High-altitude emergency rescue system and high-speed impact robust control method based on multi-unmanned aerial vehicle cooperative flexible network rescue

Abstract

The invention provides a high-altitude emergency rescue system and a high-speed impact robust control method based on multi-unmanned aerial vehicle cooperative flexible network rescue, and the system comprises a cooperative formation composed of N unmanned aerial vehicles, and each unmanned aerial vehicle is provided with a flight controller; the flexible net is hung and connected to the N unmanned aerial vehicles through N ropes and is used for intercepting a target object; the flight controller of each unmanned aerial vehicle comprises a nonlinear interference observer and a robust controller; the nonlinear interference observer is used for estimating transient pulse type disturbance and disturbance torque in real time in an impact event that a target object is intercepted by the flexible net, and generating an estimated interference signal; the robust controller is used for generating a control instruction to drive an execution mechanism of the unmanned aerial vehicle. According to the method, the defects of a traditional linear model and an undamped model are overcome, so that model prediction better conforms to the real physical characteristics of the high-molecular flexible network, and wrong prediction of system oscillation is avoided.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, specifically to a high-altitude emergency rescue system based on a multi-UAV cooperative flexible network rescue and a high-speed impact robust control method. Background Technology

[0002] High-altitude emergency rescue, especially in complex urban environments such as high-rise buildings (e.g., 50-meter-high buildings), poses a significant challenge to traditional rescue methods. Traditional ground rescue equipment (such as fire ladders) has limited height, while helicopter rescue is constrained by airflow, space, and response time.

[0003] In recent years, the development of drone technology has provided new possibilities for high-altitude rescue. Existing technologies include solutions for using drones to deliver rescue supplies or conduct reconnaissance. However, there is a fundamental technological gap in the direct interception and capture of people or objects falling from heights. Current technologies often assess drone capture capabilities based on simplified linear models or only consider the static gravity load of the target object. This assessment severely underestimates the real physical challenges of high-altitude interception. For example, a 70 kg adult falling from a height of 10 meters will reach a speed of [missing value] upon impact. During capture, the system does not withstand a static weight of 70 kg, but rather a massive impact load converted from kinetic energy.

[0004] Existing technologies, when dealing with multi-UAV cooperative flexible net interception systems, generally employ lumped mass models or static energy conservation models for force analysis. These models assume that the tension of the net cable is spatially uniformly distributed at the moment of impact and treat the UAVs as ideal force sources or fixed anchor points. However, based on in-depth research in continuum mechanics and wave dynamics, the inventors have discovered that in high-speed impacts (velocity... Under extreme operating conditions, the above assumptions are not only severely distorted, but also the fundamental physical reason for the system's interception failure. In reality, when a high-speed falling object impacts the flexible net, the resulting tension is not transmitted instantaneously, but propagates along the net cables in the form of a stress wave. When the stress wave reaches the drone's connection point, the drone is neither a physically "fixed end" (mechanically impeded) nor... It is not a "free end" (or "free end"). Instead, it is a "Controlled Finite Mass Moving Boundary." The dynamic characteristics of this boundary are highly frequency-dependent. In the initial transient phase of the impact (typically within the first 50ms), due to the motor response time and the inherent phase lag of the flight control loop, the controller cannot provide active damping in time to match the surge in characteristic impedance of the cable. At this point, the boundary of the UAV is primarily dominated by its physical mass *m*, manifesting as a massive inertial barrier. According to wave reflection theory, when a stress wave encounters a high-impedance inertial boundary, its reflection coefficient... Approaching: .

[0005] Where R is the reflection coefficient of the stress wave at the boundary; It is the equivalent mechanical impedance "seen" by the stress wave propagation medium at the boundary; It is the characteristic impedance of stress waves propagating in a flexible mesh; It is the characteristic impedance constant of the stress wave.

[0006] This means that the reflected wave and the incident wave will undergo coherent superposition (constructive interference), causing the local tension at the junction to instantaneously reach twice the amplitude of the incident wave. This physical phenomenon is known as "pulse force peak caused by boundary reflection." Existing technologies, by neglecting this wave effect, often result in controllers that underestimate the load at the moment of impact, leading to insufficient output and subsequent system instability during the rebound phase due to integral saturation. This invention is proposed to address this transient dynamics caused by "moving boundary wave reflection."

[0007] Existing UAV control systems (such as traditional PID controllers) and structural designs are completely incapable of withstanding such high-speed impact events. More seriously, existing technologies fail to accurately identify and model the key failure mechanisms during the impact process. Specifically, the impact load is not a smooth peak, but a complex, multi-stage process. After the target object touches the net, high-speed stress waves propagate through the flexible net, reflect upon reaching the UAV's connection point (i.e., the boundary), and undergo complex interference and superposition with subsequent incoming waves. This superposition effect generates a sharp, pulsed force peak within an extremely short time window (i.e., the transient loading and peak impact phase), far exceeding the average peak tensile force calculated based on energy conservation. This pulsed force peak poses the most severe test to the instantaneous response capability of the UAV motors, the structural strength of the arm, and the robustness of the flight control system. Due to a lack of deep understanding and modeling of this critical physical phase, existing technologies' control strategies (usually based on passive responses to error feedback) fail instantly in the face of such pulsed disturbances, leading to UAV attitude instability, structural damage, and even complete mission failure.

[0008] Therefore, there is an urgent need in this field for a new technical solution that can not only accurately model the coupling dynamics of multiple UAVs and flexible nets, but more importantly, can identify and actively suppress the huge pulse disturbances generated by stress wave superposition during the transient loading stage of high-speed impact, thereby ensuring the integrity and stability of the system. Summary of the Invention

[0009] Purpose of the invention: The technical problem to be solved by the present invention is to address the shortcomings of the existing technology by providing a high-altitude emergency rescue system based on multi-UAV cooperative flexible net rescue, including a cooperative formation of N UAVs, each UAV being equipped with a flight controller; and a flexible net suspended by N ropes and connected to the N UAVs for intercepting target objects;

[0010] Each drone's flight controller includes: a nonlinear disturbance observer (NDO) and a robust controller;

[0011] The nonlinear disturbance observer is used to estimate in real time the transient pulse disturbance caused by stress wave superposition transmitted from the flexible net to the UAV during an impact event in which the target is intercepted by the flexible net. and disturbance torque And generate an estimated interference signal;

[0012] The robust controller (e.g., an adaptive sliding mode controller, SMC) is used to receive the estimated interference signal and use the interference signal as a feedforward compensation input to generate a control command to drive the actuators of the UAV, thereby actively suppressing transient pulse disturbances.

[0013] The system is controlled based on a unified coupled dynamics model, which includes equations describing the translation of the UAV's center of mass, equations describing the UAV's rotation around its center of mass, and a direction cosine matrix. The equation.

[0014] The equation describing the translation of the UAV's center of mass is in the following form:

[0015] (1),

[0016] in, It is the mass of the i-th drone; is the acceleration vector of the center of mass of the i-th UAV in the inertial coordinate system; g is the acceleration due to gravity; It is the direction cosine matrix of the i-th UAV from the machine system to the inertial frame; It is the total thrust vector generated by the i-th UAV under the system; It is the aerodynamic disturbance experienced by the i-th UAV; The resultant force exerted by the flexible net and tethering ropes through the i-th drone:

[0017] (2),

[0018] in, It is the shape variable of the network; It is the time-varying rate of change of the deformation of a flexible net or rope, that is, the equivalent representation of strain rate; It is a first-order linear elastic stiffness coefficient; It is a third-order linear elastic stiffness coefficient; It is the viscoelastic damping coefficient; It is the viscoelastic damping coefficient; It is a symbolic function.

[0019] The equation describing the rotation of the UAV around its center of mass is in the form of:

[0020] (3),

[0021] in Let be the rotational inertia matrix of the i-th UAV about its center of mass; It is the angular acceleration of the i-th UAV; It is the angular velocity of the i-th drone; It is the controllable input torque generated by the rotor system of the i-th UAV; It is the external torque generated by aerodynamic disturbance; It is the additional external torque exerted by the flexible net and tethering rope on the i-th UAV;

[0022] The direction cosine matrix The equation is in the form of:

[0023] (4),

[0024] Among them, and Represent and .

[0025] The unified coupled dynamics model also includes a modified energy conservation model describing the impact process. This modified energy conservation model integrates the nonlinear force-deformation relationship and energy dissipation terms, and its specific form is as follows:

[0026] (5),

[0027] in, and These are the target mass and initial velocity, respectively. This refers to the height of the fall. It is a constant, representing the maximum value of the network's deformation. This represents the energy dissipation caused by viscoelastic damping; d represents the differential symbol.

[0028] The suspension geometry and flexible net parameters of the system are configured to satisfy the following constraints based on peak deceleration analysis:

[0029] Peak deceleration estimation:

[0030] (6),

[0031] in, It is the peak deceleration;

[0032] Total tensile force of the system Estimate:

[0033] (7),

[0034] Single-unit pull of the i-th drone for:

[0035] (8),

[0036] in, It is the uneven distribution factor of the i-th UAV; The angle between the rope and the vertical direction (set to approximately) );

[0037] And minimum network shape variable satisfy:

[0038] (9),

[0039] in, It is an uneven distribution factor; This represents the maximum allowable peak tensile force for a single unit.

[0040] The present invention also provides a high-speed impact robust control method for the said system, comprising the following steps:

[0041] Based on mechanical analysis, the impact process of the target object is identified as:

[0042] Stage 1: A localized impact stage; Stage 2: A transient loading and peak impact stage; Stage 3: A damped oscillation stage, represented as:

[0043] (10)

[0044] in, It is the first arrival time of the stress wave from the center of the net to the connection point of the UAV; t represents the system running time, which is a continuous time variable starting from the initial moment of contact between the target and the net.

[0045] The second confirmation phase involves the reflection and superposition of stress waves at the boundary of the flexible net (i.e., the connection point of the UAV), thereby generating a sharp, pulse-like disturbance. and disturbance torque ;

[0046] During Phase 2, a nonlinear disturbance observer (NDO) is used to estimate the pulsed disturbance force and disturbance torque in real time to generate an estimated disturbance signal.

[0047] The estimated disturbance signal is fed forward to a robust controller (e.g., an adaptive sliding mode controller, SMC), which generates control commands to actively compensate for the pulsed disturbances, thereby forcing the UAV's system state to remain stable.

[0048] The present invention also provides an electronic device, including a processor and a memory, the memory storing program code that, when executed by the processor, causes the processor to perform the method described thereon.

[0049] The present invention also provides a storage medium storing a computer program or instructions that, when the computer program or instructions are run on a computer, execute the method described thereon.

[0050] The core technical problem to be solved by this invention is: how to effectively suppress the huge, transient, pulse-like force peak generated when the target is intercepted by the flexible net, and ensure that the system can maintain structural integrity and flight stability when subjected to this impact load, so as to avoid mission failure.

[0051] More specifically, this invention aims to address the pulsed peak tensile force, far exceeding the target's gravity or average impact force, generated by stress wave reflection and superposition at the boundary during the transient loading and peak impact phases (i.e., phase two) of the impact process. This pulsed force poses the most severe challenge to the UAV's motors, structure, and flight control system. This invention achieves active compensation and suppression of this pulsed disturbance through advanced control strategies.

[0052] Compared with the prior art, the present invention has the following significant advantages:

[0053] 1. Improved model fidelity and design accuracy. This invention overcomes the shortcomings of traditional linear and undamped models by introducing nonlinear force-deformation relationships and energy dissipation terms into the impact energy model, making the model predictions more consistent with the real physical properties (geometric stiffening and viscoelastic damping) of the polymer flexible mesh, and avoiding erroneous predictions of system oscillations.

[0054] 2. The balance between system design and operational efficiency has been optimized. Through in-depth analysis of the suspension geometry (angle), this invention reveals the inherent contradiction between mechanical efficiency and interception area, providing a constrained optimization criterion for system design that balances ensuring the safety of individual unit loads with maximizing interception success rate.

[0055] 3. Robust suppression of transient pulse impacts is achieved. By finely dividing the impact process into three stages, this invention accurately identifies the transient loading and peak impact stage (stage two) as the stage that is caused by stress wave superposition and poses the greatest threat to the system.

[0056] 4. An active compensation control strategy is provided. The control strategy of Nonlinear Disturbance Observer (NDO) + ​​Adaptive Sliding Mode Controller (SMC) adopted in this invention is valuable for its ability to handle strong nonlinearity, fast time-varying, and large-amplitude uncertainty disturbances. The fast estimation and feedforward active compensation mechanism of NDO can actively cancel the disturbance when the impulse force peak (stage two) occurs, instead of passively responding after a huge error occurs, as is the case with traditional PID. The robustness of SMC ensures the final stability of the system under extreme disturbances.

[0057] In summary, this invention solves the robustness problem of multi-UAV systems under high-speed impact by combining in-depth mechanical analysis with advanced control strategies, making it possible to safely and stably capture high-altitude falling targets using UAV swarms. Attached Figure Description

[0058] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0059] Figure 1 This is a schematic diagram of the rescue operation process and robust control logic of the present invention.

[0060] Figure 2 This is a tension distribution diagram of representative working conditions (S1-S6) according to an embodiment of the present invention, showing the relationship between the total tension of the system and the peak tension of a single unit as a function of the drop height and the net shape.

[0061] Figure 3 This is a steady-state circumferential tension distribution diagram according to an embodiment of the present invention, showing the distribution of circumferential tension in a flexible net as a function of the equivalent radius r under specific working conditions.

[0062] Figure 4 The vertical total thrust curve output by the controller demonstrates the system's disturbance rejection capability. Detailed Implementation

[0064] This invention provides a high-altitude emergency rescue system based on a multi-UAV cooperative flexible network. The rescue system of this invention is a highly coupled multibody dynamics system, and the flowchart is as follows. Figure 1 As shown. To implement the control method of this invention, an accurate system model must first be established.

[0065] To accurately describe the motion of the i-th UAV in the system, a standard six-degree-of-freedom nonlinear dynamic model is established. It is assumed that the UAV is a rigid body with symmetrical mass distribution. The translational motion of the UAV's center of mass is described in the inertial coordinate system {I} and follows Newton's second law:

[0066] (1),

[0067] It is the direction cosine matrix from the body coordinate system {B} to the inertial coordinate system {I}, which represents the attitude angles. The function, specifically in the form of:

[0068] (2),

[0069] The total thrust generated by the four rotors, along the body coordinate system Acting in the negative direction of the axis; Air resistance, which can usually be simplified and modeled as a damping force proportional to velocity, i.e. ,in The overall aerodynamic damping coefficient; This is the tension vector applied to the drone by a flexible net via ropes; this is a key factor in achieving system coupling. This is particularly relevant for polymer materials operating at high strain rates. To address the viscoelastic characteristics under these conditions, this invention introduces a simplified form based on the Zhu-Wang-Tang nonlinear constitutive equation and replaces the traditional linear viscous damping with a power-law damping model:

[0070] (3),

[0071] This reflects the "dynamic hardening" phenomenon caused by the impeded slippage of molecular chains during high-speed stretching of materials. Compared to traditional... Compared to damping, this model can more accurately describe the hysteresis loop of energy dissipation during impact.

[0072] The above nonlinear forces A strong coupling torque is generated by applying a lever arm r to the drone. .because Includes high-order terms of speed This results in a coupling term with extremely large time-varying gain in the UAV attitude dynamics equations. At the moment of impact, this term acts as a parameter excitation, easily inducing nonlinear oscillations in the UAV attitude loop. The NDO+SMC control architecture adopted in this invention is precisely designed to estimate and compensate for this strong nonlinear disturbance caused by the complex material constitutive model in real time, which is impossible with conventional PID control.

[0073] The rotation of the UAV around its center of mass is described in the body coordinate system {B}, and its expression is:

[0074] (4),

[0075] The core characteristic of this system lies in the bidirectional dynamic coupling between the UAV subsystem and the flexible net subsystem. On one hand, the tension applied by the flexible net... and The external time-varying disturbance is input into formulas (1) and (4). On the other hand, the dynamic behavior of the flexible net itself (such as geometry and internal tension) depends entirely on its boundary conditions, namely the mounting points defined by the positions of each UAV.

[0076] To improve model fidelity, this invention makes two key modifications to the traditional energy conservation equation:

[0077] (5),

[0078] Secondly, an energy dissipation term must be introduced. In a real impact, not all initial kinetic and potential energy is converted into elastic potential energy. A portion of the energy is converted into heat through the viscoelastic damping (internal friction) of the material. This energy dissipation mechanism is crucial for the stability of the system after the impact; it is the physical basis for the system's vibration to decay and eventually reach a new equilibrium state. (Ignoring...) This can lead to models predicting that the system will oscillate indefinitely after the shock, which contradicts reality.

[0079] The suspension geometry, especially the angle between the rope and the vertical direction. It is a key design parameter that determines the mechanical performance and operational efficiency of the system.

[0080] Under equivalent linear elasticity Then the peak velocity is approximately:

[0081] (6),

[0082] Total tensile force of the system:

[0083] (7),

[0084] Considering uneven angles in the four-corner distribution:

[0085] (8),

[0086] This reflects transient non-uniformity, from which we can know Enlarge Reduce Decrease.

[0087] If the maximum allowable peak tensile force for a single machine is known. (Limited by motor, arm, rope, or safety redundancy), it can be deduced that satisfying the requirements can be achieved. Minimum required network shape :

[0088] (9),

[0089] This invention reveals a profound inherent contradiction in system design:

[0090] Mechanical efficiency: to reduce It is necessary to make Increase, that is The angle decreases (the rope becomes more vertical).

[0091] Operational efficiency: To increase the interception area and improve the success rate, drones must be dispersed, leading to... The angle increases.

[0092] As shown in Table 1, with a target mass m = 70 kg and a falling height h = 10 m, therefore... =14.01m / s, number of drones N=4, uneven distribution factor For example, under the conditions of a uniform deformation of 4.75m and a single-machine ultimate tensile force of 800N, we analyze different included angles. Impact on system design parameters.

[0093] Table 1: Influence of suspension angle on minimum deformation and peak tensile force

[0094]

[0095] Table 1 clearly shows the data, if It would require a deformation of over 10 meters to control the tensile force within 800N, which is practically impossible in engineering. Only 4.75 meters of deformation is required to meet the requirements. Therefore, a preferred embodiment of the present invention selects... As a reasonable balance between mechanical safety and working area.

[0096] This invention decomposes the unmanned aerial vehicle (UAV) system interception process into the following three stages with distinct physical characteristics:

[0097] Phase 1 (Local Impact and Wavefront Propagation Phase) From the instant the target object contacts the center of the net (t=0) until the stress wave first propagates to the UAV's connection point, this process takes approximately 31.6 milliseconds. During this extremely short time, the UAV is completely unaware of the impact, and the load it bears remains under pre-tension. All impact energy is absorbed and converted by localized areas of the flexible net; the main risk at this point is localized tearing or puncture of the net.

[0098] Phase Two (Transient Loading and Peak Impact Phase) When the high-speed propagating stress wave reaches the UAV's connection point (i.e., the structural boundary), the UAV begins to bear the load. Because the reflection of the stress wave at the boundary causes complex interference and superposition with subsequent incoming waves, the instantaneous peak tensile force experienced by the UAV at this stage may be far higher than the average peak tensile force calculated based on energy conservation. This sharp, pulsed force peak is the most severe test of the UAV's motor instantaneous response capability, arm structural strength, and flight control system robustness.

[0099] Phase Three (Damped Oscillation and Quasi-Static Stability Phase) Following the initial violent impact, the entire "target-flexible net-UAV" system enters a damped oscillation process. The system's stored elastic potential energy and remaining kinetic energy are mutually converted and gradually dissipated through the viscoelastic damping of the materials and air resistance. After multiple reflections and energy dissipation, the system vibration gradually decays, and the tension on each UAV eventually tends to a quasi-static value that balances the weight of the target object. .

[0100] Based on the above mechanical analysis, the control strategy of this invention aims to effectively suppress the huge pulse-like disturbances generated in the second stage (transient loading stage) of the interception process. and .

[0101] To achieve this objective, this invention employs an advanced control algorithm that combines an adaptive sliding mode controller (SMC) with a nonlinear disturbance observer (NDO). The value of such algorithms lies precisely in their ability to handle strongly nonlinear, rapidly time-varying, and large-amplitude uncertainty disturbances.

[0102] The operating mechanism of this control scheme is as follows:

[0103] The core function of NDO is to quickly and accurately estimate the unknown impact force caused by the superposition of stage two stress waves. It solves for this unknown, pulse-like external disturbance in real time by comparing the difference between the actual state of the system and the predicted state of the model.

[0104] The disturbance force estimated by the NDO is fed forward to the controller for active compensation. Traditional PID controllers are passive; they must wait until the impact force has caused a significant position or attitude error before adjusting, by which time it is too late and the system may have already become unstable. The method of this invention is active compensation. It anticipates the disturbance force itself and instructs the motor to output an opposite torque and thrust before the disturbance force causes a significant error, thereby actively counteracting the impact pulse.

[0105] The role of a sliding mode controller (SMC) is to provide robustness. SMC ensures that the system state is constrained to the desired sliding surface even when the model is inaccurate (e.g., inaccurate estimation) or there are unmodeled dynamics (e.g., air disturbances), thus guaranteeing the stability of the system.

[0106] To verify the technical solution of the present invention, a specific embodiment is provided, and its parameter settings and analysis are as follows.

[0107] Mass m = 70 kg; gravitational acceleration g = 9.81 m / s²; maximum deformation of the net. m; drop height m; Number of drones N=4; Angle between rope and vertical Uneven distribution factor Table 2 shows representative operating conditions. The velocity before touching the net (approximate for free fall); This is the equivalent peak deceleration for linear elasticity.

[0108] Table 2 Representative Operating Conditions

[0109]

[0110] Under the conditions given in Table 2, the total tensile force of the system and the peak tensile force of a single unit are shown in Table 3 and... Figure 2 As shown.

[0111] Table 3 Total tensile force and peak tensile force of a single unit

[0112]

[0113] Under operating condition S4, the target object's contact speed with the net. The average peak deceleration is 9.90 m / s². The average peak deceleration is calculated based on formulas (5) and (6). The total tensile force of the system is 19.60 m / s². The average peak tensile force per unit is 2058N. It is 772N.

[0114] This 772N value is lower than that in Table 1. The calculated ultimate tensile force is 828N or the tensile force limit set in the engineering.

[0115] However, according to transient dynamics analysis, in stage two (transient loading stage), the actual instantaneous peak tensile force... It will be higher than 772N due to the superposition of stress waves.

[0116] In this embodiment, the objective of the NDO+SMC robust control system of the present invention is that the NDO quickly and accurately estimates the instantaneous pulse disturbance caused by the stress wave. The controller feeds forward the estimated disturbance value to actively compensate for the impact force. SMC ensures system stability, successfully suppressing the actual instantaneous peak thrust experienced by the UAV within the upper limit of the permissible peak thrust for a single unit (e.g., 1500N), thereby guaranteeing the success of the rescue mission. Figure 3 As shown.

[0117] In stage three (damped oscillation and quasi-static stability stage), after the system stabilizes, the circumferential tension distribution of the flexible net is shown in Table 4 and... Figure 4 As shown (based on operating condition S4 data, steady-state single-machine tension) Steady-state total load ).

[0118] Table 4 Steady-state circumferential tension distribution

[0119]

[0120] This invention provides a high-altitude emergency rescue system and a high-speed impact robust control method based on a multi-UAV cooperative flexible network rescue system. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment of the invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A high-altitude emergency rescue system based on multi-unmanned aerial vehicle cooperative flexible network rescue, characterized in that, The system comprises a cooperative formation of N unmanned aerial vehicles (UAVs), each equipped with a flight controller; and a flexible net suspended by N ropes and connected to the N UAVs for intercepting a target object; The flight controller of each UAV comprises a nonlinear disturbance observer and a robust controller; The nonlinear disturbance observer is used to estimate in real time the transient impulse disturbance caused by stress wave superposition transmitted to the UAV by the flexible net in the impact event of the target being intercepted by the flexible net and the disturbance torque and generates an estimated disturbance signal; The robust controller is configured to receive an estimated disturbance signal and input the disturbance signal as a feedforward compensation to generate a control command for driving an actuator of the UAV to actively suppress the transient impulse disturbance.

2. The system of claim 1, wherein, The system is controlled based on a unified coupled dynamics model that includes equations describing the translational motion of the center of mass of the UAV, equations describing the rotational motion of the UAV about the center of mass, and equations for the direction cosine matrix .

3. The system of claim 2, wherein, The equation describing the translational motion of the center of mass of the UAV is of the form: （1）， wherein, is the mass of the ith UAV; is the acceleration vector of the mass center of the ith UAV in the inertial frame; g is the gravitational acceleration; is the direction cosine matrix of the ith UAV from the body frame to the inertial frame; is the total thrust vector of the ith UAV in the body frame; is the aerodynamic force disturbance on the ith UAV; is the resultant force of the flexible net and the tether through the ith UAV: （2）， wherein is the deformation of the net; is the rate of change of the deformation of the flexible net or rope with time; is the first order linear elastic stiffness coefficient; is the third order linear elastic stiffness coefficient; is the viscoelastic damping coefficient; is the viscoelastic damping coefficient; is the sign function.

4. The system of claim 3, wherein, The equation describing the rotational motion of the center of mass of the UAV is of the form: （3）， wherein is the rotational inertia matrix of the i-th UAV about the center of mass; is the angular acceleration of the i-th UAV; is the angular velocity of the i-th UAV; is the controllable input moment of the i-th UAV generated by the rotor system; is the external moment generated by the aerodynamic force disturbance; is the additional external moment exerted on the i-th UAV by the flexible net and tether.

5. The system of claim 4, wherein, The direction cosine matrix The equation is of the form: （4）， wherein, wherein and respectively represent and .

6. The system of claim 5, wherein, The unified coupled dynamic model further comprises a modified energy conservation model describing the impact process, which integrates a nonlinear force-deformation relationship and an energy dissipation term, and is of the form: （5）， wherein, and are the target mass and initial velocity, respectively; is the drop height; is a constant; is the energy dissipation due to viscoelastic damping; d denotes the differential symbol.

7. The system of claim 6, wherein, The suspension geometry and the flexible net parameters of the system are configured to satisfy the following constraint relationship based on peak deceleration analysis: Peak deceleration estimation: （6）， wherein is the peak deceleration; Total system pull force Estimate: （7）， Single drone pull of the ith drone is: （8）， wherein, is the uneven distribution factor of the i-th drone; is the angle between the rope and the vertical direction; and a minimum mesh variable satisfies: （9）， wherein, is a non-uniform distribution factor; is a single machine allowed peak tension upper limit.

8. A high speed impact robust control method for the system of any one of claims 1 to 7, characterized in that, The method comprises the following steps: Based on mechanical analysis, the impact process of the target object is identified as: Phase 1: a local impact phase, phase 2: a transient loading and peak impact phase, and phase 3: a damped oscillation phase, represented as: （10）， wherein, is the first arrival time of the stress wave propagating from the web core to the drone connection point; t represents the system running time; The second phase of validation is the reflection and superposition of the stress waves at the boundary of the flexible net, thus generating a sharp, impulsive disturbing force and disturbing moment ; During phase 2, a nonlinear disturbance observer is used to estimate the impulse disturbance force and torque in real time to generate an estimated disturbance signal; The estimated disturbance signal is fed forward to a robust controller, and the robust controller is used to generate a control command to actively compensate for the impulse disturbance to force the system state of the UAV to remain stable.

9. An electronic device, comprising: The system comprises a processor and a memory, and the memory stores program code, which, when executed by the processor, causes the processor to execute the method of claim 8.

10. A storage medium, characterized by The computer program or instructions are stored, and when executed on a computer, perform the method of claim 8.

Images