arranged in chronological order to generate the spatiotemporal trajectory of the scheme corresponding to the candidate scheduling scheme. Read the irreversible barrier graph constraint representation and obtain the triggering conditions and action time windows corresponding to each barrier node contained therein; Within each work time allocation interval corresponding to the spatiotemporal trajectory of the scheme, for work time allocation intervals that overlap with the time window of the barrier node, it is determined whether the corresponding spatiotemporal trajectory of the scheme meets the triggering condition of the barrier node. When the spatiotemporal trajectory of the scheme meets the triggering condition of at least one barrier node within any job time allocation interval, the barrier avoidance verification result corresponding to the candidate scheduling scheme is generated as a trigger. If the triggering conditions of any barrier node are not met within any of the job time allocation intervals corresponding to the spatiotemporal trajectory of the scheme, the barrier avoidance verification result corresponding to the candidate scheduling scheme is "not triggered".

7. The intelligent scheduling method for agricultural robots based on big data and spatiotemporal coupling according to claim 1, characterized in that, The generation of the intelligent scheduling results for agricultural robots specifically includes: Read the barrier avoidance verification results corresponding to each candidate scheduling scheme in the candidate scheduling scheme set; When the barrier avoidance verification result is triggered, locate the operation time allocation interval of the first triggered barrier node in the spatiotemporal trajectory of the candidate scheduling scheme; Using the first triggered operation time allocation interval as the termination position, trace back the continuous operation time allocation intervals related to the barrier node in the spatiotemporal trajectory of the scheme, and extract them to form the shortest counterexample trajectory. In the shortest counterexample trajectory, select the tasks, agricultural robots and their corresponding relationships that participate in the triggering of the barrier node, remove the remaining tasks and their corresponding relationships that do not participate in the triggering of the barrier node, and extract the minimum contradiction substructure. Associate the shortest counterexample trajectory and the minimum contradictory substructure with the corresponding candidate scheduling schemes, generate and bind an inexecutability certificate, and exclude the corresponding candidate scheduling schemes. When the barrier avoidance verification result is not triggered, an executable certificate corresponding to the candidate scheduling scheme is generated and bound to form the intelligent scheduling result of agricultural robots.

Images