Cluster unmanned aerial vehicle task cooperation semi-physical simulation method

Abstract

The invention provides a cluster unmanned aerial vehicle task cooperation semi-physical simulation method in order to solve the problem that in the prior art, a task intention is usually input in a static parameter form and is difficult to adapt to a complex dynamic task scene. The invention belongs to the technical field of unmanned aerial vehicles. The method comprises the steps of user expected intention determination, initial representation, preliminary calculation, intention integration, hardware control and physical simulation, deep reduction, physical simulation, dynamic intention representation and iterative optimization. By introducing the dynamic intention representation mechanism of the task intention, the expected intention input by the user and the initial intention representation generated by the system can be integrated in real time, the dynamic change of the task demand is effectively coped with, and the accuracy of task allocation and execution is improved. Through a closed-loop feedback mechanism, a simulation result can react to adjustment and optimization of a task intention. Physical simulation is carried out based on the physical model and the kinetic equation, and the simulation result is more accurate.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicles, and in particular relates to a method for semi-physical simulation of cluster unmanned aerial vehicle task collaboration. Background Art

[0002] Semi-physical simulation technology combines the advantages of physical hardware and virtual simulation and is an important tool for verifying the performance of UAV systems. For example, NASA's Unmanned Aircraft Systems Traffic Management (UTM) system optimizes the traffic scheduling and collaboration strategies of multiple UAV missions through semi-physical simulation.

[0003] In the existing technology, task intention is usually input in the form of static parameters, which is difficult to adapt to complex dynamic task scenarios. Summary of the invention

[0004] In order to solve the problem in the prior art that mission intent is usually input in the form of static parameters and is difficult to adapt to complex dynamic mission scenarios, the present invention provides a semi-physical simulation method for cluster UAV mission collaboration.

[0005] The technical solution adopted by the present invention is: A method for semi-physical simulation of swarm UAV mission collaboration, comprising the steps of: Step S1, determine the user's desired intention e and initial representation o ' ; Among them, the user's expected intention e is the user's expected drone position, attitude, and speed; Initial Characterization ' This is the initial visualization simulation result; Step S2: the user's expected intention e and the initial representation o ' Combine and generate comprehensive intention input i; Step S3, the comprehensive intention is inputted into i for preliminary solution to generate a paradigm intention i ' ; Step S4, the paradigm intention i ' Perform deep calculations to generate composite paradigm intent '' ; The composite paradigm intends '' Expressed as: '' ={τ1,τ2,…,τ k ,…,τ n}; Where n is the number of tasks; τ n represents the nth task; τ k is the kth task in the composite paradigm intention, including the following attributes, expressed as: τ k= (Location k ,Payload k ,Priority k ); Location k is the task location coordinate; Payload k For the mission load; Priority k Prioritize the tasks; Step S5, the composite paradigm intention i '' Convert to command list CommandList; Step S6, constructing a hardware control module and a physical simulation module; The hardware control module generates a flight control command signal u(t) at time t according to the command list CommandList; The hardware control module sends the flight control command signal u(t) to the physical simulation module, and the physical simulation module processes the flight control command signal u(t) and converts the real-time status information x of the drone into state (t) feeding back to the hardware control module via the network; Step S7, according to the composite paradigm intention i '' Perform deep restoration and convert it into physical parameters p; The physical parameter p includes the real-time status information x of the drone. state (t), Step S8, based on the flight control command signal u(t) at time t and the UAV dynamics model, the real-time state information x is calculated. state (t); Step S9: The real-time status information x of the drone is state (t) Convert into visual simulation result o and output; Step S10, repeatedly executing the steps S2 to S9, performing simulation at time t+1, and obtaining the visual simulation result o at time t+1; wherein the real-time status information x of the drone is used state (t) replace the initial characterization in step S2 ' .

[0006] Furthermore, in step S2, when the user expected intention e and the initial representation o' are combined, the combination process is to balance the user expected intention e and the initial representation o' in a linear weighted manner. ' The influence of is expressed as: i=w e ×e+wo' ×o ' ; Among them, w e is the weight of the user's desired intention e, w o' For the initial characterization ' The weight of the user's desired intention e e and the initial characterization o ' The weight w o' Calculated according to the optimization goal.

[0007] Furthermore, the weight w of the user's expected intention e is e and the initial characterization o ' The weight w o' The calculation expression is: ; The constraints are expressed as: ; Among them, Sim is the similarity function, Expressed as: ; in, for The norm of ; for The norm of ; e is the user's expected intention; o ' For initial characterization; w e is the weight of the user's desired intention e; w o' For the initial characterization ' The weight of .

[0008] Furthermore, in step S3, when performing a preliminary solution to the comprehensive intention input i, the method used is: using a rule matching algorithm to map the non-standard comprehensive intention input i into the paradigm intention definition domain.

[0009] Furthermore, in step S4, the paradigm intent i ' Perform deep calculations to generate composite paradigm intent '' Method used: Paradigm intention ' Split into tasks and generate composite paradigm intentions i '' , expressed as: ; Among them, T k is the kth task; the expression represents the composite paradigm intention i '' It is task T k The union of .

[0010] Furthermore, the step S5 specifically performs the following steps: Get the environment parameter Env, which is expressed as: Env = {ε1, ε2, …, ε i , …, ε m}; Among them, m is the number of environmental variables that affect the execution of UAV missions, ε i is the i-th environmental variable that affects the execution of the UAV mission; ε m is the mth environmental variable that affects the execution of the UAV mission; Get the resource parameter Res, which is expressed as: Res = {ρ1, ρ2, …, ρ j ,…,ρ p}; Among them, ρ j represents the available resources of the jth UAV cluster; ρ p represents the available resources of the p-th drone cluster; Get the mission instructions of each drone and construct the mission set Cmd of the i-th drone i , expressed as: Cmd i ={σ i1 , σ i2 , …, σ ij , …, σ im}; Among them, σ ij The mission set Cmd for the drone i The jth task of the i-th UAV in ; σ im The mission set Cmd for the drone i The mth mission of the i-th UAV; Get the i-th UAV to perform the mission σ ij Completion time T ij ; Get the i-th UAV to perform the mission σ ij Resource consumption R ij ; Get task σ ij The completion status Completed(σ ij ); Among them, the completion status Completed(σ ij ) is 1, indicating completion. The completion status Completed(σ ij ) is 0, indicating that it is not completed; According to the kth task τ in the composite paradigm intentionk , the environment parameter Env, and the resource parameter Res construct the task feasibility constraint condition f feasible , expressed as: f feasible (τ k ,Env,Res)=1; Among them, the expression represents that for the k-th task τ k It is feasible under the environment parameter Env and the resource parameter Res; According to the available resources ρ of the j-th UAV cluster j , and the i-th UAV performs mission σ ij Resource consumption R ij Construct resource limit constraints, expressed as: ; The expression represents the i-th UAV performing mission σ ij Resource consumption R ij The available resources ρ of the i-th drone cluster must not exceed i ; For all i; According to the kth task τ k , and the mission set Cmd of the i-th UAV i Construct resource limit constraints, expressed as: ; Where n is the number of tasks, and the expression means that each task must be assigned to at least one UAV; According to the i-th UAV performing the task σ ij Completion time T ij , the i-th UAV performs mission σ ij Resource consumption R ij , the task σ ij The completion status Completed(σ ij ), task feasibility constraints f feasible , the i-th UAV performs mission σ ij Resource consumption R ij Construct resource restriction constraints and the task set Cmd of the i-th UAV i Build resource limit constraints to build command list CommandList, expressed as: ; The constraints are expressed as: ; ; ; Among them, α1, α2, and α3 are , ,as well as The weight is adjusted according to the task requirements.

[0011] Furthermore, the command list CommandList, the flight control command signal u(t), and the real-time status information x of the drone involved in step S6 state (t) and the process can be expressed by the following expression: ; Among them, u prew (t) is the flight control command signal at the previous time of time t, that is, the flight control command signal at time t-1; ∆u(t) is the real-time status information x of the UAV based on the physical simulation module. state (t) a correction amount fed back to the hardware control module; K is the adjustment gain matrix, which controls the adjustment speed and amplitude of the drone; f feedback Indicates calculation of the current state x state (t) and the target state x target Deviation of (t); g is a deviation conversion function used to convert the deviation f feedback Converted into the correction value ∆u(t).

[0012] Furthermore, the hardware control module includes an ARM architecture computer, and the ARM architecture computer is used to parse task instructions.

[0013] Furthermore, the weight w of the user's expected intention e is e The initial value is set to 0.5; The initial characterization ' The weight w o' The initial value is set to 0.5.

[0014] Furthermore, the step S10 also iteratively updates the weight w of the user's expected intention e e and the initial characterization o ' The weight w o' , the update method is to use gradient descent or Lagrange multiplier method to optimize the weight, the weight w of the user's expected intention e e The update direction is: ;in, express right Find partial derivatives; Initial Characterization ' The weight w o'The update direction is: ;in, express right Find the partial derivative.

[0015] The beneficial effects of the present invention are: By introducing a dynamic intent representation mechanism for task intent, the present invention can integrate the desired intent input by the user and the initial intent representation generated by the system in real time, and effectively respond to dynamic changes in task requirements. Improve the accuracy of task allocation and execution, and reduce resource waste and task failures caused by intention understanding deviations. Through a closed-loop feedback mechanism, the present invention enables the simulation results to react to the adjustment and optimization of task intent. The simulation system can adaptively adjust simulation parameters during operation, significantly enhancing the intelligence level of the system. The present invention performs physical simulation based on precise physical models and dynamic equations, and the simulation results are closer to the task execution status of a real drone cluster. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 The flowchart of the collaborative semi-physical simulation method for swarm UAV missions is provided; Figure 2 Intent control flow graph for collaborative hardware-in-the-loop simulation method for swarm UAV missions. DETAILED DESCRIPTION

[0018] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0019] The disclosure below provides many different embodiments or examples to realize different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention.

[0020] The embodiments of the invention are described in detail below with reference to the accompanying drawings.

[0021] This embodiment aims at the expansion scale of the number of swarm UAVs from 10 to 100, and verifies the feasibility and performance of the swarm UAV task collaboration semi-physical simulation system proposed in the present invention in terms of task solving, simulation efficiency and realism.

[0022] The method for semi-physical simulation of cluster UAV task collaboration disclosed in this embodiment is as shown in the attached flowchart. Figure 1 As shown, the steps include: Step S1: prepare 10 to 100 groups of ARM architecture computers, each group corresponds to a drone, determine the user's expected intention e and the initial representation o ' ; Among them, the user's expected intention e is the user's expected drone position, attitude, and speed; Initial Characterization ' This is the initial visualization simulation result; Step S2: user expected intention e and initial representation o ' Combine and generate comprehensive intention input i; Step S3, import the comprehensive intention input i into the computer front end for preliminary solution to generate the paradigm intention i ' ; Step S4, the paradigm intention i ' The computer backend performs deep calculations to generate composite paradigm intent i '' ; Composite Paradigm Intention '' Expressed as: i '' ={τ1,τ2,…,τ k ,…,τ n}; Where n is the number of tasks; τ n represents the nth task; τ k is the kth task in the composite paradigm intention, including the following attributes, expressed as: τ k = (Location k ,Payload k ,Priority k ); Location k is the task location coordinate; Payload k For the mission load; Priority k Prioritize the tasks; Step S5, the composite paradigm intention i '' The computing system is converted into a command list CommandList; Step S6, constructing a hardware control module and a physical simulation module; The hardware control module generates a flight control command signal u(t) at time t according to the command list CommandList; the hardware control module sends the flight control command signal u(t) to the physical simulation module, and the physical simulation module processes the flight control command signal u(t) and converts the real-time status information x of the drone into real-time status information x. state (t) feeding back to the hardware control module via the network; Among them, real-time status information x state (t) including position, attitude and speed; Step S7, according to the composite paradigm intention i '' Perform deep restoration and convert it into physical parameters p; Among them, the physical parameter p includes the real-time status information x of the drone state (t), Step S8, based on the flight control command signal u(t) at time t and the UAV dynamics model, calculate the real-time state information x state (t); Step S9: The real-time status information of the drone is x state (t) Convert into visual simulation result o and output; Step S10, repeatedly executing the steps S2 to S9, performing simulation at time t+1, and obtaining the visual simulation result o at time t+1; wherein the real-time status information x of the drone is used state (t) replace the initial characterization in step S2 ' .

[0023] Step S10 has a closed-loop feedback mechanism, which means that the current output replaces the output of the previous state, so that the dynamic optimization parameters are more in line with the current task requirements. This process reflects the characteristics of a closed-loop control system, that is, the system dynamically adjusts according to real-time feedback during operation.

[0024] The intention control flow chart of the cluster UAV task collaborative semi-physical simulation method disclosed in this embodiment is as shown in the attached figure. Figure 2As shown in the figure, the operation domain is responsible for intent input and calculation, completing the whole process from the user's expected intent to solving the composite paradigm intent, including the front-end, back-end and computing system; the control domain is responsible for instruction integration and hardware instruction generation, completing the transformation of intent solution results into hardware control signals, including the computing system and hardware control module; the execution domain is responsible for task execution and simulation feedback, simulating the task behavior of the UAV in the actual environment, and outputting the visualization result intent, including the hardware control module and the physical simulation module.

[0025] Furthermore, in step S2, when the user's expected intention e and the initial representation o' are combined, the combination process is to balance the user's expected intention e and the initial representation o in a linear weighted manner. ' The influence of is expressed as: i=w e ×e+w o' ×o ' ; Among them, w e is the weight of the user's expected intention e, w o' For initial characterization ' The weight of the user's expected intention e; the weight w e and initial characterization ' The weight w o' Calculated according to the optimization goal.

[0026] Furthermore, the weight w of the user's expected intention e e and initial characterization ' The weight w o' The calculation expression is: ; The constraints are expressed as: ; Among them, Sim is the similarity function, Expressed as: ; in, for The norm of ; for The norm of ; e is the user's expected intention; o ' For initial characterization; w e is the weight of the user's desired intention e; w o' For the initial characterization ' The weight of .

[0027] Furthermore, in step S3, when the comprehensive intent input i is initially solved, the method used is: using a rule matching algorithm to map the non-standard comprehensive intent input i into the paradigm intent definition domain.

[0028] Further, in step S4, the paradigm intention i ' Perform deep calculations to generate composite paradigm intent '' Method used: Paradigm intention ' Split into tasks and generate composite paradigm intentions i '' , expressed as: ; Among them, T k is the kth task; the expression represents the composite paradigm intention i '' It is task T k The union of .

[0029] Furthermore, step S5 specifically performs the following steps: Get the environment parameter Env. The environment parameter Env is expressed as: Env = {ε1, ε2, …, ε i , …, ε m}; Among them, m is the number of environmental variables that affect the execution of UAV missions, ε i is the i-th environmental variable that affects the execution of the UAV mission (such as wind speed, temperature, terrain, etc.); ε m is the mth environmental variable that affects the execution of the UAV mission; Get the resource parameter Res. The resource parameter Res is expressed as: Res = {ρ1, ρ2, …, ρ j ,…,ρ p}; Among them, ρ j represents the available resources of the j-th UAV cluster (such as power, load capacity, communication bandwidth, etc.); ρ p represents the available resources of the p-th drone cluster; Get the mission instructions of each drone and construct the mission set Cmd of the i-th drone i , expressed as: Cmd i ={σ i1 , σ i2 , …, σ ij , …, σ im}; Among them, σ ij The mission set Cmd for the drone i The jth task of the i-th UAV in ; σ im The mission set Cmd for the dronei The mth mission of the i-th UAV; Get the i-th UAV to perform the mission σ ij Completion time T ij ; Includes flight time and execution time; Get the i-th UAV to perform the mission σ ij Resource consumption R ij ; Including power consumption and communication bandwidth occupancy; Get task σ ij The completion status Completed(σ ij ); Among them, the completion status Completed(σ ij ) is 1, indicating completion, and the completion status is Completed(σ ij ) is 0, indicating that it is not completed; According to the kth task τ in the composite paradigm intention k , environment parameters Env, and resource parameters Res to construct task feasibility constraints f feasible , expressed as: f feasible (τ k ,Env,Res)=1; The expression represents that for the kth task τ k It is feasible under the condition of environment parameter Env and resource parameter Res; According to the available resources ρ of the j-th drone cluster j , and the i-th UAV performs mission σ ij Resource consumption R ij Construct resource limit constraints, expressed as: ; The expression represents the i-th UAV performing mission σ ij Resource consumption R ij The available resources ρ of the i-th drone cluster must not exceed i ; Mathematical symbols For all i; According to the kth task τ k , and the mission set Cmd of the i-th UAV i Construct resource limit constraints, expressed as: ; Where n is the number of tasks, and the expression means that each task must be assigned to at least one UAV; According to the i-th UAV to perform the task σ ij Completion time T ij , the i-th UAV performs the mission σ ij Resource consumption Rij , Task σ ij The completion status Completed(σ ij ), task feasibility constraints f feasible , the i-th UAV performs the mission σ ij Resource consumption R ij Construct resource restriction constraints and the task set Cmd of the i-th drone i Build resource limit constraints to build command list CommandList, expressed as: ; The constraints are expressed as: ; ; ; Among them, α1, α2, and α3 are , ,as well as The weights are adjusted according to task requirements.

[0030] Furthermore, the command list CommandList, the flight control command signal u(t), and the real-time status information x of the drone involved in step S6 state (t) and the process can be expressed by the following expression: ; Among them, u prew (t) is the flight control command signal at the previous time of time t, that is, the flight control command signal at time t-1; ∆u(t) is the real-time status information x of the UAV based on the physical simulation module. state (t) Correction amount fed back to the hardware control module; K is the adjustment gain matrix, which controls the adjustment speed and amplitude of the drone; f feedback Indicates calculation of the current state x state (t) and the target state x target Deviation of (t); g is the deviation conversion function, which is used to convert the deviation f feedback Converted into correction value ∆u(t); Dynamic adjustment refers to a mechanism that adaptively adjusts the flight control command signal u(t) according to the real-time status and environmental changes during the execution of the UAV mission. The correction amount adjusted at each time is ∆u(t). The goal of dynamic adjustment is to maximize the mission objectives, reduce deviations and conflicts during the execution of the mission, and improve the stability of the system in a dynamic environment.

[0031] Furthermore, the hardware control module includes an ARM architecture computer, and the ARM architecture computer is used to parse the task instructions.

[0032] Furthermore, the weight w of the user's expected intention e e The initial value is set to 0.5; Initial Characterization ' The weight w o' The initial value is set to 0.5.

[0033] Furthermore, step S10 also iteratively updates the weight w of the user's desired intention e e and initial characterization ' The weight w o' The update method is to use gradient descent or Lagrange multiplier method to optimize the weights, the weight w of the user's expected intention e e The update direction is: ;in, express right Find partial derivatives; Initial Characterization ' The weight w o' The update direction is: ;in, express right Find the partial derivative.

[0034] Simulation Results and Analysis The mission completion rate increases with the number of drones, reaching 99% with a configuration of 100 drones.

[0035] The task decomposition time increases linearly from 50 seconds (10 aircraft) to 80 seconds (100 aircraft), maintaining high solution efficiency. The simulation execution time increases slightly with the increase in the number of drones, but does not exceed 10% at most.

[0036] Through dynamic optimization, the resource utilization rate of 20 drones is 85%, and the resource utilization rate of 100 drones is 92%.

[0037] By combining hardware control with physical simulation, the deviation of simulation results is controlled within 3%, verifying the realism of large-scale clusters.

[0038] Implementation Effect

[0039] This embodiment verifies the feasibility and superior performance of the present invention at a scale of 10 to 100 drones, and fully demonstrates the expansion capability and application potential of the system.

[0040] This embodiment introduces a dynamic intent representation mechanism for task intent, which can integrate the expected intent input by the user and the initial intent representation generated by the system in real time, and effectively respond to dynamic changes in task requirements. Improve the accuracy of task allocation and execution, and reduce resource waste and task failures caused by intention understanding deviations. This embodiment uses a closed-loop feedback mechanism to enable the simulation results to react to the adjustment and optimization of task intent. The simulation system can adaptively adjust simulation parameters during operation, significantly enhancing the intelligence level of the system. This embodiment performs physical simulation based on precise physical models and dynamic equations, and the simulation results are closer to the task execution status of a real drone cluster.

Claims

1. A method for semi-physical simulation of swarm UAV mission collaboration, characterized in that: Includes steps: Step S1, determine the user's desired intention e and initial representation o ' ; Among them, the user's expected intention e is the user's expected drone position, attitude, and speed; Initial Characterization ' This is the initial visualization simulation result; Step S2: the user's expected intention e and the initial representation o ' Combine and generate comprehensive intention input i; Step S3, the comprehensive intention is inputted into i for preliminary solution to generate a paradigm intention i ' ; Step S4, the paradigm intention i ' Perform deep calculations to generate composite paradigm intent '' ; The composite paradigm intends '' Expressed as: I '' ={τ1,τ2,…,τ k ,…,t n }; Where n is the number of tasks; τ n represents the nth task; τ k is the kth task in the composite paradigm intention, including the following attributes, expressed as: τ k = (Location k ,Payload k ,Priority k ); Location k is the task location coordinate; Payload k For the mission load; Priority k Prioritize the tasks; Step S5, the composite paradigm intention i '' Convert to command list CommandList; Step S6, constructing a hardware control module and a physical simulation module; The hardware control module generates a flight control command signal u(t) at time t according to the command list CommandList; The hardware control module sends the flight control command signal u(t) to the physical simulation module, and the physical simulation module processes the flight control command signal u(t) and converts the real-time status information x state (t) feeding back to the hardware control module via the network; Wherein, the real-time status information x state (t) including position, attitude and speed; Step S7, according to the composite paradigm intention i '' Perform deep restoration and convert it into physical parameters p; The physical parameter p includes the real-time status information x of the drone. state (t), Step S8, based on the flight control command signal u(t) at time t and the UAV dynamics model, the real-time state information x is calculated. state (t); Step S9: The real-time status information x of the drone is state (t) Convert into visual simulation result o and output; Step S10, repeatedly executing the steps S2 to S9, performing simulation at time t+1, and obtaining the visual simulation result o at time t+1; wherein the real-time status information x of the drone is used state (t) replace the initial characterization in step S2 ' .

2. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: In step S2, when the user's expected intention e and the initial representation o' are combined, the combination process is to balance the user's expected intention e and the initial representation o' in a linear weighted manner. ' The influence of is expressed as: i=w e ×e+w o' o ' ; Among them, w e is the weight of the user's desired intention e, w o' For the initial characterization ' The weight of the user's desired intention e e and the initial characterization o ' The weight w o' Calculated according to the optimization goal.

3. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 2 is characterized in that: The weight w of the user's expected intention e e and the initial characterization o ' The weight w o' The calculation expression is: ; The constraints are expressed as: ; Among them, Sim is the similarity function, Expressed as: ; in, for The norm of ; for The norm of e is the user's expected intention; o ' For initial characterization; w e is the weight of the user's desired intention e; w o' For the initial characterization ' The weight of .

4. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: In the step S3, when the comprehensive intention input i is initially resolved, the method used is: using a rule matching algorithm to map the non-standard comprehensive intention input i into the paradigm intention definition domain.

5. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: In step S4, the paradigm intent i ' Perform deep calculations to generate composite paradigm intent '' Method used: Paradigm intention ' Split into tasks and generate composite paradigm intentions i '' , expressed as: ; Among them, T k is the kth task; the expression represents the composite paradigm intention i '' It is task T k The union of .

6. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: The step S5 specifically performs the following steps: Get the environment parameter Env, which is expressed as: Env={ε1,ε2,…,ε i ,…,eh m }; Among them, m is the number of environmental variables that affect the execution of UAV missions, ε i is the i-th environmental variable that affects the execution of the UAV mission; ε m is the mth environmental variable that affects the execution of the UAV mission; Get the resource parameter Res, which is expressed as: Res={ρ1,ρ2,…,ρ j ,…,r p }; Among them, ρ j represents the available resources of the jth drone cluster; ρ p represents the available resources of the p-th drone cluster; Get the mission instructions of each drone and construct the mission set Cmd of the i-th drone i , expressed as: Cmd i ={σ i1 ,s i2 ,…,s ij ,…,s im }; Among them, σ ij The mission set Cmd for the drone i The jth mission of the i-th UAV in; σ im The mission set Cmd for the drone i The mth mission of the i-th UAV; Get the i-th UAV to perform the mission σ ij Completion time T ij ; Get the i-th UAV to perform the mission σ ij Resource consumption R ij ; Get task σ ij The completion status Completed(σ ij ); Among them, the completion status Completed(σ ij ) is 1, indicating completion. The completion status Completed(σ ij ) is 0, indicating that it is not completed; According to the kth task τ in the composite paradigm intention k , the environment parameter Env, and the resource parameter Res construct the task feasibility constraint condition f feasible , expressed as: f feasible (t) k ,Env,Res)=1; Among them, the expression represents that for the k-th task τ k It is feasible under the environment parameter Env and the resource parameter Res; According to the available resources ρ of the j-th UAV cluster j , and the i-th UAV performs mission σ ij Resource consumption R ij Construct resource limit constraints, expressed as: ; The expression represents the i-th UAV performing mission σ ij Resource consumption R ij The available resources ρ of the i-th drone cluster must not exceed i ; For all i; According to the kth task τ k , and the mission set Cmd of the i-th UAV i Construct resource limit constraints, expressed as: ; Where n is the number of tasks, and the expression means that each task must be assigned to at least one UAV; According to the i-th UAV performing the task σ ij Completion time T ij , the i-th UAV performs mission σ ij Resource consumption R ij , the task σ ij The completion status Completed(σ ij ), task feasibility constraints f feasible , the i-th UAV performs mission σ ij Resource consumption R ij Construct resource restriction constraints and the task set Cmd of the i-th UAV i Build resource limit constraints to build command list CommandList, expressed as: ; The constraints are expressed as: ； ； ； Among them, α1, α2, and α3 are , ,as well as The weight is adjusted according to the task requirements.

7. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: The command list CommandList, the flight control command signal u(t), and the real-time status information x of the drone involved in step S6 state (t) and the process can be expressed by the following expression: ; Among them, u prew (t) is the flight control command signal at the previous time of time t, that is, the flight control command signal at time t-1; ∆u(t) is the real-time status information x of the UAV based on the physical simulation module. state (t) a correction amount fed back to the hardware control module; K is the adjustment gain matrix, which controls the adjustment speed and amplitude of the drone; f feedback Indicates calculation of the current state x state (t) and the target state x target Deviation of (t); g is a deviation conversion function used to convert the deviation f feedback Converted into the correction value ∆u(t).

8. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 1 is characterized in that: The hardware control module includes an ARM architecture computer, and the ARM architecture computer is used to parse task instructions.

9. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 3 is characterized in that: The weight w of the user's expected intention e e The initial value is set to 0.5; The initial characterization ' The weight w o' The initial value is set to 0.

5.

10. The method for semi-physical simulation of swarm UAV mission collaboration according to claim 9, characterized in that: The step S10 also iteratively updates the weight w of the user's expected intention e e and the initial characterization o ' The weight w o' , the update method is to use gradient descent or Lagrange multiplier method to optimize the weight, the weight w of the user's expected intention e e The update direction is: ;in, express right Find partial derivatives; Initial Characterization ' The weight w o' The update direction is: ;in, express right Find the partial derivative.

Images