Active anti-swing control method, device and system for unmanned aerial vehicle to hoist load

Abstract

The application provides a method, device and system for active anti-swing control of a payload hoisted by a UAV, and relates to the field of industrial automation, comprising: obtaining dynamic parameters of the UAV, mass parameters of the payload, and structure parameters representing the flexibility of the soft ladder; constructing a rigid-flexible coupled dynamic model of the UAV, the soft ladder and the payload based on the above parameters; identifying time-varying structure parameters of the soft ladder based on the model during hoisting; and dynamically updating an active anti-swing control strategy based on the rigid-flexible coupled dynamic model and the time-varying structure parameters to control the UAV to perform an anti-swing action. The application solves the problem of model mismatch caused by the fact that the prior art does not consider the structure parameters of the soft ladder, improves the control accuracy and efficiency of active anti-swing, and ensures hoisting safety.

Description

Technical Field

[0001] This application relates to the field of industrial automation, specifically to an active anti-sway control method, device, and system for unmanned aerial vehicles (UAVs) lifting loads. Background Technology

[0002] In drone-based lifting operations such as power line inspection and logistics delivery, a mounting system is typically used in conjunction with a rope ladder to lift payloads such as maintenance robots. Existing drone-based anti-sway control technologies are mostly based on the assumption of rigid ropes or fixed-length pendulums to establish dynamic models, and input shaping or feedback control strategies are designed accordingly. However, the rope ladders used in actual operations have significant axial elasticity, structural damping, and distributed mass, and their structural parameters exhibit time-varying nonlinear characteristics due to factors such as tension changes and wear during lifting. Current technologies do not incorporate the rope ladder's own structural parameters into the active anti-sway control model, leading to a mismatch between the dynamic model and the actual physical system. This results in insufficient anti-sway control accuracy, low sway suppression efficiency, and even safety hazards during large-scale maneuvers due to the inability to accurately predict the rope ladder tension. Therefore, there is an urgent need for a technical solution that can integrate the rope ladder's own structural parameters into the control loop to improve anti-sway accuracy and safety. Summary of the Invention

[0003] To address the problems of mismatch in active anti-sway control models and insufficient control accuracy and safety caused by the failure to consider the structural parameters of the rope ladder itself in existing technologies, this application proposes an active anti-sway control method, device, and system for unmanned aerial vehicle (UAV) lifting loads to improve the control accuracy and efficiency of active anti-sway and ensure lifting safety.

[0004] To achieve the above objectives, this application adopts the following technical solution: An active anti-sway control method for unmanned aerial vehicle (UAV) lifting loads includes: acquiring UAV dynamic parameters, load mass parameters, and structural parameters of a rope ladder, wherein the structural parameters characterize the flexibility of the rope ladder; constructing a rigid-flexible coupled dynamic model of the UAV, the rope ladder, and the load based on the UAV dynamic parameters, load mass parameters, and rope ladder structural parameters; identifying time-varying structural parameters of the rope ladder based on the rigid-flexible coupled dynamic model during UAV lifting; and dynamically updating an active anti-sway control strategy based on the rigid-flexible coupled dynamic model and the time-varying structural parameters to control the UAV to perform anti-sway actions.

[0005] Preferably, the structural parameters include axial elastic parameters; the construction of a rigid-flexible coupled dynamic model consisting of the UAV, the rope ladder, and the load based on the UAV dynamic parameters, the load mass parameters, and the structural parameters of the rope ladder includes: constructing an elastic constraint equation considering axial elastic deformation based on the axial elastic parameters, so as to characterize the coupling relationship between the elastic deformation of the rope ladder and the motion of the load in the rigid-flexible coupled dynamic model.

[0006] Preferably, the structural parameters further include: the original length of a single rope ladder, axial elastic modulus, equivalent load-bearing cross-sectional area, axial stiffness, material structural damping coefficient, mass per unit length, ladder pitch, and equivalent diameter of the ladder rod; the elastic constraint equation is specifically: the distance between the load position and the connection point at the top of each rope ladder is equal to the sum of the original length of a single rope ladder and the elastic elongation of the rope ladder, wherein the elastic elongation is the ratio of the real-time tension to the axial stiffness of the rope ladder; the equivalent swing frequency corresponding to the rigid-flexible coupling dynamic model is jointly determined by the gravitational acceleration, the equivalent pendulum length, and the geometric relationship between the distance between the connection points at the top of the two rope ladders and the original length of a single rope ladder; wherein, the equivalent pendulum length is the sum of the original length of a single rope ladder and the ratio of the static tension and axial stiffness of the load.

[0007] Preferably, during the UAV hoisting process, identifying the time-varying structural parameters of the rope ladder based on the rigid-flexible coupling dynamic model includes: constructing an extended state observer by taking the equivalent axial stiffness and equivalent structural damping of the rope ladder as an extended state; and using the extended state observer to jointly estimate the time-varying structural parameters with the swing angle of the load and environmental wind disturbance.

[0008] Preferably, the state vector of the extended state observer includes: load longitudinal swing angle, load lateral swing angle, longitudinal swing angular velocity, lateral swing angular velocity, equivalent wind disturbance force component, equivalent axial stiffness, and equivalent structural damping; a coupling force term of the elastic deformation of the rope ladder and the equivalent structural damping is added to the observation model of the extended state observer, and the real-time tension of the rope ladder is calculated using the thrust estimate of the UAV as an auxiliary observation; during the takeoff phase, the initial structural parameters of the rope ladder are self-calibrated by executing a preset maneuvering excitation.

[0009] Preferably, the step of dynamically updating the active anti-sway control strategy based on the rigid-flexible coupling dynamic model and the time-varying structural parameters includes: calculating the system's equivalent sway frequency and total equivalent damping ratio under the current operating condition based on the equivalent axial stiffness in the time-varying structural parameters identified in real time and the real-time tension of the rope ladder; and dynamically updating the pulse parameters of the feedforward input shaper based on the system's equivalent sway frequency and the total equivalent damping ratio to generate a shaped reference trajectory.

[0010] Preferably, the dynamic updating of the pulse parameters of the feedforward input shaper includes: updating the pulse amplitude and time interval of the feedforward input shaper cycle by control period, and introducing a swing angle phase-frequency prediction stage to correct the pulse parameters in advance.

[0011] Preferably, the dynamic updating of the active anti-sway control strategy based on the rigid-flexible coupling dynamic model and the time-varying structural parameters further includes: adding a rope ladder allowable tension constraint to the reference trajectory planning, so that the upper limit of trajectory acceleration is dynamically adjusted with the real-time tension of the rope ladder to avoid the rope ladder tension exceeding the limit and reduce elastic oscillations caused by dynamic tension; introducing a rope ladder maximum allowable tension constraint in the feedback control, and constructing an extended obstacle function together with the rope ladder maximum allowable tension constraint and the swing angle constraint, so that the swing angle and the rope ladder tension are simultaneously constrained within the safety threshold, and the rope ladder tension is monitored in a closed loop; calculating the rope ladder wind resistance based on the structural parameters of the rope ladder, and incorporating the rope ladder wind resistance and load wind disturbance into the feedforward compensation term.

[0012] Preferably, the method further includes a phased control strategy that switches control parameters according to the operation phase, wherein the operation phase includes a takeoff cruise phase, a tower approach phase, a precise loading or unloading phase, and an emergency hovering phase; in the takeoff cruise phase, the rope ladder parameters are self-calibrated, and a feedforward input shaper configured with zero vibration derivative is used; in the tower approach phase, the high-order pulse parameters of the feedforward input shaper are dynamically updated, and active enhancement logic for rope ladder sway damping is added; in the precise loading or unloading phase, the feedforward input shaper is updated at high frequency, and closed-loop monitoring of the rope ladder tension is activated, and the loading action is suspended when the tension fluctuation exceeds a preset threshold; in the emergency hovering phase, optimal energy dissipation control logic based on the structural damping characteristics of the rope ladder is added.

[0013] In addition, this application also provides an active anti-sway control device for unmanned aerial vehicle (UAV) lifting loads, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the active anti-sway control method for UAV lifting loads as described above.

[0014] Furthermore, this application also provides a drone hoisting rack device, applicable to the active anti-sway control method for drone load hoisting as described in the above scheme, comprising: a flexible rope fixing plate, the flexible rope fixing plate having a hollowed-out bracing structure; a drone connecting frame mounting plate, disposed on the back of the flexible rope fixing plate; two insulated rope ladders, one end of which is connected to the flexible rope fixing plate; a rope ladder connecting frame, respectively connecting the other ends of the two insulated rope ladders; a universal maintenance robot hook, connected to the rope ladder connecting frame, the end of the universal maintenance robot hook having an opening structure for mounting; the two insulated rope ladders, the flexible rope fixing plate, and the rope ladder connecting frame form a double rope ladder triangular hoisting configuration.

[0015] In addition, this application also provides a drone hoisting system, including: a drone; a drone hoisting rack device as described above, installed on the drone; and a flight control system configured to execute the active anti-sway control method for drone load hoisting as described above.

[0016] Beneficial effects: The technical solution provided in this application constructs a rigid-flexible coupled dynamic model that incorporates structural parameters characterizing the flexibility of the rope ladder into the modeling process. This model includes the coupling relationship between the elastic deformation of the rope ladder and the load motion, overcoming the model mismatch defects caused by the rigidity assumption in existing technologies and providing an accurate model foundation for high-precision anti-sway control. Based on this, by identifying the time-varying structural parameters of the rope ladder online during hoisting, it is possible to perceive the characteristic drift of the rope ladder caused by wear, aging, or tension changes in real time, ensuring that the control strategy always matches the actual physical system.

[0017] Meanwhile, this application dynamically updates the active anti-sway control strategy based on real-time identified parameters, achieving adaptive adjustment of feedforward shaping and feedback control, significantly improving the anti-sway control accuracy and response efficiency under complex working conditions. Furthermore, this application combines allowable tension constraints on the rope ladder with a phased control strategy, effectively avoiding the risk of rope ladder overload and ensuring the safety of the entire hoisting operation.

[0018] The mounting device used in this application adopts a double-ladder triangular configuration and a lightweight hollow design, which provides passive stability and a high load-bearing ratio from a physical structural perspective. This, combined with the control method, creates a synergistic effect and further enhances the overall performance of the UAV hoisting system. Attached Figure Description

[0019] Figure 1 This is a flowchart of the active anti-sway control method for hoisting load according to an embodiment of this application; Figure 2 This is a schematic diagram of a flexible rope fixing plate with a hollowed-out tie structure according to an embodiment of this application; Figure 3 This is a schematic diagram of the overall structure of the drone hoisting rack device according to an embodiment of this application; Figure 4 This is a flowchart of the autonomous and precise mounting control method according to an embodiment of this application; Figure 5 This is a flowchart of the multi-objective optimization method for the hanging structure parameters according to an embodiment of this application.

[0020] Among them, 1-flexible rope fixing plate, 2-UAV connecting frame mounting plate, 3-lifting eye bolt fixing nut, 4-insulated rope ladder, 5-lifting eye bolt, 6-rope ladder connecting frame, 7-general maintenance robot hook. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0023] Example 1 like Figure 1 As shown, this embodiment provides an active anti-sway control method for unmanned aerial vehicles (UAVs) lifting loads. This method is executed by the flight control system or a dedicated control device onboard the UAV, aiming to solve the problem of anti-sway model mismatch caused by neglecting the inherent flexibility of the rope ladder in existing technologies. The method mainly includes the following steps.

[0024] Step S1 involves acquiring the UAV's dynamic parameters, payload mass parameters, and the ladder's structural parameters. These structural parameters characterize the ladder's flexibility. Specifically, the UAV's dynamic parameters may include its mass, moment of inertia, thrust coefficient, and aerodynamic damping coefficient, describing its motion response. The payload mass parameters refer to the real-time or nominal weight of objects lifted by a general-purpose maintenance robot. Unlike existing technologies that treat slings as fixed-length rigid constraints, the ladder's structural parameters in this embodiment not only include geometric dimensions but also, more importantly, encompass flexible properties characterizing its material mechanical behavior, such as axial elastic modulus, equivalent load-bearing cross-sectional area, material structural damping coefficient, and mass per unit length. These parameters collectively define the ladder's elongation, contraction, and energy dissipation characteristics under stress. In practical applications, some structural parameters can be pre-stored in memory through factory calibration, while payload mass parameters and some environmentally sensitive parameters can be acquired or estimated in real-time using onboard sensors. By introducing structural parameters characterizing flexibility, necessary physical inputs are provided for establishing a high-fidelity dynamic model, avoiding inherent model defects caused by simplification assumptions.

[0025] Step S2 involves constructing a rigid-flexible coupled dynamic model of the UAV, load mass parameters, and ladder structural parameters. Specifically, this model treats the UAV and load as rigid bodies, and the ladder as a flexible body with distributed mass and axial elasticity, coupling them into a unified system through elastic constraint equations. In this model, the ladder is no longer simplified as an inextensible geometric segment but is modeled as a flexible connector capable of axial elastic deformation and generating corresponding restoring and damping forces. This means that the oscillating motion of the load and the expansion and contraction of the ladder are coupled: the load oscillation causes periodic changes in the ladder's tension and length, while the ladder's elastic deformation, in turn, affects the load's oscillation frequency and damping characteristics. This rigid-flexible coupled modeling method accurately reflects the unique dynamic behavior of the double-ladder triangular hoisting configuration, particularly by precisely calculating the equivalent oscillation frequency of the system after considering the ladder's elastic effect, thus overcoming the problem of large frequency prediction deviations in traditional rigid single pendulum models under high-frequency maneuvers or heavy load conditions.

[0026] Step S3 involves identifying the time-varying structural parameters of the rope ladder during the drone lifting process based on a rigid-flexible coupling dynamics model. Specifically, because the rope ladder is subjected to repeated stretching, friction and wear, changes in ambient temperature and humidity, and long-term aging during actual operation, its key structural parameters, such as axial stiffness and damping coefficient, are not constant but exhibit time-varying nonlinear characteristics. If the control algorithm consistently uses the initial nominal parameters, the mismatch between the model and the actual physical object will gradually increase over time, leading to a decline in anti-sway performance and even resonance. This step continuously monitors the drone's attitude, acceleration, and motor speed during flight, utilizing an extended state observation mechanism to jointly identify the rope ladder's equivalent axial stiffness and equivalent structural damping as state variables to be estimated, along with load swing angle and environmental wind disturbance, in an online manner.

[0027] Step S4 involves dynamically updating the active anti-sway control strategy based on the rigid-flexible coupling dynamic model and time-varying structural parameters to control the UAV to perform anti-sway actions. Specifically, the active anti-sway control strategy typically includes two aspects: feedforward trajectory shaping and feedback error suppression. In traditional solutions, controller parameters are often based on fixed values ​​calibrated offline, which cannot adapt to real-time changes in system characteristics. In this embodiment, however, the control strategy parameters are dynamically generated: the system calculates the current equivalent swing frequency and total damping ratio of the system online based on the equivalent stiffness, damping, and current tension of the rope ladder identified in real time in step S3, and adjusts the pulse sequence parameters of the feedforward input shaper and the gain of the feedback controller cycle by cycle accordingly. For example, when it is detected that the stiffness of the rope ladder decreases and the swing frequency decreases due to wear, the system automatically extends the time interval of the shaper to match the new low-frequency characteristics, preventing residual vibration due to frequency misalignment. At the same time, the upper limit of the acceleration of the trajectory planning can be dynamically adjusted according to the real-time tension to avoid overloading of the rope ladder.

[0028] Example 2 In some embodiments, structural parameters include axial elastic parameters. Specifically, axial elastic parameters are key physical quantities describing the ease with which a rope ladder deforms in the direction of force, directly determining its dynamic expansion and contraction behavior during hoisting. Unlike existing technologies that simplify slings into inextensible rigid links, this embodiment introduces axial elastic parameters, enabling the dynamic model to capture minute changes in the rope ladder length caused by load acceleration / deceleration or wind disturbance. Based on the axial elastic parameters, an elastic constraint equation considering axial elastic deformation is constructed to characterize the coupling relationship between the elastic deformation of the rope ladder and the load motion in a rigid-flexible coupled dynamic model. This elastic constraint equation replaces the traditional fixed-length geometric constraint, establishing a functional mapping between the rope ladder tension and its real-time length. This means that when load oscillation causes rope ladder tension fluctuations, the model can automatically calculate the corresponding elastic elongation, thereby correcting the actual position coordinates of the load; conversely, changes in load position also affect the tension distribution through feedback via this equation. This two-way coupling mechanism truly reflects the energy exchange process of the flexible hoisting system, effectively avoiding prediction errors in the system resonance frequency caused by neglecting elastic potential energy.

[0029] As a preferred option, to more accurately quantify the aforementioned flexibility characteristics, the structural parameters also include: original length of a single rope ladder, axial modulus of elasticity, equivalent load-bearing cross-sectional area, axial stiffness, material structural damping coefficient, mass per unit length, ladder pitch, and equivalent diameter of the ladder rods. These eight parameters together constitute the complete mechanical fingerprint of the rope ladder. Among them, the original length of a single rope ladder refers to the natural length of the rope ladder in a stress-free state; the axial modulus of elasticity and the equivalent load-bearing cross-sectional area jointly determine the rope ladder's ability to resist axial tension; the axial stiffness is usually obtained by multiplying the modulus of elasticity and the cross-sectional area by the original length, and is a core parameter directly used to calculate the amount of elastic deformation; the material structural damping coefficient characterizes the ability of the rope ladder's internal friction and the friction at the ladder rod hinges to dissipate vibration energy, which is crucial for suppressing high-frequency oscillations; the mass per unit length reflects the rope ladder's own inertial effect, which cannot be ignored in long-span hoisting; the ladder rod pitch and the equivalent diameter of the ladder rods are structural parameters unique to rope ladders. The former affects the rope ladder's bending compliance and local modes, while the latter not only relates to bending stiffness but also directly determines the aerodynamic drag characteristics of the rope ladder's windward side. By incorporating these fine-grained parameters into the model, the control system can distinguish the dynamic differences of rope ladders of different specifications, wear levels, and even materials, significantly improving the model's generalization ability and robustness.

[0030] Based on this, the elastic constraint equation is specifically as follows: the distance between the load location and the connection point at the top of each rope ladder is equal to the sum of the original length of a single rope ladder and its elastic elongation, where the elastic elongation is the ratio of the real-time tension to the axial stiffness of the rope ladder. In mathematical terms, if we assume the load center of mass vector is... , No. The vector of the position of the upper connection point of the rope ladder is The original length of a single rope ladder is Real-time tension is axial stiffness is Then the elastic constraint equation can be expressed as: ; In this formula, It is not a constant value, but a dynamic component that changes in real time with flight maneuvers and wind disturbances. In the numerical solution process, this equation is embedded as an algebraic constraint in the differential algebraic equation system (DAE), ensuring that geometric compatibility and mechanical equilibrium are satisfied simultaneously in simulation and control calculations.

[0031] Furthermore, the equivalent swing frequency corresponding to the rigid-flexible coupling dynamic model is determined by the gravitational acceleration, the equivalent pendulum length, and the geometric relationship between the distance between the upper connection points of the two rope ladders and the original length of a single rope ladder. For the triangular hoisting configuration of the two rope ladders used in this application, its equivalent swing frequency... The calculation formula is as follows: ; in, It is the acceleration due to gravity; The spacing between the connection points at the top of the double rope ladder; This is the original length of a single rope ladder; This is the equivalent pendulum length. It is particularly important to note that the equivalent pendulum length... Defined as the sum of the ratios of the original length of a single rope ladder to the static tension of the load and the axial stiffness, i.e. ,here This represents the static tension borne by a single strand of a rope ladder when the load is in a stationary, suspended state. The formula includes two important physical corrections: the first is a correction for the pendulum length, expressed in the denominator... This indicates that the system's natural frequency depends not only on its geometric length but also on the combined modulation of the ladder's stiffness and static load. The smaller the stiffness or the heavier the load, the longer the equivalent pendulum length, and the lower the frequency. The second layer involves configuration modifications, using a multiplier factor. This demonstrates the geometric constraint effect of the double-rope triangular structure compared to single-rope suspension. The factor is consistently less than 1, indicating that the triangular configuration further reduces the system's equivalent oscillation frequency. Using this analytical formula, the system can quickly and accurately calculate the theoretical reference frequency given known structural parameters and current load, providing a precise basis for the subsequent parameter design of the input shaper. This fundamentally solves the oscillation suppression failure problem caused by model simplification in traditional methods.

[0032] Example 3 In some embodiments, to address the characteristic drift of the rope ladder due to wear, aging, or environmental factors during long-term operation, time-varying structural parameters of the rope ladder are identified based on a rigid-flexible coupled dynamic model during UAV hoisting. Specifically, this step involves constructing an extended state observer by treating the equivalent axial stiffness and equivalent structural damping of the rope ladder as an extended state. Traditional state observers typically focus only on the system's motion state, treating model parameters as fixed constants. However, in this application, considering that the axial stiffness and structural damping of the rope ladder are core physical quantities determining the system's oscillation frequency and energy decay rate, and that these two parameters are most prone to nonlinear time-varying under actual operating conditions, they are elevated to an extended state with equal status to the motion state for real-time estimation. Through the extended state observer, the time-varying structural parameters are jointly estimated with the load's sway angle and environmental wind disturbance. The technical value of this joint estimation mechanism lies in its ability to effectively decouple the influence of external disturbances and internal parameter changes on the system response. For example, when the load sway amplitude abnormally increases, it may be due to a sudden gust of wind, or it may be due to a decrease in the rope ladder's stiffness leading to a shift in its natural frequency and thus resonance. If only the swing angle is estimated without estimating the parameters, the controller is very prone to misjudgment; however, through joint estimation, the system can distinguish which responses are transient excitations caused by wind disturbances and which are modal changes caused by parameter drift, thus providing an accurate attribution basis for subsequent adaptive control.

[0033] Preferably, to ensure the observability and convergence accuracy of the joint estimation, the extended state observer's state vector includes seven components: load longitudinal swing angle, load lateral swing angle, longitudinal swing angular velocity, lateral swing angular velocity, equivalent wind disturbance force component, equivalent axial stiffness, and equivalent structural damping. The first four components describe the swing kinematics of the load in three-dimensional space; the equivalent wind disturbance force component simplifies complex atmospheric turbulence into a lumped disturbance force acting on the load's center of mass, compensating for unmodeled aerodynamic effects; and the latter two components track the evolution of the ladder's mechanical properties in real time. These seven components together constitute a complete minimal implementation system, ensuring that the state space has sufficient dimensionality to cover all key dynamic features while avoiding the computational burden and noise sensitivity caused by excessively high dimensionality.

[0034] Building upon this, to enhance the observer's ability to detect minute deformations of the rope ladder, a coupling force term representing the elastic deformation of the rope ladder and the equivalent structural damping is added to the observation model of the extended state observer. Specifically, the acceleration signal measured by the UAV's onboard IMU is actually a mixture of the UAV's own motion, the load's oscillating inertial force, and the rope ladder's elastic restoring force. If the coupling force generated by elastic deformation and damping is ignored in the observation equation, and the IMU reading is directly mapped to the swing angle, it will lead to significant estimation bias. This embodiment explicitly introduces the coupling force term derived from the elastic constraint equation described in Embodiment 2 into the observation matrix, enabling the observer to accurately extract the mechanical components caused by the rope ladder's expansion and contraction and internal friction from the mixed acceleration signal, thereby deducing the true swing angle state and parameter values. Simultaneously, the real-time tension of the rope ladder is calculated using the UAV's thrust estimate, which is injected into the observer as an auxiliary observation. This is a low-cost implementation path that does not require installing a tension sensor on the rope ladder. The principle is that, according to Newton's second law, the total lift generated by the UAV's rotor is equal to the vector sum of the UAV's weight, aerodynamic drag, and sling tension. Given the aerodynamic model of the UAV's body mass and current attitude, the tension component along the rope ladder direction can be indirectly calculated through vector decomposition. Introducing this real-time tension as an auxiliary observation provides the observer with a strong constraint condition directly related to the rope ladder stiffness, significantly improving the observability of the equivalent axial stiffness as a state variable. This is particularly effective in preventing drift or divergence of parameter estimates during steady-state flight when the load swing angle is small and elastic deformation is not significant.

[0035] Furthermore, considering the potential for significant parameter uncertainties in the observer during the initial startup phase, this embodiment also incorporates a dedicated initialization calibration process. Specifically, during takeoff, the initial structural parameters of the rope ladder are self-calibrated by executing a preset maneuvering excitation. This preset maneuvering excitation typically involves the UAV performing a set of amplitude-limited sinusoidal sweep or step maneuvers covering the system's equivalent bandwidth within a safe altitude. For example, controlling the UAV to perform a sinusoidal reciprocating motion with an amplitude of 0.5m at a frequency of 0.5Hz to 2Hz in the horizontal plane for 3 to 5 cycles. This specific excitation signal can fully excite the axial elastic mode and oscillation mode of the rope ladder, enabling the system to generate a dynamic response with a sufficient signal-to-noise ratio. During the excitation process, the estimated values ​​of the equivalent axial stiffness and equivalent structural damping output by the observer are compared in real time with the pre-stored nominal parameters. If the deviation exceeds a preset threshold, the baseline parameters of the subsequent control algorithm are corrected using the observed values; if the deviation is within the threshold, the nominal parameters are confirmed to be valid. Meanwhile, strict physical constraints were set throughout the identification process. For example, the estimated equivalent axial stiffness was limited to 50% to 120% of the material's theoretical stiffness, and the equivalent structural damping ratio was limited to 0.01 to 0.2. These boundary conditions, based on the physical limits of the rope ladder, prevent non-physical jumps in the estimated values ​​caused by sensor noise or instantaneous disturbances, ensuring the numerical stability and engineering reliability of the online identification process. Through the aforementioned self-calibration process, the system can automatically eliminate initial parameter errors caused by assembly differences, transportation damage, or changes in ambient temperature before each operation begins, laying a solid foundation for high-precision anti-sway control throughout the entire process.

[0036] Example 4 In some embodiments, an active anti-sway control strategy is dynamically updated based on a rigid-flexible coupling dynamic model and time-varying structural parameters. Specifically, this step aims to address the problem that traditional input shaping techniques, which rely on fixed model parameters, cannot adapt to the real-time changes in the flexible characteristics of the rope ladder. Based on the equivalent axial stiffness in the time-varying structural parameters identified in real time and the real-time tension of the rope ladder, the system's equivalent sway frequency and total equivalent damping ratio under the current operating conditions are calculated. The emphasis on real-time is because the equivalent axial stiffness of the rope ladder slowly drifts with wear and aging, while the real-time tension of the rope ladder... This will cause drastic fluctuations within milliseconds due to load swings and flight maneuvers. According to the rigid-flexible coupling dynamic model established in Example 2, the system's equivalent oscillation frequency is not a constant value, but rather a function of the aforementioned time-varying parameters. The system's equivalent oscillation frequency under the current operating condition... The calculation formula is as follows: ; in, It is the acceleration due to gravity; The load mass estimate obtained in real time is used for identification. This provides an estimated value of the equivalent axial stiffness of a single rope ladder obtained in real time. Through this online calculation method, the system can accurately capture the instantaneous frequency shift caused by tension fluctuations, as well as the long-term frequency drift caused by material aging, providing a real-time physical reference for subsequent precise sway suppression.

[0037] Based on this, the pulse parameters of the feedforward input shaper are dynamically updated according to the system's equivalent oscillation frequency and total equivalent damping ratio to generate a shaped reference trajectory. The core principle of the input shaper is to convolve the original command signal with a series of pulse sequences, using the destructive interference between the pulses to eliminate system resonance. The parameters of the pulse sequence depend entirely on the system's natural frequency and damping ratio. For a zero-vibration derivative (ZVD) shaper, the parameter calculation formulas for its three pulses are shown in the following example: ; ; ; in, These represent three pulses respectively. These represent the corresponding times. This is the attenuation factor related to the damping ratio. It should be understood that although a ZVD shaper is used as an example here, in other implementations, zero-vibration derivative (ZVDD) shapers or additional insulation (EI) shapers can be used depending on the requirements for robustness or response speed. Their pulse parameters are also dynamically calculated according to the mapping relationship between the real-time frequency and the damping ratio. This dynamic update mechanism ensures that even in extreme conditions where large fluctuations in the rope ladder tension cause the system frequency to deviate from the nominal value by more than 20%, the shaper can still effectively suppress residual vibrations, avoiding the risk of failure due to frequency mismatch in traditional fixed-parameter shapers.

[0038] Preferably, to further enhance dynamic adaptability, the pulse parameters of the feedforward input shaper are dynamically updated, including: updating the pulse amplitude and time interval of the feedforward input shaper cycle by control cycle, and introducing a swing angle phase-frequency prediction stage to correct the pulse parameters in advance. Specifically, cycle-by-control update means that the flight control system recalculates a set of the latest pulse parameters in each control cycle, rather than updating only once during mission switching. However, real-time updates alone still have limitations: due to the high coupling between the cable ladder tension and the swing angle, the tension is at its maximum and the frequency is highest when the load swings to its maximum amplitude, while the tension is at its minimum and the frequency is lowest at the equilibrium position. This periodic fluctuation in frequency means that by the time the controller detects the current frequency, completes parameter calculation, and then performs convolution operations, the actual frequency of the system may have changed, resulting in the control action lagging behind the physical process. To address this, this embodiment introduces a swing angle phase-frequency prediction stage. This stage uses the longitudinal and lateral swing angles of the load output by the observer and their rates of change to estimate the phase of the current swing cycle, and combines this with a rigid-flexible coupling dynamic model to predict the expected frequency in the next control cycle. For example, when the swing angle is detected to be moving from zero to its maximum value with a positive velocity, the prediction module will determine that the tension will increase in the next moment, thus adjusting the design frequency parameters of the shaper in advance. This correction based on a physical model effectively offsets the combined time delay caused by sensor sampling, algorithm calculation, and actuator response, enabling the input shaper to maintain precise synchronization with the system modes during dynamic processes with rapidly changing frequencies, significantly improving the sway suppression accuracy and stability under complex maneuvers.

[0039] Example 5 In some embodiments, to further enhance the safety and environmental adaptability of the hoisting process, the dynamic update of the active anti-sway control strategy based on the rigid-flexible coupling dynamic model and time-varying structural parameters also includes multiple safety constraints and disturbance compensation mechanisms. Specifically, a rope ladder allowable tension constraint is added to the reference trajectory planning, so that the upper limit of trajectory acceleration is dynamically adjusted with the real-time tension of the rope ladder to avoid the rope ladder tension exceeding the limit and reduce elastic oscillations caused by dynamic tension. The essence of this mechanism is to bring the mechanical safety boundary of the rope ladder forward to the motion planning layer. Traditional methods often only limit the maximum acceleration of the UAV body, while ignoring the instantaneous peak value of the sling tension. In this embodiment, the system predicts the tension change trend in the near future based on the real-time identified equivalent axial stiffness of the rope ladder and the current swing angle state. When the predicted tension approaches the allowable tension threshold of the rope ladder, the acceleration command of the reference trajectory is automatically reduced to suppress the propagation of elastic waves excited by violent maneuvers from the source.

[0040] Based on this, a maximum allowable tension constraint for the rope ladder is introduced into the feedback control. This maximum allowable tension constraint, together with the swing angle constraint, constructs an extended barrier function to ensure that both the swing angle and the rope ladder tension are simultaneously constrained within a safe threshold, and to perform closed-loop monitoring of the rope ladder tension. Specifically, this extended barrier function... The potential field is expressed in logarithmic form as follows: ; in, The longitudinal swing angle of the load. This is the preset maximum allowable swing angle threshold; For the first Real-time tension estimate of the rope ladder. This represents the maximum allowable tension of the rope ladder. The physical meaning of this function lies in the angle of swing when... or tension As the system approaches its respective safety boundary, the function value tends towards infinity, generating a huge repulsive torque in the control law, forcing the system state away from the danger zone. Compared to traditional saturation limiting or penalty function methods, this logarithmic barrier function is continuously differentiable and has a steep boundary gradient. It can maximize the use of control capabilities within the safety envelope while ensuring system stability, avoiding performance waste caused by conservative design. Simultaneously, by performing closed-loop monitoring of the rope ladder tension, the controller can detect abnormal tension fluctuations in real time and trigger protection logic when necessary, ensuring structural safety throughout the hoisting process.

[0041] Furthermore, for complex high-altitude wind fields, the wind resistance of the rope ladder is calculated based on its structural parameters, and the wind resistance and load disturbance are jointly incorporated into the feedforward compensation term. Specifically, existing technologies typically only consider the wind resistance of the load itself, neglecting the significant aerodynamic drag generated by the slender, flexible rope ladder in the wind. This embodiment utilizes the aforementioned structural parameters such as the equivalent diameter of the ladder rod, the ladder rod pitch, and the total length of the rope ladder, combined with real-time wind speed and direction information, to integrally calculate the aerodynamic drag force distributed along the length of the rope ladder. For example, for cylindrical ladder rods, the drag coefficient can be obtained by looking up the Reynolds number in a table, then multiplied by the projected area to obtain the wind resistance per unit length, and finally integrated along the current attitude curve of the rope ladder to obtain the total wind resistance vector of the rope ladder. This force vector is superimposed with the load disturbance force and injected into the control loop as a feedforward quantity, which can effectively offset wind-induced steady-state errors and significantly improve the fixed-point hovering accuracy and trajectory tracking performance under high wind conditions.

[0042] As a preferred approach, to adapt to the diverse operational needs throughout the entire drone lifting process, this method also includes a phased control strategy that switches control parameters according to the operational stage. Specifically, the operational stages include the takeoff and cruise stage, the tower approach stage, the precision loading or unloading stage, and the emergency hovering stage. Each stage employs a targeted control configuration to achieve the optimal balance between safety and efficiency.

[0043] During the takeoff and cruise phase, the rope ladder parameters are self-calibrated, and a feedforward input shaper with zero vibration derivative configuration is used. The main task of this phase is to establish a stable lifting reference. As described in Example 3, the initial parameters of the rope ladder are identified and corrected online through preset motorized excitation to eliminate assembly errors. At the same time, since the cruise phase has high requirements for sway robustness but relatively relaxed requirements for response speed, the use of a zero vibration derivative (ZVD) shaper can maintain a good sway suppression effect within a certain frequency deviation range. Combined with a relatively relaxed sway angle constraint (e.g., ±15°), the system can quickly complete parameter convergence and state stabilization within a safe range.

[0044] During the approach phase to the tower, the higher-order pulse parameters of the feedforward input shaper are dynamically updated, and active damping enhancement logic for the rope ladder oscillation is incorporated. As the UAV approaches the target tower, space becomes limited and obstacle risks increase, necessitating a tighter constraint on the oscillation angle. At this point, the system upgrades the shaper to a higher-order type with greater tolerance for frequency errors to address frequency uncertainties caused by airflow disturbances due to distance variations. Simultaneously, the active damping enhancement logic artificially increases the system's equivalent damping by applying a small control torque opposite to the oscillation velocity, accelerating the decay of residual oscillations and preparing for entry into the high-precision operation window.

[0045] During the precise mounting or dismounting phase, a high-frequency updated feedforward input shaper is employed, and closed-loop monitoring of the rope ladder tension is activated. The mounting action is paused when tension fluctuations exceed a preset threshold. This is the most critical step in the entire operation, requiring extremely high positioning accuracy and contact safety. An extra-insulated (EI) shaper is selected for this phase because it has the strongest robustness to model frequency mismatch and can maintain sway suppression performance even when impedance changes cause sudden frequency shifts at the moment of contact. More importantly, the system monitors the rope ladder tension fluctuation rate in real time. Once an abnormal tension jump caused by hook collision or jamming is detected, the robotic arm or hook's action sequence is immediately paused, and the action continues only after the swaying has converged and the tension has returned to normal.

[0046] During the emergency hovering phase, energy dissipation optimal control logic based on the structural damping characteristics of the rope ladder is incorporated. When encountering sudden faults or requiring rapid load stabilization for emergency obstacle avoidance, conventional time-optimal control may induce excessive transient tension. The control objective in this phase shifts to maximizing the energy dissipation rate. The controller utilizes the identified equivalent structural damping characteristics of the rope ladder and designs feedback gain to ensure that the poles of the closed-loop system are configured in directions that maximize the excitation and utilization of the rope ladder's internal friction energy dissipation. This scheme cleverly guides the swing energy to be converted into heat energy dissipated within the rope ladder, thereby achieving rapid cessation of oscillation while ensuring structural safety. Through the adaptive switching of the above four stages, the phased control strategy of this application achieves full-scenario coverage from extensive transportation to precision operations, taking into account operational efficiency, control accuracy, and inherent safety.

[0047] Example 6 This embodiment provides an active anti-sway control device for UAV payload lifting, comprising: a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the active anti-sway control method for UAV payload lifting as described in any of Embodiments 1 to 5 above. Specifically, this device is the physical execution entity of the aforementioned active anti-sway control method, transforming abstract control logic into specific electrical signal processing and instruction output processes. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The memory can include volatile memory, such as random access memory (RAM); it can also include non-volatile memory, such as flash memory, hard disk, or solid-state drive. In practical engineering applications, this device is typically integrated into the flight control system of the UAV or mounted on the UAV fuselage as a separate mission computer. For example, in a preferred embodiment, the processor employs a high-performance microcontroller based on the ARM Cortex-M7 core, with a clock frequency of no less than 400MHz, to meet the real-time requirements of the extended state observer's cycle-by-cycle iterative calculations; the memory uses a combination of SRAM and Flash with power-loss protection, used to cache runtime state variables and store control algorithm code, respectively. It should be understood that although an embedded microcontroller has been used as an example here, in other embodiments, if the anti-sway algorithm involves complex deep learning inference or large-scale optimization solutions, the processor can also employ a GPU acceleration unit or a multi-core heterogeneous SoC platform, as long as the update frequency of the control loop meets the system stability requirements.

[0048] During operation, the processor reads pre-stored computer program instructions from the memory via a bus or dedicated interface, and sequentially executes steps such as parameter acquisition, model building, online identification, and dynamic strategy updating. Specifically, the processor first acquires data from the UAV's sensors and payload information through the communication interface, completing step S100 (data acquisition); then, it calls the dynamic model library and numerical integration routines stored in ROM to instantiate a rigid-flexible coupling dynamic model in RAM, completing step S200 (model building); next, it uses the on-chip floating-point unit (FPU) to perform matrix operations of the extended state observer at high speed, solving the time-varying parameters of the soft ladder in real time, completing step S300 (parameter identification); finally, based on the identification results, it looks up a table or calculates the shaper pulse sequence online to generate anti-sway control commands and outputs them to the UAV's power system via PWM or CAN bus, completing step S400 (strategy updating and execution). This integrated hardware and software design ensures that the complex adaptive anti-sway algorithm can operate stably within millisecond-level control cycles, avoiding phase lag and control divergence caused by computational delays. Meanwhile, by encapsulating the core algorithm in an independent control device, the anti-sway function and the underlying attitude control of the UAV are decoupled, enabling the device to be flexibly adapted to different UAV platforms. It can be reused by simply adjusting the interface protocol and calibration parameters, which significantly improves the portability and engineering practical value of the technical solution.

[0049] Example 7 like Figure 2 and Figure 3 As shown, this embodiment provides a drone hoisting rack device, applicable to the active anti-sway control method for drone load hoisting as described in any of the preceding embodiments one to six. Specifically, the drone hoisting rack device includes a flexible rope fixing plate 1, a drone connecting frame mounting plate 2, two insulated rope ladders 4, a rope ladder connecting frame 6, and a universal maintenance robot hook 7.

[0050] The flexible rope fixing plate 1 features a perforated bracing structure. Specifically, the flexible rope fixing plate 1 is generally rectangular, with multiple radially arranged triangular perforations distributed in its central area, and downward extending wall structures at the edges of the perforations. This specific perforated bracing topology is not a simple weight-reduction hollowing out, but a reinforcing rib network based on optimized mechanical transmission path design. It can significantly reduce the weight of the plate while maintaining or even improving the stiffness and bending and torsional resistance in key stress directions. In a preferred embodiment, the flexible rope fixing plate 1 is made of 6061 aluminum alloy, with a surface treated by sandblasting and bright black anodizing. This material has a high strength-to-weight ratio, excellent corrosion resistance, and good processability. Combined with the anodized layer, it can further improve surface hardness and wear resistance, adapting to harsh outdoor working environments. A set of circular mounting holes is provided at each of the four corners of the plate for mounting eye bolts 5 using eye bolts to fix nuts 3; small circular holes arranged in a square array are also provided at the geometric center, which can be used to integrate sensing modules or other auxiliary equipment. Through the above-mentioned synergistic design of lightweight and high strength, the flexible rope fixing plate 1 effectively reduces the ineffective payload of the UAV while ensuring the safety redundancy of hoisting, and improves the endurance and effective payload ratio. At the same time, it provides a high-rigidity reference anchor point for the dynamic modeling based on the structural parameters of the rope ladder in the aforementioned embodiment, and reduces the model error introduced by the elastic deformation of the mounting plate itself.

[0051] The UAV connector mounting plate 2 is located on the back of the flexible rope fixing plate 1. Specifically, the UAV connector mounting plate 2 serves as the mechanical interface between the hanger and the UAV fuselage. Its positional layout ensures that the center of gravity of the entire lifting system is as close as possible to the UAV's center of mass, thereby reducing the coupling torque of the lifting load on the UAV's attitude control. In practical applications, the UAV connector mounting plate 2 typically features multiple sets of adjustable-spacing mounting holes or groove structures to accommodate the spacing differences of carbon fiber tubes in the frames of different brands and models of UAVs, achieving universal compatibility of the hanger device. For the connection method, a distributed fixing using multiple bolt groups, combined with reinforced welded joints or an integrated molding process, ensures that the connection interface does not loosen or fail due to fatigue under high-dynamic flight maneuvers. This modular and standardized interface design not only improves the interchangeability and maintenance convenience of the equipment but also provides definite mechanical boundary conditions for obtaining accurate UAV dynamic parameters in the aforementioned embodiments.

[0052] Two insulated rope ladders 4 are connected at one end to a flexible rope fixing plate 1. Specifically, the two insulated rope ladders 4 are arranged in parallel, and their upper ends are fastened to the mounting holes at both ends of the flexible rope fixing plate 1 by eye bolts 5 and eye bolt fixing nuts 3, respectively. The annular head of the eye bolt 5 provides a flexible hinge fulcrum for the rope ladder, allowing the ladder to swing freely within a certain range to adapt to load movement, while limiting lateral slippage. The insulated rope ladder 4 itself is woven or crimped with high-strength insulating rope and ladder rods, possessing both electrical insulation performance and flexible load-bearing capacity. It should be understood that although this embodiment uses two insulated rope ladders as an example, in other embodiments, depending on the load weight and stability requirements, three or more rope ladders can be used to form a polygonal hoisting configuration, as long as the functional requirements of flexible connection and geometric stability can be met. The core advantage of using two insulated rope ladders is that they, together with the upper and lower connection points, form a natural planar triangular stable system. Compared with single rope suspension, this configuration geometrically eliminates the degree of freedom of the load rotating around the vertical axis, significantly suppressing torsional oscillation modes, and providing a physical basis for simplifying the dynamic model dimensions and improving real-time calculation efficiency in the aforementioned embodiments.

[0053] The rope ladder connecting frame 6 connects to the other ends of the two insulated rope ladders 4. Specifically, the rope ladder connecting frame 6 is located at the lower end where the two insulated rope ladders 4 meet, serving as a force convergence and rigid transition. In a preferred embodiment, the rope ladder connecting frame 6 is integrally formed from Q235 structural steel using a precision welding process, and its surface is coated with a composite protective coating system of spray paint and baking paint. The high yield strength of the steel ensures that the connecting frame does not undergo plastic deformation under rated load and impact conditions, while the composite coating effectively resists the erosion of the welded joints by rainwater, ultraviolet rays, and chemical corrosion, fundamentally eliminating the risk of structural strength degradation due to rust. The design of the rope ladder connecting frame 6 also needs to consider the reliability of the connection with the insulated rope ladders 4. For example, the ends of the rope ladders can be firmly locked through methods such as compression sleeves, braiding, or special clamps to prevent loosening under long-term alternating loads.

[0054] The general-purpose maintenance robot hook 7 is connected to the rope ladder connecting frame 6, and its end has an opening structure for mounting. Specifically, the general-purpose maintenance robot hook 7 is the final interface between the mounting device and the load being lifted. One end is securely mounted on the robot's aluminum frame, and the other end is connected to the rope ladder connecting frame 6. The connection point is typically reinforced with a triangular structure to enhance overall stability. The opening structure at the end of the hook 7 features a large bevel design, providing ample guide clearance. This feature is crucial for achieving autonomous and precise mounting: when the UAV approaches the target guide wire under visual servo guidance, the large bevel can tolerate a certain range of positioning deviations, using the beveled guide effect to automatically correct the hook to the correct mounting posture, reducing the extreme dependence on flight control precision. Simultaneously, the hook 7 itself is also made of Q235 material and reinforced with a double-protection process, ensuring that it maintains structural integrity and functional reliability even under repeated hooking / unhooking friction and outdoor exposure.

[0055] Two insulated rope ladders 4, a flexible rope fixing plate 1, and a rope ladder connecting frame 6 form a double-rope ladder triangular hoisting configuration. Specifically, this configuration means that, on the side view projection plane, the two insulated rope ladders 4 serve as the two hypotenuses of the triangle, the line connecting the two connection points on the flexible rope fixing plate 1 serves as the top edge, and the rope ladder connecting frame 6 serves as the base point, together forming a stable triangular geometric frame. This passive stabilization structure and the active anti-sway control method in the aforementioned embodiment form a deep synergistic effect: on the one hand, the geometric constraints provided by the triangular configuration naturally suppress the lateral sway and torsion of the load, reducing the initial disturbance energy that the active control system needs to overcome; on the other hand, the equivalent swing length and geometric factor determined by this configuration directly participate in the construction of the rigid-flexible coupling dynamic model, enabling the control algorithm to accurately predict the system frequency and implement targeted sway suppression. The hanging device provided in this embodiment is not only a physical load-bearing component but also an indispensable part of the active anti-sway control strategy. Together, they constitute a guarantee system for passive structure sway suppression and active control sway elimination, ensuring that the general-purpose maintenance robot is safely, smoothly, and accurately hoisted to the target work point under complex weather conditions.

[0056] Example 8 This embodiment provides a drone hoisting system, which integrates the hardware structure and software algorithms of the aforementioned embodiments to form a complete intelligent hoisting operation platform. Specifically, the drone hoisting system includes a drone, a drone hoisting rack device as described in Embodiment 7, and a flight control system.

[0057] The UAV hoisting rack device is securely mounted to the underside of the UAV fuselage via its UAV connecting frame mounting plate 2, forming a rigid connection interface. The flight control system is configured to execute the active anti-sway control method for UAV load hoisting as described in any of Embodiments 1 to 5 above. In a preferred embodiment, the flight control system can directly reuse the UAV's original flight control computer, implementing the anti-sway control function through firmware upgrades or loading dedicated algorithm modules; alternatively, the flight control system can be mounted on the UAV fuselage as an independent mission computer, communicating with the UAV's underlying flight control system via a CAN bus or serial port. Regardless of the hardware deployment method, the flight control system serves as the decision-making core of the entire hoisting system, receiving data streams in real time from the UAV's own sensors and optional external sensing devices.

[0058] During operation, the flight control system first constructs and maintains a rigid-flexible coupled dynamic model reflecting the current physical state based on acquired sensor data and pre-stored rope ladder structural parameters. Subsequently, the system continuously runs an extended state observer to identify the time-varying equivalent axial stiffness and equivalent structural damping of the rope ladder online, and dynamically updates the pulse parameters of the feedforward input shaper and the gain of the feedback controller cycle by cycle based on the identification results. The generated anti-sway control commands are ultimately converted into UAV motor speed adjustment signals or attitude correction commands, driving the UAV to perform corresponding flight maneuvers. Simultaneously, the pylon device installed below the UAV, with its double-rope ladder triangular hoisting configuration and the hollowed-out bracing structure of the flexible rope fixing plate 1, provides the system with passive geometric stability and a lightweight load-bearing foundation. This deep synergy of "active algorithm + passive structure" allows the system to maximize the suppression of load sway while ensuring hoisting safety.

[0059] Taking the hoisting of a general maintenance robot to an eight-split conductor in a power line inspection scenario as an example, the typical workflow of this UAV hoisting system is as follows: During the takeoff and cruise phase, the flight control system automatically triggers a preset maneuvering excitation to complete the self-calibration of the rope ladder parameters, and uses a zero-vibration derivative shaper to ensure basic sway suppression performance during rapid transfer; when the system enters the near-tower approach phase, the flight control system dynamically switches to a higher-order shaper based on the real-time identified parameters, and actively enhances the swing damping, in conjunction with the triangular configuration constraint of the mounting device, to converge the load swing angle to a safe range; during the precise loading phase, the flight control system activates a high-frequency updated additional insulation shaper and tension closed-loop monitoring logic. Once abnormal tension fluctuations caused by contact collisions are detected, the loading action is immediately suspended until the state stabilizes, and the large bevel design at the end of the hook 7 of the general maintenance robot provides fault-tolerant guidance for the final mechanical docking; if a sudden fault occurs and the system enters the emergency hovering phase, the flight control system switches to the optimal energy dissipation control mode, using the rope ladder's own structural damping to quickly calm the swaying vibration. Through the above-mentioned integrated hardware and software system design, the UAV hoisting system provided in this application effectively solves the problems of low anti-sway accuracy and poor safety caused by model mismatch in the prior art, and realizes the autonomous, precise and safe hoisting of general maintenance robots under complex working conditions.

[0060] Example 9 like Figure 4 As shown, this embodiment provides an autonomous and precise mounting control method based on the aforementioned mounting device. This method is executed by the UAV flight control system in conjunction with the onboard sensing module, aiming to achieve autonomous docking and eliminate reliance on manual remote control. The method mainly includes the following steps.

[0061] Step S5, Sensing System Configuration. Specifically, a sensing module is integrated below the flexible rope fixing plate 1, including a binocular stereo vision camera, a single-line LiDAR, and an LED illumination module. The binocular stereo vision camera is mounted on both sides of the bottom surface of the flexible rope fixing plate 1, with a baseline distance of 200mm and a resolution of 1280×720@30fps, used to acquire stereo image information of the target guide wire. The single-line LiDAR is installed at the central axis of the flexible rope fixing plate 1 for accurate distance measurement, compensating for the shortcomings of pure vision solutions in depth estimation. The LED illumination module is installed in a ring around the camera to ensure high-quality image data can still be obtained under adverse lighting conditions such as backlight, shadow, or nighttime. The above three sensing devices work together to provide multi-source redundant sensing input for subsequent guide wire identification and positioning.

[0062] Step S6, Conductor Recognition and Localization. Specifically, a lightweight semantic segmentation network, MobileNetV3-DeepLabV3+, is used to run in real-time on an airborne embedded platform to perform pixel-level segmentation of conductors in the binocular images. This network has been trained on a large amount of power line inspection scenario data and can accurately extract conductor regions in complex backgrounds. Based on the segmentation results, the centerline coordinates of the conductors in the left and right views are extracted respectively, and the 3D point cloud data of the conductors is obtained through a stereo matching algorithm. Subsequently, the depth estimation error of the point cloud is corrected by fusing the ranging data of the LiDAR, and the corrected point cloud is parabolic fitted to account for the sag effect of the conductors, thereby improving the conductor localization accuracy. On this basis, for the special configuration of the eight-split conductor, the system automatically identifies the arrangement configuration of the sub-conductors in the 3D point cloud and calculates the target pose of the optimal mounting point, providing a precise target reference for subsequent visual servo control.

[0063] Step S7, three-stage visual servo control. Specifically, based on the distance between the UAV and the target guideline, the mounting process is divided into three progressive control stages, each employing a different control strategy to balance efficiency and accuracy.

[0064] The first stage is the coarse alignment stage. This stage primarily uses GPS / RTK high-precision positioning, supplemented by visual perception, to guide the UAV to fly directly above the target guideline. During this stage, the vision system locks onto and tracks the target guideline, establishing a stable visual servo reference. The main task of this stage is to rapidly reduce the spatial distance between the UAV and the target while ensuring the vision system can continuously and stably acquire the target.

[0065] The second stage is the precision alignment stage. In this stage, the system switches to pure vision servo control, and the control law is as follows: ; in, For drone speed commands, The positional error between the current pose and the target pose. The rate of change of position error, This is an estimate of the velocity of the target conductor. , , These are the proportional gain, derivative gain, and feedforward gain, respectively. The speed command output by this control law is then fed into the flight control system after speed limiting and acceleration smoothing, driving the UAV to approach the target attitude in a smooth and controllable manner.

[0066] The third stage is the compliant loading stage. In this stage, the system switches to impedance control mode and establishes a virtual spring-damper model. ; in, For virtual contact force, , , These are virtual mass, virtual damping, and virtual stiffness parameters, respectively. This refers to the positional deviation between the hook and the conductor. When the hook contacts the conductor, impedance control allows for a compliant deflection of the hook, absorbing contact impact energy and preventing equipment damage or conductor damage caused by rigid collisions. The contact force is indirectly estimated through a sudden change in flight control thrust, eliminating the need for a force sensor on the hook.

[0067] Step S8: Loading Confirmation and Anomaly Handling. Specifically, the system employs a dual mechanism to determine successful loading. The visual confirmation mechanism requires the overlap between the hook area and the guide wire area in the image to remain stable for more than 2 seconds to eliminate momentary misjudgments. The mechanical confirmation mechanism determines this by monitoring changes in the estimated thrust of the UAV—when the hook is successfully loaded, the load weight disappears from the UAV thrust estimate, and the thrust requirement decreases significantly. If both confirmations pass, the loading is considered successful, and the process ends. If confirmation fails, the system automatically executes a safe retreat action, then feeds back to the precision alignment stage via a dotted-line loop to retry, with a maximum of 3 retries. If it still fails, the return-to-home logic is triggered to ensure operational safety.

[0068] Example 10 like Figure 5 As shown, this embodiment provides a multi-objective optimization design method for the structural parameters of the aforementioned hanger device. This method aims to achieve optimal lightweighting of the hanger through systematic optimization design, while meeting strength, stiffness, and assembly requirements. The method mainly includes the following steps.

[0069] Step S9, define design variables. Specifically, extract the following 7 key structural parameters as optimization variables: flexible rope fixing plate, main plate thickness. , tensioner thickness Width of tie rod Hole diameter Number of perforated holes Effective length of rope ladder Spacing between the connection points of the double rope ladder These variables cover the core geometric parameters that affect the weight, stiffness, and dynamic characteristics of the mounting bracket, and their values ​​are determined based on engineering experience and the mechanical properties of materials.

[0070] Step S10: Construct the objective function. Specifically, three optimization objectives are established: Objective 1 is to minimize the total mass of the structure, and its expression is: ; in, and The densities of aluminum alloy and steel are respectively. and The volumes of the flexible rope fixing plate and the rope ladder connecting frame are respectively. For fastener quality.

[0071] Objective 2 is to maximize the equivalent stiffness, that is: ; in, It is the reciprocal of the maximum deformation of the critical node under rated load, characterizing the ability of the hanger to resist elastic deformation.

[0072] Target 3 is to avoid the inherent frequency, that is: ; in, The first-order bending natural frequency of the pylon is the frequency difference between the frequency through which the UAV rotor passes; the goal is to keep the first-order bending natural frequency of the pylon away from the frequency band through which the UAV rotor passes, thus avoiding resonance. There is an inherent contradiction among the three objectives, requiring multi-objective optimization to find the Pareto optimal solution set.

[0073] Step S11: Set constraints. Specifically, these include four types of constraints: Strength constraints: Under a safety factor of 1.5, the maximum Von Mises stress shall not exceed 220.8 MPa; Fatigue constraints: Under 10^6 cycles of loading, the stress amplitude of the welded joint shall not exceed the fatigue limit of Q235 steel; Geometric constraints: The minimum bridge width between the hollow holes shall not be less than twice the width of the tie rod to prevent local buckling; Assembly constraints: The spacing between connection points must be compatible with the carbon tube spacing range of the target UAV to ensure the universal compatibility of the mounting bracket.

[0074] Step S12, Improved NSGA-II Solution. Specifically, the improved Non-Dominated Sorting Genetic Algorithm II (NSGA-II) is used to solve the above multi-objective optimization problem. Improvements include: using adaptive crossover and mutation probabilities with exponential distributions to dynamically adjust the search intensity based on population diversity; employing a continuous-integer hybrid encoding strategy to handle both continuous and discrete variables; constructing a finite element surrogate model based on Kriging interpolation, initializing the training set with 120 sets of Latin hypercube sampling, and correcting the surrogate model every 20 generations using finite element analysis results, significantly reducing computational costs; and using an adaptive penalty function to handle constraints, transforming the degree of constraint violation of infeasible solutions into a penalty term added to the fitness function to guide the population towards the feasible region.

[0075] Step S13, Design Output. Specifically, the algorithm outputs a Pareto optimal solution set, with each solution corresponding to a specific combination of structural parameters. Engineers can select a suitable solution from the solution set according to actual needs. Simultaneously, the algorithm outputs parameterized design criteria, including the optimal ratio range of plate thickness to tie rod width, recommended perforation rate range, and the optimal ratio between rope ladder length and connection point spacing. Compared to traditional experience-based design, this design effectively avoids the risk of resonance.

[0076] The three control / optimization methods mentioned above work synergistically with the aforementioned hanger structure: the length of the rope ladder and the spacing between connection points output by the structural parameter optimization method directly determine the equivalent swing frequency of the hoisting system. This frequency parameter is used to configure the input shaper in the active anti-sway control method. After the active anti-sway control converges the load swing angle to ±3°, the autonomous precision mounting control method can then initiate the precise docking procedure. These three methods, together with the hanger structure device, constitute a complete UAV hoisting general-purpose maintenance robot technology system.

[0077] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An active anti-sway control method for unmanned aerial vehicle (UAV) payload lifting, characterized in that, include: The dynamic parameters of the UAV, the payload mass parameters, and the structural parameters of the rope ladder are obtained, wherein the structural parameters characterize the flexibility of the rope ladder; Based on the UAV's dynamic parameters, load mass parameters, and the rope ladder's structural parameters, a rigid-flexible coupled dynamic model is constructed, consisting of the UAV, the rope ladder, and the load. The structural parameters include axial elastic parameters. The process of constructing a rigid-flexible coupled dynamic model of the UAV, the rope ladder, and the load based on the UAV's dynamic parameters, the load's mass parameters, and the rope ladder's structural parameters includes: Based on the axial elastic parameters, an elastic constraint equation considering axial elastic deformation is constructed to characterize the coupling relationship between the elastic deformation and load motion of the ladder in the rigid-flexible coupling dynamic model. During the drone lifting process, the time-varying structural parameters of the rope ladder are identified based on the rigid-flexible coupling dynamic model, including: The equivalent axial stiffness and equivalent structural damping of the rope ladder are used as the extended state to construct an extended state observer; The time-varying structural parameters are jointly estimated using the extended state observer, along with the swing angle of the load and environmental wind disturbance. Based on the rigid-flexible coupling dynamic model and the time-varying structural parameters, the active anti-sway control strategy is dynamically updated to control the UAV to perform anti-sway actions; the dynamic updating of the active anti-sway control strategy based on the rigid-flexible coupling dynamic model and the time-varying structural parameters includes: Based on the equivalent axial stiffness in the time-varying structural parameters identified in real time and the real-time tension of the rope ladder, the system's equivalent oscillation frequency and total equivalent damping ratio under the current working condition are calculated. Based on the system's equivalent oscillation frequency and the total equivalent damping ratio, the pulse parameters of the feedforward input shaper are dynamically updated to generate a shaped reference trajectory.

2. The active anti-sway control method for unmanned aerial vehicle (UAV) load lifting according to claim 1, characterized in that, The structural parameters also include: original length of a single rope ladder, axial elastic modulus, equivalent load-bearing cross-sectional area, axial stiffness, material structural damping coefficient, mass per unit length, ladder pitch, and equivalent diameter of the ladder rod; The elastic constraint equation is as follows: the distance between the load position and the connection point at the top of each rope ladder is equal to the sum of the original length of a single rope ladder and the elastic elongation of the rope ladder, wherein the elastic elongation is the ratio of the real-time tension to the axial stiffness of the rope ladder. The equivalent swing frequency corresponding to the rigid-flexible coupling dynamic model is determined by the geometric relationship between the gravitational acceleration, the equivalent pendulum length, and the distance between the upper connection points of the double rope ladder and the original length of the single rope ladder; wherein, the equivalent pendulum length is the sum of the ratios of the original length of the single rope ladder to the static tension of the load and the axial stiffness.

3. The active anti-sway control method for unmanned aerial vehicle (UAV) load lifting according to claim 1, characterized in that, The state vector of the extended state observer includes: load longitudinal swing angle, load lateral swing angle, longitudinal swing angular velocity, lateral swing angular velocity, equivalent wind disturbance force component, equivalent axial stiffness, and equivalent structural damping; The coupling force term of the elastic deformation of the rope ladder and the equivalent structural damping is added to the observation model of the extended state observer, and the real-time tension of the rope ladder is calculated using the thrust estimate of the UAV as an auxiliary observation. During takeoff, the initial structural parameters of the rope ladder are self-calibrated by performing preset maneuvering excitation.

4. The active anti-sway control method for unmanned aerial vehicle (UAV) load lifting according to claim 1, characterized in that, The pulse parameters of the dynamically updated feedforward input shaper include: The pulse amplitude and time interval of the feedforward input shaper are updated cycle by cycle, and the swing angle phase-frequency prediction stage is introduced to correct the pulse parameters in advance.

5. The active anti-sway control method for unmanned aerial vehicle (UAV) load lifting according to claim 1, characterized in that, The active anti-sway control strategy, which is dynamically updated based on the rigid-flexible coupling dynamic model and the time-varying structural parameters, further includes: In the reference trajectory planning, allowable tension constraints for the rope ladder are added, so that the upper limit of trajectory acceleration is dynamically adjusted according to the real-time tension of the rope ladder, in order to avoid the rope ladder tension exceeding the limit and reduce elastic oscillations caused by dynamic tension; in the feedback control, the maximum allowable tension constraint of the rope ladder is introduced, and the maximum allowable tension constraint of the rope ladder and the swing angle constraint are used together to construct an extended obstacle function, so that the swing angle and the rope ladder tension are simultaneously constrained within the safety threshold, and the rope ladder tension is monitored in a closed loop. The wind resistance of the rope ladder is calculated based on its structural parameters, and the wind resistance of the rope ladder and the load wind disturbance are jointly included in the feedforward compensation term.

6. The active anti-sway control method for unmanned aerial vehicle (UAV) load lifting according to claim 5, characterized in that, The method also includes a phased control strategy that switches control parameters according to the operation phase, which includes the takeoff cruise phase, the tower approach phase, the precision loading or unloading phase, and the emergency hovering phase. During the takeoff and cruise phase, the rope ladder parameters are self-calibrated, and a feedforward input shaper with zero vibration derivative configuration is used. During the near-tower approach phase, the higher-order pulse parameters of the feedforward input shaper are dynamically updated, and active enhancement logic for the rope ladder swing damping is added. During the precise mounting or unmounting phase, the feedforward input shaper is updated at a high frequency, and the closed-loop monitoring of the rope ladder tension is activated. When the tension fluctuation exceeds a preset threshold, the mounting action is paused. During the emergency hovering phase, optimal energy dissipation control logic based on the structural damping characteristics of the rope ladder is incorporated.

7. An active anti-sway control device for unmanned aerial vehicle (UAV) payload lifting, characterized in that, include: A memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the active anti-sway control method for unmanned aerial vehicle (UAV) lifting loads as described in any one of claims 1 to 6.

8. A drone hoisting system, characterized in that, include: Drones; A drone lifting rack device is installed on the drone, and the drone lifting rack device includes: A flexible rope fixing plate, wherein the flexible rope fixing plate is provided with a hollowed-out tensioning structure; The drone connector mounting plate is located on the back of the flexible rope fixing plate; Two insulated rope ladders, one end of which is connected to the rope fixing plate; A rope ladder connecting frame connects to the other ends of the two insulated rope ladders respectively; A universal maintenance robot hook is connected to the rope ladder connecting frame, and the end of the universal maintenance robot hook is provided with an opening structure for hanging; The two insulated rope ladders, the flexible rope fixing plate, and the rope ladder connecting frame form a double rope ladder triangular hoisting configuration; The flight control system is configured to perform the active anti-sway control method for unmanned aerial vehicle (UAV) payload lifting as described in any one of claims 1 to 6.

Images