Rotor unmanned aerial vehicle cluster real-time simulation method

Abstract

The invention discloses a rotor unmanned aerial vehicle cluster real-time simulation method. According to the technology, a regional sub-cluster edge distributed interaction rule is designed, and the problems of overlarge simulation communication pressure and cluster scale amplification are solved; designing a cluster grouping hardware simulation environment and a software function according to a target simulation scene requirement; a clock synchronization method for master clock synchronization and sub-clock interval verification is provided, and the clock synchronization problem in the cluster simulation process is solved. Through a database storage mechanism in which a cache array and an inline array are fused, data partition management and data interaction of multiple unmanned aerial vehicles are realized; and a sub-task operation mechanism is planned according to frequencies and execution points, so that multi-sub-task cooperative operation is realized, and the requirement of large-scale cluster unmanned aerial vehicle simulation is met.

Description

Technical Field

[0001] The invention relates to a real-time simulation method for a rotor UAV cluster, mainly to a cluster real-time simulation method, and belongs to the technical field of real-time simulation. Background Art

[0002] In recent years, drones have been used more and more widely, playing an important role in military, civil, firefighting and other fields. With the change of application environment and the diversification of tasks, drones have gradually transformed from single-machine platforms to multi-machine cluster operations. Drone cluster flight requires reasonable design of cluster control algorithms and repeated testing and optimization to ensure that drone clusters can perfectly complete flight missions. Therefore, the study of real-time simulation systems is of great significance for high-confidence verification of drone cluster algorithms.

[0003] At present, there are two common cluster control methods: distributed and centralized. In the centralized mode, the information exchange and control adjustment of the cluster drones are carried out through the central drone, and the cluster has a high degree of information acquisition. However, the centralized mode has many defects. If the central drone is paralyzed or the information is interrupted, the entire cluster will lose control. In the centralized mode, the amount of information interaction increases geometrically with the expansion of the group size, which is very unfavorable for the expansion of the cluster. In contrast, the distributed mode has many advantages. The drones exchange information with their neighbors around themselves and make autonomous decisions. The cluster has good scalability.

[0004] Relevant teams at home and abroad have carried out a lot of research work on the construction of verification systems for cluster control methods. Considering the needs of UAV cluster confrontation evaluation and the requirements of "high efficiency and low cost" for experiments, a real-time simulation system running in Windows environment is built. In this paper, the particle model is used as the cluster control object, and the UAV cluster performs distributed interaction and performs route tracking flight under the guidance of the virtual pilot speed. The invention designs the distributed interaction rules at the edge of the regional sub-cluster to solve the problems of excessive simulation communication pressure and cluster scale expansion; according to the requirements of the target simulation scenario, the cluster grouping hardware simulation environment and software functions are designed, and software development is carried out through rapid prototyping; based on the clock synchronization method of main clock synchronization and sub-clock interval verification, the clock consistency of different programs is achieved; the method of combining cache array and inline array is used for data storage, and the data area is partitioned and managed, realizing the data storage and high-efficiency indexing of the simulation system; according to the requirements of the simulation system, the sub-task operation planning is carried out based on the frequency division and execution point division method to ensure the rational execution of the task. Finally, a comprehensive simulation verification is carried out, and the results show that the real-time simulation system can verify the cluster control algorithm, and the performance of the simulation system meets the real-time requirements. Summary of the invention

[0005] In view of the above technical problems, the purpose of the present invention is to design a cluster real-time simulation technology, which can realize high-precision real-time simulation of cluster algorithms. The method includes: designing regional sub-cluster edge distributed interaction rules to solve the problems of excessive simulation communication pressure and cluster scale expansion; designing cluster grouping hardware simulation environment and software functions according to the target simulation scenario requirements; based on the clock synchronization method of main clock synchronization and sub-clock interval verification, the clocks of different programs are consistent; using a method of merging cache arrays and inline arrays for data storage, and partitioning the data area to realize simulation system data storage and high-efficiency indexing; according to the simulation system requirements, subtask operation planning is carried out based on the frequency division and execution point division method to ensure the rational execution of tasks.

[0006] A real-time simulation method for a rotorcraft drone cluster, the method comprising the following steps:

[0007] Step 1: Design the distributed interaction rules at the edge of regional sub-clusters;

[0008] Step 2: Design the cluster grouping hardware simulation environment and software function implementation according to the target simulation scenario requirements, and develop software through rapid prototyping;

[0009] Step 3: Design a clock synchronization method for main clock synchronization and sub-clock interval verification according to the software function to solve the problem of clock consistency between different programs;

[0010] Step 4: Design a data management mechanism based on array partitioning according to the software functions to solve the problem of drone data storage and data management;

[0011] Step 5: Design a task operation allocation method based on frequency and time points according to the software functions to achieve reasonable planning of task operation.

[0012] Furthermore, the step 1 is specifically as follows:

[0013] The cluster is divided into multiple sub-cluster modules according to the sub-cluster division rule. The sub-cluster division rule is as follows: taking the cluster center as the symmetry center, the surrounding area is divided into squares;

[0014] Each sub-cluster module performs distributed information sharing according to edge key information, and each sub-cluster module performs distributed interaction. The sub-cluster only interacts with the neighboring sub-cluster. During the simulation process, the sub-cluster only shares edge drone information.

[0015] Furthermore, the step 2 is specifically as follows:

[0016] Each sub-cluster module is simulated by the corresponding sub-cluster computer, and the central node computer acts as a central server to receive and share the data of each drone in the sub-cluster and the host computer commands; the host computer is used for simulation parameter display, cluster trajectory display and simulation process control.

[0017] The software function implementation includes a main program and a sub-cluster program. The main program is used to control the navigation tasks of the cluster, data transfer and storage records, main clock scale and host computer command broadcast; the sub-cluster program is responsible for cluster drone simulation and cluster control tasks.

[0018] Rapid prototyping technology is used for software development, and the Simulink model and manual code are cross-compiled to generate an application that can run in real time on a PC, realizing real-time simulation of UAV clusters.

[0019] Furthermore, the step 3 is specifically as follows:

[0020] The clock feedback value inside Windows is used for clock counting. The formula is:

[0021] t clock =(C start -C end ) / f cpu

[0022] Where: f cpu The computer's internal clock frequency, C start is the current clock count, C end Count the clock for the next beat;

[0023] The clock formula of each sub-cluster computer is as follows:

[0024] T main =k flag_start (T main +T scale )

[0025]

[0026] Where: T main is the master clock, k flag_start It is the simulation start mark, T scale is the clock base scale, T sub is the sub-clock, k flag_syn It is the clock synchronization flag. is the current clock count value, It is the clock count value of the previous step.

[0027] Furthermore, the step 4 is specifically as follows:

[0028] A data storage method combining a cache array and an inline array is used, wherein the cache array is responsible for data transmission and reception, and the inline array is responsible for data communication between the manual code and the internal data of the Simulink model;

[0029] The cache array and the inline array are both one-dimensional arrays. The data inside the array is stored in the data area according to the drone number and signal type number. The data positioning formula is:

[0030]

[0031] Where: U l is the length of the drone data area, S l is the maximum length of the signal data, and n is the sequence number of the data in the signal data area;

[0032] The data management mechanism has a preset flag to distinguish the read and write functions of the cache array and the inline array.

[0033] Furthermore, the step 5 is specifically as follows:

[0034] According to the importance of the task and the way the task is triggered, the task types are divided into three types: high-frequency tasks, low-frequency tasks, and triggered tasks;

[0035] The high-frequency tasks are basic tasks, including model data writing, model solving, model data reading, drone status sending, navigation information publishing, and drone status information sending;

[0036] The low-frequency tasks are tasks that assist the simulation system, including data txt recording and host computer data sending;

[0037] The trigger task refers to a task that is triggered when conditions are met, including clock synchronization and simulation command modules;

[0038] According to the task accuracy requirements, high-frequency tasks are executed at a high frequency, and low-frequency tasks are executed at a low frequency. Tasks at the same frequency are assigned different execution time points according to the execution time requirements.

[0039] Beneficial effects:

[0040] A real-time simulation method for a rotorcraft drone cluster of the present invention designs distributed interaction rules at the edge of regional sub-clusters to solve the problems of excessive simulation communication pressure and cluster scale expansion; designs cluster grouping hardware simulation environment and software functions according to the requirements of the target simulation scenario; realizes the consistency of clocks of different programs based on the clock synchronization method of master clock synchronization and sub-clock interval verification; adopts a method of integrating cache array and inline array for data storage, and partitions the data area to realize simulation system data storage and high-efficiency indexing; according to the requirements of the simulation system, sub-task operation planning is carried out based on the frequency division and execution point division method to ensure the rational execution of tasks. Finally, comprehensive simulation verification shows that the real-time simulation system can meet the performance requirements of cluster simulation.

[0041] Description of the attached figures and tables

[0042] Figure 1 A schematic diagram of the speed-guided cluster route tracking provided by the present invention;

[0043] Figure 2 A schematic diagram of the hardware network structure of the simulation system provided by the present invention;

[0044] Figure 3 A schematic diagram of the main program function design provided by the present invention;

[0045] Figure 4 A schematic diagram of the functional design of the sub-cluster program provided by the present invention;

[0046] Figure 5 A schematic diagram of rapid prototyping provided by the present invention;

[0047] Figure 6 A schematic diagram of a data read-write protection mechanism provided by the present invention;

[0048] Figure 7 A schematic diagram of data zone partition management provided by the present invention;

[0049] Figure 8 A schematic diagram of sub-cluster boundary points provided by the present invention;

[0050] Fig. 9 The cluster route tracking flight trajectory diagram provided by the present invention;

[0051] Fig.10 A curve diagram of cluster order parameter variation provided by the present invention;

[0052] Fig.11 A curve diagram of the minimum neighbor distance of the drone cluster provided by the present invention;

[0053] Fig.12 A clock delay curve diagram provided by the present invention;

[0054] Fig.13 A communication delay curve diagram provided by the present invention;

[0055] Table 1 is the main program task frequency allocation parameters provided by the present invention;

[0056] Table 2 is a sub-cluster program task frequency allocation parameter diagram provided by the present invention;

[0057] Table 3 shows the mission waypoint parameters provided by the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further described below in conjunction with the accompanying drawings.

[0059] The real-time simulation technology of rotorcraft drone clusters needs to solve the problems of cluster scale expansion, software and hardware design and development, clock consistency, data storage management, and task operation planning. The present invention designs distributed interaction rules at the edge of regional sub-clusters to solve the problems of excessive simulation communication pressure and cluster scale expansion; according to the requirements of the target simulation scenario, the cluster grouping hardware simulation environment and software functions are designed to achieve rational division of system modules; a clock synchronization method based on master clock synchronization and sub-clock interval verification is adopted to achieve clock consistency of different programs; a method combining cache arrays and inline arrays is used for data storage, and the data area is partitioned and managed to achieve data storage and efficient indexing of the simulation system; according to the requirements of the simulation system, sub-task operation planning is carried out based on the frequency division and execution point division method to ensure the rational execution of tasks.

[0060] Furthermore, the method comprises:

[0061] Step 1: Design distributed interaction rules at the edge of regional sub-clusters to solve the problems of excessive simulation communication pressure and cluster scale expansion;

[0062] Step 2: According to the target simulation scenario requirements in step 1, design the cluster grouping hardware simulation environment and software functions, and develop the software through rapid prototyping;

[0063] Step 3: According to the software function requirements designed in step 2, a clock synchronization method for checking the interval between the main clock synchronization and the sub-clock is designed to solve the problem of clock consistency between different programs;

[0064] Step 4: Design a data management mechanism based on array partitioning according to the software functional requirements designed in step 2 to solve the problem of drone data storage and data management;

[0065] Step 5: Design a task operation allocation method based on frequency and time points according to the software functional requirements designed in step 2 to achieve reasonable planning of task operation.

[0066] The following is a detailed description of the above solution with reference to specific cases:

[0067] Step S1: Design distributed interaction rules at the edge of regional sub-clusters to solve the problems of excessive simulation communication pressure and cluster scale expansion.

[0068] First, cluster control uses a speed-consistent cluster control method and a speed-guided route tracking method:

[0069] The double integral particle model is:

[0070]

[0071] Where: p m (t), v m (t) are the position and velocity vector of the UAV respectively; u m (t) is the control input vector of the UAV.

[0072] The cluster control goal is to keep the positions of drones at a safe distance and at the same speed. The formula is:

[0073]

[0074] Where: d min is the minimum safe distance of the drone, d max is the communication distance between UAVs.

[0075] According to the cluster aggregation, alignment and attraction behavior rules, a three-layer action area is designed and the upper limit of interaction neighbors is limited to obtain the neighbors for drone interaction.

[0076] Assume that the drone communication area R c , the area around the drone is divided into three layers of areas according to the circle, namely, the exclusion area R rep , Alignment area R ali , and attraction region R att , the area radius relationship is as follows:

[0077] R rep <R ali <R att =R c

[0078] In actual flight simulation, the drone only needs to interact with its nearest neighbors. Here, drones with a fixed upper limit on the number of neighbors are selected as the interaction objects. In the separation region, the neighbor set of the region that attracts the region can be expressed as

[0079] for:

[0080]

[0081] The expected speed of the drone is designed according to the three rules of repulsion, alignment and attraction of the cluster:

[0082] v d (t) = v rep (t)+v ali (t)+v att (t)+v nav (t)

[0083] In the formula, v rep (t), v ali (t), v att (t) and v nav (t) represents the repulsion velocity term, alignment term, attraction velocity term and guidance velocity term respectively; the guidance velocity plays the role of guiding the overall movement direction of the cluster.

[0084] Specific formulas for each speed term:

[0085]

[0086] Where: k rep , k att , k ali They are the separation term, aggregation term and speed coordination term coefficients between UAVs, and the weights of the corresponding terms can be adjusted.

[0087] A cluster route tracking method using virtual pilot speed guidance. Figure 1 As shown, the virtual pilot aircraft tracks the position of the route, and the drone cluster tracks the speed of the virtual pilot aircraft. In order to prevent the drone cluster from deviating from the route, the route deviation is fed back to the virtual pilot aircraft, and the virtual pilot aircraft corrects the guidance speed to ensure that the drone cluster accurately tracks the route.

[0088] The above speed consistency cluster control method and speed guidance route tracking method are used to achieve cluster coordination and flight along the route. The following is the design of regional sub-cluster edge distributed interaction rules:

[0089] In order to alleviate the computational pressure that increases exponentially with the expansion of the cluster size, the cluster is divided into multiple sub-cluster modules. Sub-cluster division rules: Take the cluster center as the symmetry center, divide the surrounding area into squares, and reasonably plan the number of squares and the length of the area to divide the drone cluster into multiple sub-cluster modules. The drones in the corresponding area are used as simulated drones of the corresponding sub-cluster.

[0090] Subclusters share distributed information according to edge key information. In order to solve the problem of excessive communication pressure caused by sharing all drone information between subclusters, subclusters interact in a distributed manner, and subclusters only interact with neighboring subclusters. During the simulation process, the edge drones of the subcluster have a greater impact on the neighboring drones, and the inner drones of the subcluster have little impact on the neighboring subclusters. Therefore, during the simulation process, subclusters only share edge drone information.

[0091] Step S2: design the sub-cluster simulation system hardware structure and software functions, and perform software development through rapid prototyping.

[0092] In order to alleviate the computing pressure that increases exponentially with the expansion of cluster size, the cluster is divided into multiple sub-cluster modules. The hardware structure is designed according to the performance requirements of the sub-cluster real-time simulation system. The hardware equipment of the simulation system is as follows: Figure 2 As shown in the figure, the system hardware consists of a host computer and a simulation computer. The simulation program and the host computer program are developed by the development computer, LabVIEW realizes the host computer interface design, and simulation programming software such as MATLAB, Visual Studio and Source Insight are used to develop the simulation program. The host computer is responsible for simulation process control, full process simulation parameters and trajectory display. The simulation computer consists of a central node computer and a sub-node computer. The central node computer runs the main program, acts as a central server, and is responsible for data transfer sharing, clock scale and cluster navigation tasks. The sub-node computer runs the sub-cluster program to perform sub-cluster simulation.

[0093] Design the sub-cluster simulation system software functions according to the hardware system architecture: main program and sub-cluster program. The main program functions are divided into Figure 3 As shown in the figure, the main program is responsible for the navigation task of the cluster algorithmically, and is responsible for data transfer and storage records, main clock scale and host computer command broadcasting. The cluster navigation algorithm is implemented based on the Simulink model, and the manual code is used to implement the simulation system strategy design, including clock module, communication module and data management module. The sub-cluster function is divided as follows Figure 4 As shown in the figure, the sub-cluster program is responsible for simulating some UAVs in the cluster. The cluster control algorithm and particle UAV model solution are implemented through the Simulink model. The manual code is used to implement the clock module, communication module and data management module, etc.

[0094] Use rapid prototyping techniques for software development, such as Figure 5 As shown in the figure, the Simulink model and manual code are cross-compiled to generate an application that can run in real time on a PC to realize real-time simulation of UAV clusters.

[0095] Step S3: According to the software function requirements designed in step 2, a clock synchronization method for main clock synchronization and sub-clock interval verification is designed to solve the problem of clock consistency between different programs.

[0096] The clock feedback value inside Windows is used for clock counting. The formula is:

[0097] t clock =(C start -C end ) / f cpu

[0098] Where: f cpu The computer's internal clock frequency, C start is the current clock count, C end Count the clock for the next beat.

[0099] By counting to meet the local clock requirements, different modules need to check the clocks to achieve clock synchronization. Clock synchronization is performed by master clock synchronization and sub-clock interval verification. The master clock module is selected as the clock reference. The master clock module will send a clock synchronization signal to the remaining sub-clock modules at the beginning of the simulation. After the initial clock synchronization is achieved, the clock verification phase is entered, and the clock synchronization task of the master clock module enters the conditional judgment mode. The verification clock information is sent by the sub-clock module at an interval of T. The master clock receives the verification clock and makes a time difference judgment. When the time difference is greater than the time difference margin, the master clock sends clock synchronization information to eliminate the clock error. The clock formulas for each module are as follows:

[0100] T main =k flag_start (T main +T scale )

[0101]

[0102] Where: T main is the master clock, k flag_start It is the simulation start mark, T scale is the clock base scale, T sub is the sub-clock, k flag_syn It is the clock synchronization flag. is the current clock count value, It is the clock count value of the previous step.

[0103] Step S4: Design a data management mechanism based on array partitioning according to the software functional requirements designed in step 2 to solve the problems of drone data storage and data management.

[0104] The data storage method of cache array plus inline array is adopted. The cache array is responsible for data transmission and reception, and the inline array is responsible for data communication between manual code and Simulink model. The two use mutually exclusive flags to distinguish between reading and writing to avoid reading and writing operations on the same array at the same time. Figure 6 As shown, when the flag bit is 1, the cache array is writable but not readable, and the inline array is readable but not writable. When the flag bit is 0, the cache array is readable but not writable, and the inline array is writable but not readable.

[0105] Adopt partition management data management method. Figure 7 As shown, the cache array and the inline array are both one-dimensional arrays, and the data inside the array is stored in the data area according to the drone number and signal type number. The data positioning formula is:

[0106]

[0107] Where: U l is the length of the drone data area, S l is the maximum length of the signal data, and n is the sequence number of the data in the signal data area.

[0108] Step S5: design a task operation allocation method based on frequency and time points according to the software functional requirements designed in step 2 to achieve reasonable planning of task operation.

[0109] According to the importance of the task and the task triggering method, the task types are divided into three types: high-frequency tasks, low-frequency tasks and triggered tasks. High-frequency tasks are basic tasks that have a greater impact on the real-time performance of the simulation system, including model data writing, model solving, model data reading, drone status sending, navigation information publishing, drone status information sending, etc.; low-frequency tasks are tasks that play an auxiliary role in the simulation system, including data txt recording, host computer data sending, etc.; trigger tasks refer to tasks that are triggered when conditions are met, including clock synchronization, simulation command modules, etc.

[0110] According to the task accuracy requirements, high-precision tasks are executed at a high frequency, and low-precision tasks are executed at a low frequency. Tasks at the same frequency are assigned different execution time points according to the execution time requirements. The main program task running distribution is shown in Table 1, and the sub-cluster program task running distribution is shown in Table 2.

[0111] Table 1 Main program task frequency distribution

[0112]

[0113] Table 2 Frequency distribution of sub-cluster program tasks

[0114]

[0115] Comprehensive simulation effect analysis:

[0116] The real-time simulation technology of rotary-wing UAVs proposed in the present invention ensures good real-time performance while verifying the algorithm for the UAV cluster. The experiment assumes a total of 100 UAVs. The main program performs clock synchronization and simulation command forwarding. Two subroutines simulate 50 UAVs. The simulation time is 350s. The maximum speed of the UAV is 4m / s. The 100 UAVs are initially distributed in the area of ​​[0,200m]×[0,200m]. The initial speed is 0 and the initial guidance speed is [3,3]m / s. The UAV has a volume of 1 cubic meter and an initial interval of about 20m from its neighbors. The radius of the exclusion area is 10m, the radius of the alignment area is 20m, the radius of the attraction area is 30m, and the communication radius is 30m. The initial position of the cluster and the boundary points of the sub-cluster are as follows: Figure 8 As shown in the figure, the upper part is subcluster 1, and the lower part is subcluster 2. The subclusters only exchange boundary neighbor information. The UAV cluster completes the assembly and coordinated flight under the action of the initial guidance speed and cluster control. When the cluster order parameter reaches 0.95 and moves to the route range, it enters the cluster route flight stage. A total of 7 mission waypoints are preset. The first waypoint is the mission entry waypoint, and the flight mission ends at the sixth waypoint. The waypoint information is shown in Table 3. The initial position of the drone is scattered and disordered, and the initial speed is 0. The cluster flight trajectory is shown in Fig. 9 As shown in Figure 2, it can be seen that the cluster control algorithm can well control the cluster to be consistent and orderly, and the route tracking method can accurately track the waypoints. Fig.10 As shown in the figure, under the initial guidance speed control, the cluster began to adjust the spacing and speed, and the order parameter gradually increased. At 20 seconds, the order parameter reached the critical point of 0.95, meeting the consistency requirement. In the subsequent process, the cluster order parameter continued to increase and approached 1, indicating that the cluster had a good degree of stability during flight. The minimum spacing between drones during the flight process is shown in the figure. Fig.11 As shown in the figure, the safe distance between drones is set to 10m. In the initial stage, the distance between drones is relatively close. Under cluster control, the distance between drones gradually increases. The minimum steady-state distance between drone clusters is 15m, which is greater than the safe distance. This shows that the position control effect of the drone flight process is good. The clock delay of the real-time simulation system is shown in Fig.12 As shown in Figure 2, when the system is running stably, the clock delay is maintained at a time difference margin of 5ms, which shows the effectiveness of the clock synchronization method. Fig.13 As shown, the communication delay is large at the beginning of the simulation, with a peak value of 29ms. When the system is running stably, the communication delay is 5ms on average, and the acceptable communication delay is 30ms, which is acceptable.

[0117] Table 3 Mission waypoints

[0118]

[0119]

[0120] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A real-time simulation method for a rotorcraft drone cluster, characterized in that: The method comprises the following steps: Step 1: Design the distributed interaction rules at the edge of regional sub-clusters; Step 2: Design the cluster grouping hardware simulation environment and software function implementation according to the target simulation scenario requirements, and develop software through rapid prototyping; Step 3: Design a clock synchronization method for main clock synchronization and sub-clock interval verification according to the software function to solve the problem of clock consistency between different programs; Step 4: Design a data management mechanism based on array partitioning according to the software functions to solve the problem of drone data storage and data management; Step 5: Design a task operation allocation method based on frequency and time points according to the software functions to achieve reasonable planning of task operation.

2. A real-time simulation method for a rotorcraft drone cluster according to claim 1, characterized in that: The step 1 is specifically as follows: The cluster is divided into multiple sub-cluster modules according to the sub-cluster division rule. The sub-cluster division rule is as follows: taking the cluster center as the symmetry center, the surrounding area is divided into squares; Each sub-cluster module performs distributed information sharing according to edge key information, and each sub-cluster module performs distributed interaction. The sub-cluster only interacts with the neighboring sub-cluster. During the simulation process, the sub-cluster only shares edge drone information.

3. The real-time simulation method of a rotorcraft drone cluster according to claim 1, characterized in that: The step 2 is specifically as follows: Each sub-cluster module is simulated by the corresponding sub-cluster computer, and the central node computer acts as a central server to receive and share the data of each drone in the sub-cluster and the host computer commands; the host computer is used for simulation parameter display, cluster trajectory display and simulation process control; The software function implementation includes a main program and a sub-cluster program. The main program is used to control the navigation task of the cluster, data transfer and storage records, main clock scale and host computer command broadcast; the sub-cluster program is responsible for cluster drone simulation and cluster control tasks; Rapid prototyping technology is used for software development, and the Simulink model and manual code are cross-compiled to generate an application that can run in real time on a PC, realizing real-time simulation of UAV clusters.

4. The real-time simulation method for a rotorcraft drone cluster according to claim 1, characterized in that: The step 3 is specifically as follows: The clock feedback value inside Windows is used for clock counting. The formula is: t clock =(C start -C end ) / f cpu Where: f cpu The computer's internal clock frequency, C start is the current clock count, C end Count the clock for the next beat; The clock formula of each sub-cluster computer is as follows: T main =k flag_start (T main +T scale ) Where: T main is the master clock, k flag_start It is the simulation start mark, T scale is the clock base scale, T sub is the sub-clock, k flag_syn It is the clock synchronization flag. is the current clock count value, It is the clock count value of the previous step.

5. The real-time simulation method for a rotorcraft drone cluster according to claim 1, characterized in that: The step 4 is specifically as follows: A data storage method combining a cache array and an inline array is used, wherein the cache array is responsible for data transmission and reception, and the inline array is responsible for data communication between the manual code and the internal data of the Simulink model; The cache array and the inline array are both one-dimensional arrays. The data inside the array is stored in the data area according to the drone number and signal type number. The data positioning formula is: Where: U l is the length of the drone data area, S l is the maximum length of the signal data, and n is the sequence number of the data in the signal data area; The data management mechanism has a preset flag to distinguish the read and write functions of the cache array and the inline array.

6. The real-time simulation method of a rotorcraft UAV cluster according to claim 1, characterized in that: The step 5 is specifically as follows: According to the importance of the task and the way the task is triggered, the task types are divided into three types: high-frequency tasks, low-frequency tasks, and triggered tasks; The high-frequency tasks are basic tasks, including model data writing, model solving, model data reading, drone status sending, navigation information publishing, and drone status information sending; The low-frequency tasks are tasks that assist the simulation system, including data txt recording and host computer data sending; The trigger task refers to a task that is triggered when conditions are met, including clock synchronization and simulation command modules; According to the task accuracy requirements, high-frequency tasks are executed at a high frequency, and low-frequency tasks are executed at a low frequency. Tasks at the same frequency are assigned different execution time points according to the execution time requirements.

Images