Method, device and medium for rapid assembly of a cluster of fixed-wing unmanned aerial vehicles taking into account constraints

Abstract

The present application relates to a kind of fixed-wing unmanned aerial vehicle cluster rapid assembly method, device and medium considering constraint, the formation assembly process is divided into take-off climb, hover waiting, lateral separation and formation formation four stages: by height stratified management ensures vertical safety interval;Multiple machine collaborative waiting is realized using hover waiting mechanism;Based on cut-in and cut-out strategy, the horizontal interval is accurately adjusted;Eliminate formation error by introducing master-slave collaborative speed control.Simultaneously, the composite PID controller based on ESO is designed, the disturbance estimation information of ESO is introduced and real-time compensation is carried out, and the control precision is improved.The present application improves the response speed and robustness of formation assembly process, and improves the adaptability of take-off timing change by introducing constraint, to ensure the formation safety in complex environment;Help to solve the challenges of low efficiency, poor robustness and insufficient safety in large-scale unmanned aerial vehicle cluster assembly process, and has wide application prospect.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) cooperative control technology, and in particular to a method, apparatus and medium for rapid assembly of a constrained fixed-wing UAV swarm. Background Technology

[0002] With the continuous deepening of research in modern control theory and intelligent algorithms, unmanned aerial vehicle (UAV) systems have demonstrated significant advantages in diverse application scenarios such as military reconnaissance, disaster relief, agricultural plant protection, and geographic mapping. In particular, fixed-wing UAV swarm systems, with their long endurance and high maneuverability, are irreplaceable in collaborative operations in complex environments. However, such systems face technical challenges related to multi-physics coupling in engineering practice: at the dynamic level, periodic engine vibration, nonlinear disturbances in the transmission mechanism, and aerodynamic parameter perturbations create complex mechanical-aerodynamic coupling effects; at the control level, unmodeled dynamic characteristics are superimposed with multiple uncertainties such as gusts and electromagnetic interference, resulting in strong nonlinearity and time-varying characteristics. Furthermore, fixed-wing UAV swarm systems are highly complex systems, susceptible to multi-source noise and interference during flight, including engine vibration, transmission mechanism vibration, aerodynamic parameter perturbations, unmodeled dynamics, and external disturbances.

[0003] To address the rapid formation of fixed-wing UAV swarms, scholars both domestically and internationally have proposed various formation strategies, including formation in the air after all UAVs have taken off and formation diffusion followed by convergence. Traditional UAV formation control technologies suffer from low swarm formation efficiency due to the lack of multi-stage coordination mechanisms, resulting in lengthy formation times and excessive energy consumption. Furthermore, the safety protection system has structural deficiencies, lacking a highly hierarchical control architecture, leading to a persistently high risk of flight collisions. The system's robustness is constrained by complex parameter coupling relationships, exhibiting significant inadequacy in responding to sudden environmental disturbances. Additionally, the lack of a phased, refined control strategy results in response lag and poor control accuracy during formation convergence. Summary of the Invention

[0004] Purpose of the invention: This invention addresses the challenges of low efficiency, poor robustness, and insufficient safety in existing drone swarming processes. It proposes a phased swarming strategy that considers the kinematic constraints of fixed-wing drones and a composite PID (Proportional Integral Derivative Control) swarming method based on an Extended State Observer (ESO). This effectively solves the problems of low efficiency, poor robustness, and insufficient safety in large-scale drone swarming processes.

[0005] Technical solution: The present invention provides a method for rapid assembly of a constrained fixed-wing UAV swarm, comprising the following steps:

[0006] Establish a phased assembly strategy for drone swarms; the phased assembly strategy includes:

[0007] Takeoff and climb phase: Different altitude levels are assigned to each UAV, and altitude stratification is achieved through closed-loop control;

[0008] Loitering and waiting phase: Control the drone to loiter in the designated airspace at a fixed angular rate, waiting for all drones to reach the target altitude layer;

[0009] Lateral separation phase: Adjust the projection spacing of UAVs on the side of the target formation by cutting in and cutting out control strategies, and eliminate lateral formation errors by combining master-slave cooperative control;

[0010] Formation formation stage: Construct a composite PID controller based on an extended state observer (ESO), and implement speed cooperative control based on a master-slave architecture to eliminate axial and lateral errors caused by timing uncertainties, thereby forming and maintaining the formation;

[0011] In the formation formation stage, a composite PID controller based on an extended state observer (ESO) is constructed: First, a disturbance model of the UAV position system is established, and an ESO is dynamically designed based on the UAV position to estimate the model uncertainty and the lumped disturbance caused by external disturbances; then, the control input of the UAV and the sensor measurement values ​​are input into the ESO to obtain the state estimate and lumped disturbance estimate of the fixed-wing swarm system; finally, the estimated information is input into the composite PID controller based on the extended ESO to calculate the control input.

[0012] Furthermore, the takeoff and climb phase is implemented as follows:

[0013] Based on cluster target height Minimum safe altitude difference between drones Design the expected climb altitude for each drone:

[0014]

[0015] in, The total number of drones in the cluster. Number the drone; based on the drone's current actual altitude. and expected altitude Using vertical acceleration as the control variable, a PID control method is employed to control the altitude of the UAV, thereby obtaining the required vertical acceleration. :

[0016]

[0017] in, For the number UAV altitude controller parameters.

[0018] Furthermore, the hovering and waiting phase is implemented as follows:

[0019] Based on the preset turning radius and level flight speed Calculate the target angular velocity :

[0020]

[0021] Based on the desired angular velocity Compared with the current actual angular velocity The number is obtained through PID calculation. angular acceleration of drones :

[0022]

[0023] in, It is numbered The parameters of the drone's angular velocity controller.

[0024] Furthermore, the process of implementing the lateral separation stage is as follows:

[0025] After all drones complete the hovering and waiting phase, they enter the lateral separation phase. This phase uses a five-stage approach of "cutting out hovering - lateral level flight - cutting in hovering - cutting out hovering - forward level flight" to build the expected lateral spacing of the fixed-wing swarm system.

[0026] When the drone enters the lateral separation phase after its yaw angle reaches 270° during its hovering process, it first enters the lateral level flight phase by cutting out of the hover. The duration of the lateral level flight depends on the drone's desired lateral distance. Peaceful Speed The calculation yielded: After completing the horizontal level flight mission, the UAV re-enters the yaw angle by cutting in and circling until the yaw angle becomes 0°. Then, the UAV enters the forward level flight phase by cutting out and circling, thus ending the current phase of the UAV's mission.

[0027] The process from the end of the lateral separation phase for any one drone to the end of the lateral separation phase for all drones is defined as the transition state of that drone; the motion strategy of the drone in the transition state is determined by the state of the lead drone.

[0028] Scenario 1: The lead aircraft also ends the lateral separation phase. In this case, the slave aircraft directly enters the "formation formation phase", that is, the speed closed-loop feedback control is performed with the lead aircraft's position and formation information as feedback.

[0029] Scenario 2: The lead aircraft is in the lateral separation phase. In this scenario, the lead aircraft's track angle information is known. The slave aircraft's speed is calculated using the following formula:

[0030]

[0031] in, This refers to the difference in flight time between the lead and follower aircraft due to the different target altitudes. Because of this time difference, [the following will occur]... Obtained on the axis The spacing, Similar to the first case, the target formation calculated based on the formation and aircraft sequence number is in Components on the axis.

[0032] Furthermore, the formation formation stage is implemented as follows:

[0033] After all drones have completed the lateral separation phase, all drones in the drone formation enter the formation phase. During this phase, the slave drones use speed coordination control to eliminate forward formation errors caused by takeoff timing uncertainties. When the fixed-wing swarm system enters the pursuit phase, the lead drone will accelerate to the predetermined cruise speed, while the other slave drones will adaptively adjust their speed based on the real-time distance deviation between them and the lead drone.

[0034] Record the desired formation spacing in the target formation. According to the expected formation position of the drones Current actual location of the drone Using the heading acceleration as the control variable, a PID control method is employed for UAV position control to obtain the required acceleration. :

[0035]

[0036] in, These are the parameters for the position controller.

[0037] Furthermore, the specific implementation process of constructing the composite PID controller based on the extended state observer (ESO) is as follows:

[0038] Throughout the formation process, motion status information includes horizontal velocity. Vertical velocity track angle Ground coordinate system Axis coordinates Ground coordinate system Axis coordinates Height in ground coordinate system The kinematic equations of the lead aircraft in the ground coordinate system are determined as follows:

[0039]

[0040] in, For horizontal acceleration and vertical angular velocity commands, For track angle command, The trajectory angular rate is considered; interference is taken into account. The kinematic equations of the UAV in the plane are:

[0041]

[0042] To maintain the relative distance between the lead and slave aircraft and thus ensure formation stability, a lead aircraft coordinate system is established with the lead aircraft's position as the origin, the lead aircraft's velocity direction as the positive OX axis, and the OY axis perpendicular to the OX axis and pointing to the left. In this lead aircraft coordinate system, the difference between the lead aircraft's x-axis coordinate and the slave aircraft's x-axis coordinate is defined as the longitudinal distance, and the difference between the lead aircraft's y-axis coordinate and the slave aircraft's y-axis coordinate is defined as the lateral offset. Subsequent uses... Represents the desired longitudinal distance and desired lateral offset between the lead aircraft and each slave aircraft in the lead aircraft coordinate system;

[0043] Determine the ground coordinates of each slave aircraft at each moment. Desired position of axis and Desired position of axis :

[0044]

[0045] pass Desired position of axis and Desired position of axis The formation tracking error was obtained. and the first-order dynamics of the error :

[0046]

[0047] This leads to the second-order dynamics of the formation tracking error:

[0048]

[0049] in, This is a virtual control variable, specifically in the following form:

[0050]

[0051] For measurable nonlinear terms, the specific form is as follows:

[0052]

[0053] in, This includes the lead aircraft's trajectory angular acceleration and external wind disturbances, which constitute the lumped interference of the formation system;

[0054] The complex nonlinear terms containing unknowns are treated as lumped disturbances in the system, and an extended state observer is designed. Using the observer's estimates, disturbance compensation is performed when designing the formation controller, thereby improving the formation's control performance. The resulting... Channels and The specific form of the channel expansion state observer is as follows:

[0055]

[0056] in, This is an estimate of the lumped disturbance. The parameters for the interference observer must satisfy the characteristic polynomial. For Herwitz stability;

[0057] The final formation controller design is as follows:

[0058]

[0059] in, These are the controller parameters.

[0060] The present invention provides a storage medium storing a computer program, which, when executed by at least one processor, implements the steps of the constrained rapid assembly method for fixed-wing UAV swarms as described above.

[0061] The present invention provides an electronic device comprising a memory and a processor, wherein: the memory is used to store a computer program capable of running on the processor; the processor is used to execute, when running the computer program, the steps of the constrained fixed-wing UAV swarm rapid assembly method described above.

[0062] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are as follows:

[0063] This invention decomposes the formation assembly process into four stages: takeoff and climb, hovering and waiting, lateral separation, and formation formation. It ensures vertical safety spacing through altitude-layered management; achieves multi-aircraft coordinated waiting using a hovering and waiting mechanism; precisely adjusts horizontal spacing based on a cut-in / cut-out strategy; introduces master-slave coordinated speed control to eliminate formation errors; and designs a composite PID controller based on ESO (Electronic Stability and Environment) to improve control accuracy by incorporating ESO interference estimation information and performing real-time compensation. While improving the response speed and robustness of the formation assembly process, this invention enhances adaptability to changes in takeoff timing through the introduction of constraints, thereby ensuring formation safety in complex environments. It helps address the challenges of low efficiency, poor robustness, and insufficient safety in large-scale UAV swarm assembly and has broad application prospects. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the takeoff and climb phase and the circling and waiting phase in the phased assembly strategy; where (a) is a schematic diagram of the altitude layer of each aircraft; and (b) is a schematic diagram of the circling and waiting phase.

[0065] Figure 2 This is a schematic diagram of the horizontal separation phase in the phased aggregation strategy; (a) Schematic diagram of spacing formation; (b) Schematic diagram of different spacing formation;

[0066] Figure 3 This is a schematic diagram illustrating different scenarios in the horizontal separation phase of a phased aggregation strategy;

[0067] Figure 4 This is a schematic diagram of the formation stage in a phased assembly strategy;

[0068] Figure 5 This is a diagram illustrating the formation error that occurs during the formation phase.

[0069] Figure 6 This is a schematic diagram showing the relationship between the lead aircraft coordinate system and the ground coordinate system;

[0070] Figure 7 This is a schematic diagram of the planar trajectory during formation flight;

[0071] Figure 8 The position response of each UAV under the action of the PID controller in simulation scenario 1;

[0072] Figure 9 The positional relationship between each slave unit and the master unit under the action of the PID controller in simulation scenario 1;

[0073] Figure 10 The position response of each UAV under the action of the composite PID controller in simulation scenario 1;

[0074] Figure 11The positional relationship between each slave unit and the master unit under the action of the composite PID controller in simulation scenario 1;

[0075] Figure 12 Simulation scenario 2: Position response of each UAV under the action of the PID controller;

[0076] Figure 13 The positional relationship between each slave unit and the master unit under the action of the PID controller in simulation scenario 2;

[0077] Figure 14 The position response of each UAV under the action of the composite PID controller in simulation scenario 2;

[0078] Figure 15 The positional relationship between each slave unit and the master unit under the action of the composite PID controller in simulation scenario 2;

[0079] Figure 16 The position response of each UAV under the action of the PID controller in simulation scenario 3;

[0080] Figure 17 The positional relationship between each slave unit and the master unit under the action of the PID controller in simulation scenario 3;

[0081] Figure 18 The position response of each UAV under the action of the composite PID controller in simulation scenario 3;

[0082] Figure 19 The positional relationship between each slave unit and the master unit under the action of the composite PID controller in simulation scenario 3. Detailed Implementation

[0083] The present invention will now be described in further detail with reference to the accompanying drawings.

[0084] This application discloses a method for rapid assembly of fixed-wing UAV swarms considering constraints. In this embodiment, taking a swarm of four fixed-wing UAVs as an example, the goal is to form a 45° diagonally aligned line formation with equal spacing, a target cruising altitude of 3000 meters, and a cruising speed of 70 m / s. The second UAV is designated as the lead aircraft, and the others are slave aircraft. The UAV swarm formation control method includes the following:

[0085] Establish a phased assembly strategy for drone swarms. As a complex multi-agent cooperative control system, the core function of formation assembly algorithms is to coordinate the entire process from individual takeoff to the final formation of the desired drone configuration. In practical applications, considering constraints such as airspace management and flight safety, multiple aircraft typically need to take off sequentially at staggered times. This strategy effectively reduces the risk of collisions that may result from simultaneous takeoffs and improves the safety of the entire assembly process. During the assembly process, each aircraft needs to undergo a series of strictly defined flight phases:

[0086] 1) During the takeoff and climb phase, different altitude levels are assigned to each UAV, and altitude stratification is achieved through closed-loop control;

[0087] 2) During the hovering and waiting phase, the drone is controlled to hover in the designated airspace at a fixed angular rate, waiting for all drones to reach the target altitude layer;

[0088] 3) During the lateral separation phase, the UAV's projection spacing on the side of the target formation is adjusted by the cut-in and cut-out control strategy, and the lateral formation error is eliminated by combining master-slave cooperative control;

[0089] 4) During the formation stage, the speed coordination control based on the master-slave architecture eliminates axial and lateral errors caused by timing uncertainties, thereby forming and maintaining the formation.

[0090] The actions and objectives for each stage are shown in Table 1:

[0091] Table 1. Phased Assembly Strategy

[0092]

[0093] (1) Take-off and climb phase: In a multi-UAV formation system, the hierarchical management mechanism is a key technical strategy to ensure flight safety and improve system robustness.

[0094] The safe target flight altitude level for each aircraft can be obtained using the following formula, specifically for the takeoff and climb phases, based on the cluster target altitude. Minimum safe altitude difference between drones Design the expected climb altitude for each drone:

[0095]

[0096] in, The total number of drones in the cluster. This refers to the drone numbers. In this example, there are four drones, with target altitudes of 3300m, 3100m, 2900m, and 2700m. (Based on the drones' current actual altitudes...) and expected altitude Using vertical acceleration as the control variable and employing PID control to control the altitude of the UAV, the required vertical acceleration can be obtained. :

[0097]

[0098] in, For the number UAV altitude controller parameters, .

[0099] like Figure 1 As shown in (a), based on a target formation altitude of 3000 meters, a layered structure with a 200-meter vertical interval is adopted. During the specific execution of altitude control, different target altitude layers are assigned to each aircraft according to the takeoff sequence: the first aircraft to take off is designated to reach the 3300-meter altitude layer, and the subsequent second, third, and fourth aircraft are assigned to the 3100-meter, 2900-meter, and 2700-meter altitude layers respectively. Altitude sensors are used to improve the altitude closed loop. This gradient altitude distribution strategy not only ensures a safe vertical separation but also provides ample maneuverability for subsequent formation maneuvers. During the takeoff climb phase, the aircraft adopts a constant speed climb, with a horizontal speed of 70 km / h. The climbing speed remains constant at the maximum climbing speed. Fixed and unchanging.

[0100] (2) Circling and waiting phase: Task diagram as shown Figure 1 As shown in (b), after a single aircraft completes its climb to a designated altitude, the system automatically assesses the overall situation: if other aircraft have not yet reached their respective target altitudes, the aircraft will enter a controlled circling state. In the design of the circling control parameters, based on the aircraft's performance characteristics and mission requirements, the cruise speed is set to 70 m / s, and the stability and controllability of the circling trajectory are ensured by controlling the angular rate and circling radius.

[0101] After reaching the target altitude, the drone enters a hovering and waiting phase, adjusting its turning radius according to a pre-set schedule. and level flight speed Calculate the target angular velocity :

[0102]

[0103] Based on the desired angular velocity Compared with the current actual angular velocity The number is obtained through PID calculation. angular acceleration of drones :

[0104]

[0105] in It is numbered Parameters of the drone angular velocity controller .

[0106] When the last aircraft reaches the altitude layer, since all other aircraft are already in a stable circling and waiting state, the system will directly trigger the transition to the lateral separation phase, thereby avoiding unnecessary circling and waiting processes and improving the overall mission execution efficiency.

[0107] (3) Lateral Separation Phase: After all UAVs complete the hovering and waiting phase, they enter the lateral separation phase. This phase utilizes a five-stage approach strategy of "cutting out of hovering—lateral level flight—cutting in hovering—cutting out of hovering—forward level flight" to establish the expected lateral spacing of the multi-aircraft system, laying the foundation for subsequent formation flight, such as... Figure 2 As shown in Figure (a), a schematic diagram of the UAV flight during this stage is presented. Figure 2 Figure (b) shows a schematic diagram of different spacings. Figure 2 and 3 The positions marked 1 to 3 represent the first cutout, the first cutin, and the second cutout of the drone, respectively.

[0108] When the drone enters the lateral separation phase after its yaw angle reaches 270° during hovering, such as Figure 3 As shown, the drone first enters the horizontal level flight phase by cutting out and hovering. The duration of the horizontal level flight depends on the aircraft's desired horizontal distance. Peaceful Speed The calculation yielded: After completing the horizontal level flight mission, the UAV re-enters the yaw phase by cutting in and circling until the yaw angle becomes 0°. Then, the UAV enters the forward level flight phase by cutting out and circling, thus ending the current phase of the UAV's mission.

[0109] The process from the end of the lateral separation phase for any one drone to the end of the lateral separation phase for all drones is defined as the transition state for that drone. The movement strategy of a drone in the transition state is determined by the lead drone's state:

[0110] Scenario 1: such as Figure 3 In the middle (a), the lead aircraft also ends the lateral separation phase. In this case, the slave aircraft can directly enter the "formation formation phase", that is, use the position and formation information of the lead aircraft as feedback to perform speed closed-loop feedback control.

[0111] Scenario 2: such as Figure 3 In the middle (b) phase, the lead aircraft is in the lateral separation phase. In this case, the lead aircraft's track angle information is known, and the speed of the slave aircraft can be calculated using the following formula:

[0112]

[0113] in, This refers to the difference in flight time between the lead and follower aircraft due to the different target altitudes. Because of this time difference, [the following will occur]... Obtained on the axis The spacing, Similar to the first case, the target formation calculated based on the formation and aircraft sequence number is in Components on the axis.

[0114] like Figure 4 As shown, after all UAVs complete the lateral separation phase, all UAVs in the formation enter the formation-up phase. During this phase, the slave UAVs use coordinated speed control to eliminate forward formation errors caused by takeoff timing uncertainties. Once the system enters the formation-up phase, the lead UAV accelerates to the predetermined cruise speed, while the other slave UAVs adaptively adjust their speed based on the real-time distance deviation between them and the lead UAV.

[0115] like Figure 5 As shown, due to dynamic continuity and external disturbances, formation systems are prone to errors. Let the desired formation spacing in the target formation be denoted as... According to the expected formation position of the drones Current actual location of the drone Using the heading acceleration as the control variable and employing PID control for UAV position control, the required acceleration can be obtained. :

[0116]

[0117] in, These are the position controller parameters. By implementing position loop PID control in the preceding term of the slave device, the effects of timing uncertainties are overcome, achieving the desired formation configuration and completing the entire assembly process. However, when the formation system is affected by disturbances, the formation tracking accuracy drops sharply. Therefore, it is necessary to further introduce an active anti-interference method based on a disturbance observer to improve the system's anti-interference capability.

[0118] In this way, when the cluster enters the mission flight path from the circling area, it is already in the predetermined formation state, eliminating the need for additional formation-forming stages for correction. Although this strategy introduces a longer circling waiting time compared to the direct pursuit formation method, it significantly reduces the forward flight distance required to achieve formation, making it particularly suitable for missions with limited available airspace ahead. Therefore, in specific engineering applications, a strategy trade-off can be made based on mission requirements: if the forward distance is ample and time response requirements are high, the "forward pursuit formation" method can be prioritized; if the forward distance is limited but the overall time is relatively ample, the "circling formation completion" method is more suitable. A schematic diagram of the formation flight process planar trajectory proposed in this embodiment is shown below. Figure 7 As shown.

[0119] (4) Formation formation stage: such as Figure 4 As shown, during the formation phase of a multi-UAV swarm assembly system, speed coordinated control is needed to eliminate the uncertainties caused by takeoff timing. Axial inter-machine error. This stage adopts a control architecture with the second aircraft as the lead aircraft. This design fully considers the robustness of the system and the reliability of the control. A UAV position system model is established, which includes the dynamic equations of the UAV's disturbed position under the action of control variables. Based on the dynamic equations of position and the kinematic model, an ESO is designed. The observer is used to estimate the lumped disturbances constituted by model uncertainties and external disturbances.

[0120] A composite PID controller based on ESO (Enhanced State Oscillator) is designed for a location system model. The control input from the UAV and sensor measurements are input into the ESO to obtain system state estimates and lumped disturbance estimates. Finally, the estimated information is input into the PID+ESO controller to calculate the control input.

[0121] A schematic diagram showing the relationship between the lead aircraft coordinate system and the ground coordinate system is shown below. Figure 6 As shown, during the entire formation process, the motion state information includes horizontal motion speed. Vertical velocity track angle Ground coordinate system Axis coordinates Ground coordinate system Axis coordinates Height in ground coordinate system The kinematic equations of the lead aircraft in the ground coordinate system can be determined as follows:

[0122]

[0123] in, For horizontal acceleration and vertical angular velocity commands, For track angle command, Let be the angular rate of the flight path. Consider the interference. The kinematic equations of the UAV in the plane are:

[0124]

[0125] in .

[0126] To maintain the relative distance between the lead and slave aircraft and thus ensure formation stability, a lead aircraft coordinate system needs to be defined as follows: This system uses the lead aircraft's position as the origin, the lead aircraft's velocity direction as the positive OX axis, and the OY axis perpendicular to the OX axis and pointing to the left. Within this coordinate system, the difference between the lead aircraft's x-axis coordinate and the slave aircraft's x-axis coordinate is defined as the longitudinal distance, and the difference between the lead aircraft's y-axis coordinate and the slave aircraft's y-axis coordinate is defined as the lateral offset. Representing the lead aircraft in the lead aircraft coordinate system and its number The desired longitudinal distance and desired lateral offset between the three slave aircraft are taken as follows: .

[0127] It is possible to determine the ground coordinates of each slave aircraft at each moment. Desired position of axis and Desired position of axis :

[0128]

[0129] Through the above Desired position of axis and Desired position of axis The formation tracking error was obtained. and the first-order dynamics of the error :

[0130]

[0131] This leads to the second-order dynamics of the formation tracking error:

[0132]

[0133] in, This is a virtual control variable, specifically in the following form:

[0134]

[0135] For measurable nonlinear terms, the specific form is as follows:

[0136]

[0137] The flight path angular acceleration of the lead aircraft and external wind disturbances are considered difficult to measure in the formation algorithm, meaning they cannot be fed back to each follower aircraft by the lead aircraft. Therefore, they are regarded as lumped disturbances of the formation system.

[0138] In designing the formation algorithm, complex nonlinear terms containing unknowns are treated as lumped disturbances in the system, and an ESO (Electronic Stability Occurrence) is designed. Using the observer's estimates, disturbance compensation is performed in the design of the formation controller, thereby improving the formation control performance. The resulting algorithm... Channels and The specific form of the channel's ESO is as follows:

[0139]

[0140] in This is an estimate of the lumped disturbance. The parameters for the interference observer must satisfy the characteristic polynomial. For Herwitz stability, the parameters of the interference observer are taken as follows: .

[0141] Based on the above derivation, the final design of the formation controller is as follows:

[0142]

[0143] in These are controller parameters, and their values ​​are... .

[0144] The present invention also provides a storage medium storing a computer program that, when executed by at least one processor, implements the steps of the constrained rapid assembly method for fixed-wing UAV swarms as described above.

[0145] The present invention also provides an electronic device, including a memory and a processor, wherein: the memory is used to store a computer program that can run on the processor; and the processor is used to execute, when running the computer program, the steps of the constrained fixed-wing UAV swarm rapid assembly method described above.

[0146] To illustrate the effectiveness of the constrained rapid assembly method for fixed-wing UAV swarms in this application, a UAV model was built using MATLAB R2024b for simulation verification.

[0147] The aircraft performance constraint parameters are set as shown in Table 2 below:

[0148] Table 2 Aircraft Performance Constraint Parameters

[0149]

[0150] The takeoff times of each aircraft are random but the intervals are no less than 60 seconds, and the target formation is a 45° diagonal line formation.

[0151] To simulate takeoff timing uncertainties under real-world conditions, the final values ​​of two control methods—PID and PID+ESO—were measured after formation stabilization, based on different takeoff times. , Shaft error.

[0152] Under ideal conditions in this experiment, the first unit and the second lead unit... axis, The axis spacing is Approximately 212m; Unit 3 and Unit 2 lead aircraft axis, The axis spacing is The altitude is approximately -212m; the altitude of Unit 4 and the lead unit of Unit 2... axis, The axis spacing is It is approximately -424m;

[0153] Takeoff time was set to [0 100 200 300], and simulations were performed on both methods. The simulation results are as follows. Figures 8 to 11 As shown in Table 3, the relative positions of the formations in simulation scenario 1 are the results.

[0154] Table 3. Results of relative formation positions in simulation scenario 1

[0155]

[0156] Takeoff time was set to [0 100 160 300], and simulations were performed on both methods. The simulation results are as follows. Figure 12 to

[0157] Figure 15 As shown in Table 4, the relative positions of the formations in simulation scenario two are the results.

[0158] Table 4. Results of relative positions of the formations in simulation scenario two.

[0159]

[0160] Takeoff time was set to [0 100 400 800], and simulations were performed on both methods. The simulation results are as follows. Figures 16 to 19 As shown in Table 5, the relative positions of the three formations in the simulation are as follows.

[0161] Table 5 shows the relative positions of the three formations in the simulation scenario.

[0162]

[0163] By comparing the simulation results with the target formation, and analyzing the results of three experiments with different takeoff times, Table 6 is obtained. The data in Table 6 are approximate integers of the average difference between the simulation results and the target formation.

[0164] Table 6 Results of Relative Position Error of the Formation

[0165]

[0166] It can be seen that, in the UAV swarm collaborative formation and assembly control method of this application, compared with using only PID, the PID+ESO method can complete the formation and assembly task more quickly, safely and stably.

[0167] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, such as adjusting the embedding position of the coordinate attention module, optimizing the evaluation index of the example set selection, and modifying the weight of the loss function of knowledge distillation. These improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for rapid assembly of a constrained fixed-wing UAV swarm, characterized in that, Establish a phased swarming strategy for drones; The phased assembly strategy includes: Takeoff and climb phase: Different altitude levels are assigned to each UAV, and altitude stratification is achieved through closed-loop control; Loitering and waiting phase: Control the drone to loiter in the designated airspace at a fixed angular rate, waiting for all drones to reach the target altitude layer; Lateral separation phase: Adjust the projection spacing of UAVs on the side of the target formation by cutting in and cutting out control strategies, and eliminate lateral formation errors by combining master-slave cooperative control; Formation formation stage: Construct a composite PID controller based on an extended state observer (ESO), and implement speed cooperative control based on a master-slave architecture to eliminate axial and lateral errors caused by timing uncertainties, thereby forming and maintaining the formation; In the formation formation stage, a composite PID controller based on an extended state observer (ESO) is constructed: First, a disturbance model of the UAV position system is established, and an ESO is dynamically designed based on the UAV position to estimate the model uncertainty and the lumped disturbance caused by external disturbances; then, the control input of the UAV and the sensor measurement values ​​are input into the ESO to obtain the state estimate and lumped disturbance estimate of the fixed-wing swarm system; finally, the estimated information is input into the composite PID controller based on the extended ESO to calculate the control input.

2. The method for rapid assembly of fixed-wing UAV swarms considering constraints according to claim 1, characterized in that, The takeoff and climb phase is implemented as follows: Based on cluster target height Minimum safe altitude difference between drones Design the expected climb altitude for each drone: ； in, The total number of drones in the cluster. Number the drone; based on the drone's current actual altitude. and expected altitude Using vertical acceleration as the control variable, a PID control method is employed to control the altitude of the UAV, thereby obtaining the required vertical acceleration. : ； in, For the number UAV altitude controller parameters.

3. The method for rapid assembly of fixed-wing UAV swarms considering constraints according to claim 1, characterized in that, The process of implementing the hovering and waiting phase is as follows: Based on the preset turning radius and level flight speed Calculate the target angular velocity : ； Based on the desired angular velocity Compared with the current actual angular velocity The number is obtained through PID calculation. angular acceleration of drones : ； in, It is numbered The parameters of the drone's angular velocity controller.

4. The method for rapid assembly of fixed-wing UAV swarms considering constraints according to claim 1, characterized in that, The process of the lateral separation stage is as follows: After all drones complete the hovering and waiting phase, they enter the lateral separation phase. This phase uses a five-stage approach of "cutting out hovering - lateral level flight - cutting in hovering - cutting out hovering - forward level flight" to achieve the expected lateral spacing of the fixed-wing swarm system. When the drone enters the lateral separation phase after its yaw angle reaches 270° during its hovering process, it first enters the lateral level flight phase by cutting out of the hover. The duration of the lateral level flight depends on the drone's desired lateral distance. Peaceful Speed The calculation yielded: After completing the horizontal level flight mission, the UAV re-enters the yaw angle by cutting in and circling until the yaw angle becomes 0°. Then, the UAV enters the forward level flight phase by cutting out and circling, thus ending the current phase of the UAV's mission. The process from the end of the lateral separation phase for any one drone to the end of the lateral separation phase for all drones is defined as the transition state of that drone; the motion strategy of the drone in the transition state is determined by the state of the lead drone. Scenario 1: The lead aircraft also ends the lateral separation phase. In this case, the slave aircraft directly enters the "formation formation phase", that is, the speed closed-loop feedback control is performed with the lead aircraft's position and formation information as feedback. Scenario 2: The lead aircraft is in the lateral separation phase. In this scenario, the lead aircraft's track angle information is known. The slave aircraft's speed is calculated using the following formula: ； in, This refers to the difference in flight time between the lead and follower aircraft due to the different target altitudes. Because of this time difference, [the following will occur]... Obtained on the axis The spacing, Similar to the first case, the target formation calculated based on the formation and aircraft sequence number is in Components on the axis.

5. The method for rapid assembly of fixed-wing UAV swarms considering constraints according to claim 1, characterized in that, The formation formation phase is implemented as follows: After all drones have completed the lateral separation phase, all drones in the drone formation enter the formation phase. During this phase, the slave drones use speed coordination control to eliminate forward formation errors caused by takeoff timing uncertainties. When the fixed-wing swarm system enters the pursuit phase, the lead drone will accelerate to the predetermined cruise speed, while the other slave drones will adaptively adjust their speed based on the real-time distance deviation between them and the lead drone. Record the desired formation spacing in the target formation. According to the expected formation position of the drones Current actual location of the drone Using the heading acceleration as the control variable, a PID control method is employed for UAV position control to obtain the required acceleration. : ； in, These are the parameters for the position controller.

6. The method for rapid assembly of fixed-wing UAV swarms considering constraints according to claim 1, characterized in that, The specific implementation process of constructing the composite PID controller based on the extended state observer (ESO) is as follows: Throughout the formation process, motion status information includes horizontal velocity. Vertical velocity track angle Ground coordinate system Axis coordinates Ground coordinate system Axis coordinates Height in ground coordinate system The kinematic equations of the lead aircraft in the ground coordinate system are determined as follows: ； in, For horizontal acceleration and vertical angular velocity commands, For track angle command, The trajectory angular rate is considered; interference is taken into account. The kinematic equations of the UAV in the plane are: ； To maintain the relative distance between the lead and slave aircraft and thus ensure formation stability, a lead aircraft coordinate system is established with the lead aircraft's position as the origin, the lead aircraft's velocity direction as the positive OX axis, and the OY axis perpendicular to the OX axis and pointing to the left. In this lead aircraft coordinate system, the difference between the lead aircraft's x-axis coordinate and the slave aircraft's x-axis coordinate is defined as the longitudinal distance, and the difference between the lead aircraft's y-axis coordinate and the slave aircraft's y-axis coordinate is defined as the lateral offset. Subsequent uses... Represents the desired longitudinal distance and desired lateral offset between the lead aircraft and each slave aircraft in the lead aircraft coordinate system; Determine the ground coordinates of each slave aircraft at each moment. Desired position of axis and Desired position of axis : ； pass Desired position of axis and Desired position of axis The formation tracking error was obtained. and the first-order dynamics of the error : ； This leads to the second-order dynamics of the formation tracking error: ； in, This is a virtual control variable, specifically in the following form: ； For measurable nonlinear terms, the specific form is as follows: ； in, This includes the lead aircraft's trajectory angular acceleration and external wind disturbances, which constitute the lumped interference of the formation system; The complex nonlinear terms containing unknowns are treated as lumped disturbances in the system, and an extended state observer is designed. Using the observer's estimates, disturbance compensation is performed when designing the formation controller, thereby improving the formation's control performance. The resulting... Channels and The specific form of the channel expansion state observer is as follows: ； in, This is an estimate of the lumped disturbance. The parameters for the interference observer must satisfy the characteristic polynomial. For Herwitz stability; The final formation controller design is as follows: ； in, These are the controller parameters.

7. A storage medium, characterized in that, The storage medium stores a computer program that, when executed by at least one processor, implements the steps of the constrained rapid assembly method for fixed-wing UAV swarms as described in any one of claims 1 to 6.

8. An electronic device, characterized in that, Includes memory and processor, wherein: Memory is used to store computer programs that can run on a processor; A processor, configured to, while running the computer program, perform the steps of the constrained rapid assembly method for fixed-wing unmanned aerial vehicle swarms as described in any one of claims 1 to 6.

Images