Unmanned aerial vehicle fleet task description system capable of embedding AI strategy

Abstract

The invention discloses an unmanned aerial vehicle cluster task description system capable of being embedded with an AI strategy, and relates to the technical field of computer application, and the system is characterized in that firstly, one main task is disassembled into a plurality of multi-machine tasks, the multi-machine tasks are disassembled into a plurality of single-machine behaviors, and the single-machine behaviors are disassembled into a plurality of single-machine actions; therefore, a complex task is decomposed into clear independent steps, and independent execution and overall cooperation of the unmanned aerial vehicle are facilitated; secondly, an import statement is introduced into the system, fusion of an AI strategy and rule control is achieved, and the task intelligence level is improved; meanwhile, a flow control structure and a multi-level concurrent primitive are adopted, and task sequence and multi-level concurrent execution are supported; finally, the system strengthens the autonomous cooperation capability of the unmanned aerial vehicles through a communication primitive based on a blackboard mechanism, can achieve information sharing and coordination operation without a central scheduler, and remarkably improves the cooperation efficiency and task success rate of an unmanned aerial vehicle group.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a UAV swarm mission description system capable of embedding AI strategies. Background Art

[0002] In recent years, drones have shown broad application prospects in many fields, including but not limited to aerial photography, environmental monitoring, disaster relief, agricultural production, and safety inspections. With the rapid development of artificial intelligence and drone technology, the ability of drone fleets to perform complex tasks in different fields has been significantly improved.

[0003] However, with the increasing number of drones and the increasing complexity of missions, the traditional manual control method of swarms has gradually exposed the problem of insufficient flexibility. This method has poor adaptability to dynamic environments and is difficult to effectively handle complex and changing environmental information. Therefore, in order to improve the autonomy and intelligence level of drones and enable them to better cope with dynamic environments, it is particularly important to introduce AI strategies.

[0004] Although some studies have attempted to apply rule-driven methods and AI strategies to UAV swarm missions, these studies still face a common problem: the lack of a unified task description system that can support both methods at the same time; in addition, the existing UAV swarm task description system has obvious deficiencies in describing task decomposition, concurrent execution of multiple tasks, and autonomous collaboration between UAVs, and cannot meet actual needs. Summary of the invention

[0005] The purpose of the present invention is to provide a UAV swarm mission description system that can be embedded with AI strategies to solve the problems raised in the prior art.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a drone swarm task description system that can be embedded with AI strategy, the system includes a platform description module, a dependency library import module, an action definition module, a drone type definition module, a single-machine behavior definition module, a multi-machine task definition module, a drone swarm task module, and a communication primitive module;

[0007] The platform description module is used to describe the simulation software platform on which the program is running;

[0008] Dependency library import module, used to import external dependency packages;

[0009] The drone type definition module is used to describe the actions that a drone can perform, and consists of action names;

[0010] A drone swarm task module, used to describe the drone swarm task and issue call instructions to the drone type definition module and the multi-machine task definition module respectively;

[0011] The multi-machine task definition module decomposes the UAV swarm task into multiple multi-machine tasks through the process control structure;

[0012] The single-machine behavior definition module decomposes each multi-machine task into multiple single-machine behaviors through the process control structure;

[0013] Action definition module, used to decompose each single-machine behavior into multiple single-machine actions;

[0014] The communication primitive module publishes messages on the blackboard during the execution of drone swarm tasks based on the blackboard mechanism.

[0015] Further, the process control structure includes an initialization structure, a target structure and a routine structure, wherein the initialization structure, the target structure and the routine structure are all prefixed with an @ symbol;

[0016] The initialization structure is used for initializing tasks and behaviors in the multi-machine task definition module and the single-machine behavior definition module, and the initialization is performed only once;

[0017] A target structure for determining whether the multi-machine task executed by the multi-machine task definition module has achieved the task target and determining whether the single-machine behavior executed by the single-machine behavior definition module has achieved the behavior target;

[0018] The routine structure is executed after the initialization, and determines whether to repeat the execution of the routine structure according to the target structure. The multi-machine task definition module realizes the concurrency of multi-machine group tasks through the multi-level concurrency primitives in the routine structure. The single-machine behavior definition module describes the concurrency between multiple devices on a drone through the multi-level concurrency primitives in the routine structure.

[0019] Furthermore, multi-level concurrency primitives include the each statement and the || concurrency primitive;

[0020] The each statement is used to describe that all drones in the same drone group concurrently complete a single-machine behavior, and the structure of the each statement is composed of the call statements of all drones in the same drone group and the call statement of the single-machine behavior;

[0021] The two parts of the structure adjacent to the || concurrency primitive are executed concurrently. If the || concurrency primitive is located in the routine structure of the multi-machine task definition module, it indicates the concurrency of different drone groups. If the || concurrency primitive is located in the routine structure of the single-machine behavior definition module, it indicates the concurrency between multiple devices of a single drone.

[0022] "Process control" is used to manage the execution order of tasks, which improves the readability of tasks. At the same time, in response to the multi-level concurrency requirements in tasks, each and || multi-level concurrency primitives are designed to support the concurrent description of tasks between multiple drone groups, between multiple drones within a drone group, and between multiple devices on a drone.

[0023] Further, the UAV swarm mission module includes a UAV type definition module calling unit and a multi-machine mission definition module calling unit;

[0024] The UAV type definition module call unit declares the type and number of UAVs participating in the UAV swarm mission;

[0025] The multi-machine mission definition module call unit declares the types and numbers of the drones participating in the same multi-machine mission.

[0026] Further, the multi-machine task definition module includes multi-machine task definition units, and each of the multi-machine task definition units can call the single-machine behavior definition module and the remaining multi-machine task definition units;

[0027] The structure of each of the multi-machine task definition units is a process control structure, and the multi-machine task process is defined by executing the process control structure. The definition of the multi-machine task process includes:

[0028] Step S5-1: Initialize the multi-machine task through the initialization structure;

[0029] Step S5-2: After the multi-machine task is initialized, the steps in the routine structure are executed. The routine structure includes a || concurrency primitive, an each statement, and an order statement. The || concurrency primitive is used to realize the concurrency of different drone groups. The each statement is used to realize the concurrency of multiple drones in a drone group. The order statement is used to control the drones in a drone group to perform tasks in sequence.

[0030] Step S5-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, the multi-machine task is terminated. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. The truth value of the expression is used to determine whether to repeat the steps in the routine structure. If the value of the expression is false, the steps in the routine structure are repeated. If the value of the expression is true, the multi-machine task is terminated.

[0031] Further, the stand-alone behavior definition module includes stand-alone behavior definition units, each of which can call the action definition module and other stand-alone behavior definition units;

[0032] The structure of each of the stand-alone behavior definition units is a process control structure. The stand-alone behavior process is defined by executing the process control structure. The definition of the stand-alone behavior process includes:

[0033] Step S6-1: initializing the structure to initialize the stand-alone behavior;

[0034] Step S6-2: After the stand-alone behavior is initialized, the steps in the routine structure are executed, and the concurrency between multiple devices on a single drone is realized through the || concurrency primitive in the routine structure;

[0035] Step S6-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, terminate the stand-alone behavior. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. Whether to repeat the steps in the routine structure is determined based on the truth value of the expression. If the value of the expression is false, repeat the steps in the routine structure. If the value of the expression is true, terminate the stand-alone behavior.

[0036] Furthermore, the dependency library import module describes the file name of the external dependency package through the dependency import unit;

[0037] For the above steps, the system introduces an import statement in the dependent library import module. Users can use this statement to integrate AI-based strategy modules into task steps, such as search strategies trained by deep learning. In this way, rule control (such as takeoff, landing, flying to a specific location, etc.) and AI strategies jointly participate in task execution, adapt to environmental changes and improve task success rate.

[0038] Furthermore, the action definition module includes an API calling unit, the API calling unit is used to describe the process of a single API call of a device by the drone, and the API calling unit is defined by an external dependency file;

[0039] In this system, complex swarm tasks can be decomposed into a series of steps according to the hierarchical structure of "Main-Task-Behavior-Action", so that each drone can perform independently and complete the overall task together.

[0040] Further, the communication primitive module includes a POST statement, a GET statement, a DGET statement, and a MODIFY statement;

[0041] The communication primitive module adopts the communication primitive based on the blackboard mechanism to present the messages on the blackboard in the form of key-value pairs. The key-value is not a one-to-one correspondence, and the value can be a numeric value or a string.

[0042] POST statement, used to publish key-value pairs on the blackboard;

[0043] A GET statement is used to obtain a value corresponding to a key from the blackboard and store the value in a local variable;

[0044] A MODIFY statement is used to modify the value corresponding to one of the keys on the blackboard. The execution logic of the MODIFY statement is based on the execution of the GET statement, and the local variable is used in the expression;

[0045] The system has added communication primitives such as POST, GET, DGET, MODIFY based on the black book mechanism to achieve information sharing and coordination between drones. This mechanism enhances the collaborative ability of drone clusters, making it possible to complete task information sharing and coordination operations without relying on a central scheduler.

[0046] Furthermore, the platform description module interacts with the simulation environment by configuring the platform statement. When the platform statement is not configured, only the task logic is executed without interacting with the simulation environment, and the platform statement is configured at most once.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. Clear description of task breakdown

[0049] The present invention provides an effective task decomposition method, which decomposes complex swarm tasks into independent steps that can be executed by each drone. By defining different types of drones and their capabilities, users are allowed to allocate drones that suit their task requirements through a drone type definition module. In this system, complex swarm tasks can be decomposed into a series of steps according to the hierarchical structure of "Main-Task-Behavior-Action", so that each drone can execute independently and complete the overall task together.

[0050] 2. Integration of AI strategy reasoning and rule function calls

[0051] To meet the needs of intelligent tasks, the system introduces a dependency library import module, through which users can integrate AI-based strategies into steps. In this way, rule control and AI strategies jointly participate in task execution, adapt to environmental changes and improve task success rate.

[0052] 3. Sequential and concurrent support

[0053] The system uses process control to manage the basic sequential execution of tasks, significantly improving the readability of tasks. In response to the multi-level concurrency requirements in tasks, the each statement and || concurrency primitives are designed to support concurrent description of tasks between multiple drone fleets, between multiple drones in a fleet, and between multiple devices on a drone.

[0054] 4. Collaboration mechanism supported by the blackboard system

[0055] In order to realize the autonomous collaboration of multiple UAVs, the system also provides communication primitives such as POST, GET, DGET, MODIFY based on the blackboard mechanism to realize the autonomous collaboration of UAVs. This mechanism enhances the collaborative ability of UAV clusters and can realize the sharing and coordinated operations of task information without relying on a central scheduler. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is an execution flow chart of a multi-machine task definition module and a single-machine behavior definition module in a UAV swarm task description system that can be embedded with an AI strategy according to the present invention;

[0057] Figure 2 This is a schematic diagram of the process of controlling multiple drones to concurrently execute the same behavior using each statement in a drone swarm task description system that can be embedded with AI strategies in the present invention. DETAILED DESCRIPTION

[0058] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0059] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a UAV swarm task description system that can be embedded with AI strategy, the system includes a platform description module, a dependency library import module, an action definition module, a UAV type definition module, a single-machine behavior definition module, a multi-machine task definition module, a UAV swarm task module, and a communication primitive module;

[0060] The platform description module is used to describe the simulation software platform on which the program is running;

[0061] Dependency library import module, used to import external dependency packages;

[0062] The drone type definition module is used to describe the actions that a drone can perform, and consists of action names;

[0063] A drone swarm task module, used to describe the drone swarm task and issue call instructions to the drone type definition module and the multi-machine task definition module respectively;

[0064] The multi-machine task definition module decomposes the UAV swarm task into multiple multi-machine tasks through the process control structure;

[0065] The single-machine behavior definition module decomposes each multi-machine task into multiple single-machine behaviors through the process control structure;

[0066] Action definition module, used to decompose each single-machine behavior into multiple single-machine actions;

[0067] The communication primitive module publishes messages on the blackboard during the execution of drone swarm tasks based on the blackboard mechanism.

[0068] The process control structure includes an initialization structure, a target structure and a routine structure, wherein the initialization structure, the target structure and the routine structure are all prefixed with an @ symbol;

[0069] The initialization structure is used for initializing tasks and behaviors in the multi-machine task definition module and the single-machine behavior definition module, and the initialization is performed only once;

[0070] A target structure for determining whether the multi-machine task executed by the multi-machine task definition module has achieved the task target and determining whether the single-machine behavior executed by the single-machine behavior definition module has achieved the behavior target;

[0071] The routine structure is executed after the initialization, and determines whether to repeat the execution of the routine structure according to the target structure. The multi-machine task definition module realizes the concurrency of multi-machine group tasks through the multi-level concurrency primitives in the routine structure. The single-machine behavior definition module describes the concurrency between multiple devices on a drone through the multi-level concurrency primitives in the routine structure.

[0072] Among them, multi-level concurrency primitives include each statement and || concurrency primitive;

[0073] The each statement is used to describe that all drones in the same drone group concurrently complete a single-machine behavior, and the structure of the each statement is composed of the call statements of all drones in the same drone group and the call statement of the single-machine behavior;

[0074] The two structures adjacent to the || concurrency primitive are executed concurrently. If the || concurrency primitive is located in the routine structure of the multi-machine task definition module, it indicates the concurrency of different drone groups. If the || concurrency primitive is located in the routine structure of the single-machine behavior definition module, it indicates the concurrency between multiple devices of a single drone.

[0075] Among them, the UAV swarm task module includes a UAV type definition module calling unit and a multi-machine task definition module calling unit;

[0076] The UAV type definition module call unit declares the type and number of UAVs participating in the UAV swarm mission;

[0077] The multi-machine mission definition module call unit declares the types and numbers of the drones participating in the same multi-machine mission.

[0078] In the embodiment of the present invention, Main is used as the program entry, the type of drones included in the task is declared with an Agent call statement, and the defined Task multi-machine task is called to start the task execution process.

[0079] Wherein, the multi-machine task definition module includes multi-machine task definition units, and each of the multi-machine task definition units can call the single-machine behavior definition module and the remaining multi-machine task definition units;

[0080] The structure of each of the multi-machine task definition units is a process control structure, and the multi-machine task process is defined by executing the process control structure. The definition of the multi-machine task process includes:

[0081] Step S5-1: Initialize the multi-machine task through the initialization structure;

[0082] Step S5-2: After the multi-machine task is initialized, the steps in the routine structure are executed. The routine structure includes a || concurrency primitive, an each statement, and an order statement. The || concurrency primitive is used to realize the concurrency of different drone groups. The each statement is used to realize the concurrency of multiple drones in a drone group. The order statement is used to control the drones in a drone group to perform tasks in sequence.

[0083] Step S5-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, the multi-machine task is terminated. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. The truth value of the expression is used to determine whether to repeat the steps in the routine structure. If the value of the expression is false, the steps in the routine structure are repeated. If the value of the expression is true, the multi-machine task is terminated.

[0084] In an embodiment of the present invention, after entering a multi-machine task Task, the statements in the init code block are executed to implement task initialization, and then the statements in the routine code block are executed to execute the task process, and finally, whether the routine code block needs to be repeatedly executed is determined according to the value of the expression at the end of the goal code block; in the routine code block, the || symbol is used to implement multi-machine cluster concurrency, the each statement is used to implement multi-UAV concurrency within a cluster, and the order statement is used to implement multiple UAVs within a cluster to execute certain statements sequentially.

[0085] The stand-alone behavior definition module includes stand-alone behavior definition units, and each of the stand-alone behavior definition units can call action definition modules and other stand-alone behavior definition units;

[0086] The structure of each of the stand-alone behavior definition units is a process control structure. The stand-alone behavior process is defined by executing the process control structure. The definition of the stand-alone behavior process includes:

[0087] Step S6-1: initializing the structure to initialize the stand-alone behavior;

[0088] Step S6-2: After the stand-alone behavior is initialized, the steps in the routine structure are executed, and the concurrency between multiple devices on a single drone is realized through the || concurrency primitive in the routine structure;

[0089] Step S6-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, terminate the stand-alone behavior. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. Whether to repeat the steps in the routine structure is determined based on the truth value of the expression. If the value of the expression is false, repeat the steps in the routine structure. If the value of the expression is true, terminate the stand-alone behavior.

[0090] In an embodiment of the present invention, after entering a stand-alone behavior Behavior, similar to the steps in a Task, the statements in the init code block are executed to initialize the behavior, and then the statements in the routine code block are executed to execute the behavior process. Finally, the value of the expression at the end of the goal code block is used to determine whether the routine code block needs to be repeatedly executed; in the routine code block, the || symbol is used to describe the concurrency between multiple devices on a drone. In addition to calling a stand-alone action Action, other Behaviors can also be nested.

[0091] Among them, the dependency library import module describes the file name of the external dependency package through the dependency import unit.

[0092] Among them, the action definition module includes an API calling unit, which is used to describe the process of a single API call of a drone to a certain device, and the API calling unit is defined by an external dependency file.

[0093] In the embodiment of the present invention, after entering the stand-alone as an Action, the API call unit in the external dependency package imported by the import statement is executed. Action is the smallest unit in program execution and cannot be interrupted.

[0094] Among them, the communication primitive module includes POST statement, GET statement, DGET statement and MODIFY statement;

[0095] The communication primitive module adopts the communication primitive based on the blackboard mechanism to present the messages on the blackboard in the form of key-value pairs. The key-value is not a one-to-one correspondence, and the value can be a numeric value or a string.

[0096] POST statement, used to publish key-value pairs on the blackboard;

[0097] A GET statement is used to obtain a value corresponding to a key from the blackboard and store the value in a local variable;

[0098] The MODIFY statement is used to modify the value corresponding to one of the keys on the blackboard. The execution logic of the MODIFY statement is based on the execution of the GET statement, and the local variable is used in the expression.

[0099] The platform description module interacts with the simulation environment by configuring the platform statement. When the platform statement is not configured, only the task logic is executed without interacting with the simulation environment. The platform statement is configured at most once.

[0100] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A drone swarm mission description system that can embed AI strategies, characterized by: The system includes a platform description module, a dependency library import module, an action definition module, a drone type definition module, a single-machine behavior definition module, a multi-machine task definition module, a drone group task module, and a communication primitive module; The platform description module is used to describe the simulation software platform on which the system is based when running; The dependency library import module is used to import external dependency packages; The drone type definition module is used to describe the actions that a drone has, and is composed of action names; The drone swarm task module is used to describe the drone swarm task and issue call instructions to the drone type definition module and the multi-machine task definition module respectively; The multi-machine task definition module decomposes the UAV group task into multiple multi-machine tasks through a process control structure; The single-machine behavior definition module decomposes each multi-machine task into multiple single-machine behaviors through a process control structure; The action definition module is used to decompose each single-machine behavior into multiple single-machine actions; The communication primitive module publishes messages during the execution of the UAV swarm mission on the blackboard based on the blackboard mechanism.

2. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The process control structure includes an initialization structure, a target structure and a routine structure, wherein the initialization structure, the target structure and the routine structure are all prefixed with an @ symbol; The initialization structure is used for initializing tasks and behaviors in the multi-machine task definition module and the single-machine behavior definition module, and the initialization is performed only once; The target structure is used to determine whether the multi-machine task executed by the multi-machine task definition module has achieved the task target and to determine whether the single-machine behavior executed by the single-machine behavior definition module has achieved the behavior target; The routine structure is executed after the initialization, and determines whether to repeat the execution of the routine structure according to the target structure. The multi-machine task definition module realizes the concurrency of multi-machine group tasks through the multi-level concurrency primitives in the routine structure. The single-machine behavior definition module can describe the concurrency between multiple devices on a drone through the multi-level concurrency primitives in the routine structure.

3. The UAV swarm mission description system capable of embedding AI strategy according to claim 2 is characterized by: The multi-level concurrency primitives include each statement and || concurrency primitive; The each statement is used to describe that all drones in the same drone group concurrently complete a single-machine behavior, and the structure of the each statement is composed of the call statements of all drones in the same drone group and the call statement of the single-machine behavior; The two structures adjacent to the ||concurrency primitive are executed concurrently. If the ||concurrency primitive is located in the routine structure of the multi-machine task definition module, it indicates the concurrency of different drone groups. If the ||concurrency primitive is located in the routine structure of the single-machine behavior definition module, it indicates the concurrency between multiple devices of a single drone.

4. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The UAV swarm task module includes a UAV type definition module calling unit and a multi-machine task definition module calling unit; The drone type definition module calling unit declares the type and number of drones participating in the drone swarm mission; The multi-machine mission definition module calling unit declares the types and numbers of the drones participating in the same multi-machine mission.

5. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The multi-machine task definition module includes multi-machine task definition units, and each of the multi-machine task definition units can call the single-machine behavior definition module and the remaining multi-machine task definition units; The structure of each of the multi-machine task definition units is a process control structure, and the multi-machine task process is defined by executing the process control structure. The multi-machine task process definition includes: Step S5-1: Initialize the multi-machine task through the initialization structure; Step S5-2: After the multi-machine task is initialized, the steps in the routine structure are executed. The routine structure includes a || concurrency primitive, an each statement, and an order statement. The || concurrency primitive is used to realize the concurrency of different drone groups. The each statement is used to realize the concurrency of multiple drones in a drone group. The order statement is used to control the drones in a drone group to perform tasks in sequence. Step S5-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, the multi-machine task is terminated. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. The truth value of the expression is used to determine whether to repeat the steps in the routine structure. If the value of the expression is false, the steps in the routine structure are repeated. If the value of the expression is true, the multi-machine task is terminated.

6. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The stand-alone behavior definition module includes stand-alone behavior definition units, each of which can call action definition modules and other stand-alone behavior definition units; The structure of each of the stand-alone behavior definition units is a process control structure, and the stand-alone behavior process is defined by executing the process control structure. The definition of the stand-alone behavior process includes: Step S6-1: initializing the structure to initialize the stand-alone behavior; Step S6-2: After the stand-alone behavior is initialized, the steps in the routine structure are executed, and the concurrency between multiple devices on a single drone is realized through the || concurrency primitive in the routine structure; Step S6-3: After executing the steps in the routine structure, execute the steps in the target structure. When the target structure is empty, terminate the stand-alone behavior. When the target structure is not empty, the last statement in the target structure must be an expression prefixed by $. Whether to repeat the steps in the routine structure is determined based on the truth value of the expression. If the value of the expression is false, repeat the steps in the routine structure. If the value of the expression is true, terminate the stand-alone behavior.

7. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The dependency library import module describes the file name of the external dependency package through the dependency import unit.

8. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The action definition module includes an API calling unit, which is used to describe the process of a single API call of a device by the drone, and the API calling unit is defined by an external dependency file.

9. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The communication primitive module includes POST statement, GET statement, DGET statement and MODIFY statement; The communication primitive module adopts the communication primitive based on the blackboard mechanism to present the messages on the blackboard in the form of key-value pairs, wherein the key-values ​​are not in a one-to-one correspondence, and the values ​​include numerical values ​​or character strings; The POST statement is used to publish the key-value pair on the blackboard; The GET statement is used to obtain a value corresponding to a key from the blackboard and store the value in a local variable; The DGET statement is executed after the GET statement, reads the value in the local variable and deletes the value on the blackboard; The MODIFY statement is used to modify the value corresponding to one of the keys on the blackboard. The execution logic of the MODIFY statement is based on the execution of the GET statement, and the local variable is used in the expression.

10. The UAV swarm mission description system capable of embedding AI strategy according to claim 1 is characterized by: The platform description module interacts with the simulation environment by configuring the platform statement. When the platform statement is not configured, only the task logic is executed without interacting with the simulation environment, and the platform statement is configured at most once.

Images