A flight control and load swing suppression method for a four-rotor unmanned aerial vehicle hanging load system

Abstract

The application discloses a flight control and load swing suppression method of a four-rotor unmanned aerial vehicle hanging load system, belongs to the technical field of unmanned aerial vehicle flight control, and aims at the influence of external environment uncertain factors, and proposes a robust control strategy based on adaptive multilayer neural network compensation disturbance estimator error, so that accurate tracking of a preset track is realized under the condition of load swing disturbance and wind disturbance; aiming at the accurate load transportation and dropping task demand, all states of the system are expressed by using flat output, a mapping relationship from the load movement track to the unmanned aerial vehicle movement track is established, so that the unmanned aerial vehicle drives the load to move in the expected track; aiming at the load swing existing when the system is in flight and the load residual oscillation existing when the unmanned aerial vehicle is in hovering, a load swing suppression technology based on a ZVDD shaper is proposed, so that the load swing is reduced.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) flight control technology, specifically relating to a flight control and load sway suppression method for a quadcopter UAV carrying a payload for transportation missions. Background Technology

[0002] Quadrone drones, as a highly integrated type of unmanned aerial vehicle (UAV), have advantages such as simple structure, vertical take-off and landing capability, and low-speed flight, and have been widely used in fields such as cargo handling. Among them, the use of sling loads can achieve more efficient cargo transportation, which has become a research focus of scholars at home and abroad in recent years.

[0003] The successful completion of transportation missions requires the quadcopter drone flight control system to have strong control precision and robustness. For example, when performing cargo transportation missions, quadcopter drones need to be able to withstand wind to cope with various adverse weather conditions. In addition, the swing of the load during flight will significantly affect the stability of the flight control system, and the residual oscillations after reaching the target point will also reduce the quality of the transportation mission.

[0004] Therefore, researching flight control and sway suppression methods for quadrotor UAV load-bearing systems, ensuring that the UAV accurately tracks the desired trajectory while suppressing load sway, has significant practical application value for improving the quality of operational tasks and expanding the application scenarios of UAVs. Summary of the Invention

[0005] This invention provides a method for flight control and load sway suppression of a quadcopter unmanned aerial vehicle (UAV) payload system, which ensures that the UAV accurately tracks the desired trajectory while suppressing load sway.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for flight control and load sway suppression of a quadcopter unmanned aerial vehicle (UAV) payload-carrying system includes the following steps:

[0008] Step 1: Establish a dynamic model of the quadrotor UAV's payload-carrying system using the Newton-Euler method and the Euler-Lagrange energy method;

[0009] Step 2: A cascaded trajectory tracking control strategy is adopted, considering load swing as interference to the UAV. The system is decoupled into an outer loop position control loop and an inner loop attitude control loop. For external lumped interference present in the two loops, interference estimators are designed to improve the robustness of the flight control system. To address the large interference estimation bias in the outer loop position loop interference estimator, a multi-layer adaptive neural network output compensation strategy is designed to reduce the estimation error and improve the tracking accuracy of the system.

[0010] Step 3: Using the load trajectory generation technology based on the principle of system differential flatness, the load position and the yaw angle information of the UAV are used as the flat output of the quadcopter UAV load system. The flat output and its higher-order derivatives are used to represent all states of the system, establish the mapping relationship between the load motion trajectory and the UAV motion trajectory, and generate the expected load trajectory.

[0011] Step 4: To address the load swaying issues during system flight and the residual load oscillations during UAV hovering, a ZVDD shaper is designed to suppress the residual oscillations of the system; it is then mapped to the desired trajectory of the UAV, which drives the load to move along the desired trajectory.

[0012] Beneficial Effects: This invention provides a method for flight control and load sway suppression of a quadcopter UAV payload system. Addressing the challenges of nonlinear control-based controller design, such as insufficient robustness to noise and disturbances and the inability to completely eliminate steady-state errors, this invention offers a novel solution for the successful delivery of payload systems by quadcopter UAVs. It employs inner and outer loop cascade control technology to enhance robustness against external lumped disturbances. Furthermore, to address load sway during flight and residual load oscillations during UAV hovering, the invention utilizes the system's differential flatness principle, taking the load position and UAV yaw angle information as the basis for the quadcopter UAV payload system's performance. The system's flat output and its higher-order derivatives are used to represent all system states, establishing a mapping relationship between the load's trajectory and the UAV's trajectory to generate the desired trajectory of the load. Input shaping technology is used to suppress residual oscillations in the system. The desired trajectory signal of the quadrotor UAV is calculated using the established mapping relationship. Simulation results show that for quadrotor UAV load-bearing systems, this invention can effectively achieve transportation tasks under external multi-source interference and provide good control accuracy and robustness. It can effectively achieve tracking control quality of the preset trajectory and suppress load sway, providing a complete control strategy for quadrotor UAV load-bearing systems oriented towards transportation tasks. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the rigid body model of the quadcopter UAV load-bearing system in an embodiment of the present invention;

[0014] Figure 2 This is a block diagram of the attitude loop controller in an embodiment of the present invention;

[0015] Figure 3 This is a block diagram of the position loop controller in an embodiment of the present invention;

[0016] Figure 4 This is a diagram of a multi-layer neural network structure used to compensate for lumped interference estimation errors in an embodiment of the present invention;

[0017] Figure 5 This is a simulation diagram of the attitude loop in an embodiment of the present invention;

[0018] Figure 6 This is a simulation diagram of the position loop in an embodiment of the present invention;

[0019] Figure 7 This is a schematic diagram of the shaping principle in an embodiment of the present invention;

[0020] Figure 8 This is a schematic diagram of the position tracking of a quadcopter UAV, used as a control group in this embodiment of the invention.

[0021] Figure 9 This is a schematic diagram of the position tracking of a quadcopter UAV in the embodiment of the present invention.

[0022] Figure 10 This is a schematic diagram of attitude tracking for a quadcopter UAV, used as a control group in this embodiment of the invention.

[0023] Figure 11 This is a schematic diagram of attitude tracking for the quadcopter UAV in an embodiment of the present invention.

[0024] Figure 12 The control group and experimental group φ in this embodiment of the invention L Status comparison chart;

[0025] Figure 13 The control group and experimental group θ in this embodiment of the invention L Status Comparison Chart

[0026] Figure 14 This is a schematic diagram of the integrated control technology for flight control and load sway suppression of the quadcopter UAV load-bearing system in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments:

[0028] like Figure 14 As shown, a method for flight control and load sway suppression of a quadcopter unmanned aerial vehicle (UAV) payload system specifically includes the following steps:

[0029] 1. Quadrotor UAV trajectory tracking control technology based on adaptive neural network

[0030] (1) Modeling of the quadcopter UAV's sling load system

[0031] Quadrone drones performing payload tasks involve numerous performance indicators and parameters, such as steady-state error in tracking a preset trajectory during flight and anti-interference capabilities. These performance requirements and parameter ranges directly determine whether the drone can safely and quickly transport the payload to the target location. The system rigid body model, for example... Figure 1 As shown, the six-degree-of-freedom rigid body dynamics model of the quadcopter UAV is first expressed as:

[0032]

[0033] Where F represents the vertical lift provided by the four motors; [τ φ τ θ τ ψ ] T Represents the rotational torque along the three axes of the aircraft; [xyz] T Represents the position of the quadcopter drone in an inertial coordinate system; [φ θ ψ] T m is the rotational attitude angle; Q I represents the total mass of the drone; g represents the acceleration due to gravity; I represents the total mass of the drone. xx I yy I zz These represent the inertial moments along the three directions of the body coordinate system, respectively; J represents the rotational inertia matrix, and d... φ Represents the lumped interference of the pitch channel, d θ Represents the lumped interference of the roll angle channel, d ψ Represents the lumped interference of the yaw angle channel, d x Represents the lumped interference of channel x, d y Represents the lumped interference of the y-channel, d z This represents the lumped interference in the z-channel.

[0034] Combining formula (1), the kinematic model of the quadcopter UAV load-bearing system established using the Euler-Lagrange energy method is as follows:

[0035]

[0036] Where, m Q The mass of the quadcopter drone; m L For the mass of the load; F provides vertical lift for the four motors; D Q D represents the lumped disturbance vector experienced by the quadcopter UAV during flight. L Let be the lumped disturbance vector experienced during payload flight; T be the tension vector; and G be the gravity vector [0 0g]. T Where g is the acceleration due to gravity; P = [xyz] T P is the position vector of the UAV in the geodetic coordinate system. L =[x L y Lz L ] T This is the position vector of the load in the geodetic coordinate system;

[0037] (2) Attitude loop tracking control law design

[0038] The attitude loop control block diagram is as follows: Figure 2 As shown, define attitude state variables φ, θ, ψ; the desired attitude is φ. r θ r , ψ r Define Θ = [φ θ ψ] T Θ r =[φ r θ r ψ r ] T Attitude subsystem tracking error e Θ =Θ r -Θ; The derivative of the attitude subsystem tracking error is The general form of a control system is:

[0039]

[0040] in, u Θ =[τ φ τ θ τ ψ ] T The lumped disturbance of the attitude subsystem is represented as D. Θ =[D φ D θ D ψ Specifically:

[0041]

[0042] The attitude subsystem filter tracking error is selected as follows:

[0043]

[0044] Where, λ Θ =[λ φ λ θ λ ψ ] T The design parameters are non-negative. Taking the derivative of formula (5) and combining it with formulas (3) and (4), we get...

[0045]

[0046] Design attitude robust control law Approximating the lumped disturbance of the attitude loop, substituting it into formula (6) yields...

[0047]

[0048] Design using feedback linearization Where k Θ The non-negative control parameter is used to ensure the convergence of the attitude subsystem filter tracking error. Substituting it into (7) yields...

[0049]

[0050] As shown in Equation (8), the lumped disturbance of the attitude subsystem can be estimated from the current state of the system and the desired input; further, the estimation of the lumped disturbance of the attitude subsystem is expressed as follows:

[0051]

[0052] in, For a first-order low-pass filter, a robust control law It can be expressed as

[0053]

[0054] in, Using the inverse Laplace transform operator, and combining formulas (6) to (10), the output of the attitude loop control law can be obtained as follows:

[0055]

[0056] in, It is a first-order low-pass filter;

[0057] (3) Design of position loop trajectory tracking control law

[0058] Position loop control block diagram as follows Figure 3 As shown, the desired position state is defined as P. r =[x r y r z r ] T The tracking error of the position subsystem is e P =PP r , The tracking error of the position subsystem filter is Where, λ P =[λ x λ y λ z ] T The parameters to be designed; design a robust control law for the position subsystem. Where D P =[d x d y d z ] T ,get

[0059]

[0060] Further obtain

[0061]

[0062] in, λ P The control parameters to be designed;

[0063] Define the lumped disturbance estimation error of the location subsystem as:

[0064]

[0065] External time-varying interference signals can reduce estimation accuracy. Therefore, the universal approximation property of neural networks is used to approximate the estimation error and improve the estimation accuracy. The designed neural network structure is as follows: Figure 4 As shown, the neural network includes an input layer, hidden layer 1, hidden layer 2, and an output layer. The number of neurons in each part is N1 = 6, N2 = 3, and N2+1 = 4, respectively. By selecting appropriate input information, the neural network can achieve the desired output layer. Since external lumped interference affects the state of the quadcopter UAV, the estimation error information depends on the state information of the quadcopter UAV during flight. Therefore, the input layer is selected... Includes all current state information of the position subsystem; the mapping from the input layer to the first hidden layer is defined as follows: weight matrix The activation function is used to map to the second hidden layer and a bias is added, i.e. The activation function is expressed in the form σ(k)=1 / (1+e -k ), k∈{1,...,N2}.

[0066] The output of the neural network is

[0067]

[0068] Where the weight matrix Using estimated values and To replace ideal weights and assumed and There exists a non-negative upper bound.

[0069]

[0070] Accordingly, the hidden layer is determined by the estimated value. Instead of α and β, choose the network weight update law.

[0071]

[0072] in, These are the weight update rates of the two hidden layers, δ=[ε x ε y ε z ] T The interference estimator estimation error for the three channels representing the location; γ w γ v Let κ be the parameters to be determined. By selecting appropriate parameters so that the output of the neural network can effectively compensate for the lumped disturbance estimation error, the control law of the position subsystem is obtained as follows:

[0073]

[0074] 2. System Differential Flat Load Trajectory Generation Technology

[0075] For a high-dimensional nonlinear system like a quadcopter drone's payload system, flat output can be used to effectively reduce the system's dimensionality, including the instantaneous transition from one dynamic state to another. This allows the system to still maintain control over all degrees of freedom by controlling the rotational speeds of the quadcopter's four motors.

[0076] P Q =P L -lp (19)

[0077]

[0078] Where l is the rope length, p is the unit vector pointing from the load to the drone, and further we can obtain

[0079]

[0080]

[0081] Where, φ L θ L Let P be the load swing angle, when the load position is P L Given the condition, Tp can be determined by formula (20); p can be determined by formula (21); and the position P of the UAV can be determined by formula (19). The mapping from the load motion state to the required UAV motion state can be found by utilizing the system differential flatness characteristic.

[0082] Therefore, based on the above analysis, the trajectory planning problem of the quadcopter UAV payload system can be solved in [x] L y L z LThe plan is made within the ψ] space, with the goal of generating a trajectory for stable load movement, and then obtaining a smooth trajectory that the UAV can travel on, which is then passed to the controller as a reference.

[0083] 3. Open-loop load sway suppression technology based on input shaping

[0084] The key to the input shaper is calculating the amplitude of each pulse signal and their respective time delays, the principle of which is as follows: Figure 7 The mathematical expression for the input shaper is:

[0085]

[0086] Where C(s) represents the system response, A i Let s represent the pulse at time i, and s represent the Laplace transform of the pulse at the corresponding time. Simplifying the quadcopter UAV payload system into a second-order system, we obtain its typical transfer function as follows:

[0087]

[0088] Where, ω n Let ζ be the system's natural frequency and ζ be the system's damping. Then the system's unit impulse response is:

[0089]

[0090] Where, ω d The damped oscillation frequency of the system. The shaper contains a series of pulses with different amplitudes and delays. Let the amplitude and delay of the pulse at time i be A. i and t i The transient response caused by the pulse is

[0091]

[0092] Where t represents the current time, t i Representing the response times of each impulse, the total system response is the sum of the responses of each component, expressed as:

[0093]

[0094] The percentage of residual oscillations V is defined as...

[0095]

[0096] in

[0097]

[0098] The purpose of the shaper is to divide the original signal into several segmented signals to excite the system, so that V is 0, that is, to satisfy...

[0099]

[0100] By making the percentage of residual vibration equal to zero, ∑A is guaranteed. i The constraint t1 = 1 is met, and the position of the first basic pulse is set to t1 = 0, thereby determining the amplitude of the pulse (A1 and A2) and the excitation time t2 of the second pulse.

[0101] The ZVDD shaper is designed to shape the input signal. To eliminate residual vibrations in the flexible system, y(t) should be zero in the time domain. Therefore, the selected impulse response amplitude and excitation time are expressed as follows:

[0102]

[0103] in

[0104] Simulation verification

[0105] (1) Simulation verification of control strategy

[0106] Attitude loop simulation verification: Set the initial attitude to [φ θ ψ] T =[0 0 0] T Tracking a sinusoidal signal with an amplitude of 1, the three-channel perturbation is D. φ =0.3sin(0.1*πt)+0.1,D θ =0.2cos(0.1*πt), D ψ =0.2*sin(0.1*πt)+0.2. The tracking performance of the attitude closed-loop system for the sinusoidal continuous signal is as follows: Figure 5 As can be seen, there is a small tracking error in the initial state, but it can quickly catch up with the desired sinusoidal signal and still exhibit strong robustness under disturbances. This proves that the proposed attitude control strategy can enable the attitude system to keep up with the rapidly changing desired signal, effectively eliminate the influence of variable disturbances in the control loop, and meet the requirements of actual tasks.

[0107] Position loop simulation verification: The following continuous waypoint tracking simulation experiment verifies that the proposed control algorithm can achieve the system's ability to track the desired trajectory of a large maneuver under disturbance. In the simulation, the desired trajectory is set as x. r = (1+t)cos(0.t5,y r =(4+0.1t)sin(0.5t), z r =1+0.5t; the desired yaw angle is set to 0, so obviously at t=0, the desired position is [x r y r z r ] T=[0 0 1] T That is, at this moment, the quadcopter drone's mission is to start from its initial position [x0 y0 z0]. T The desired signal was tracked, and the simulation lasted for 30 seconds. The simulation results are as follows: Figure 6 As shown, in the initial stage, due to a 1m error between the actual UAV takeoff position and the desired altitude position, there is a certain error in the initial position as seen in the z-direction position diagram. However, the error quickly converges to 0. The system maintains high robustness under simulated high-maneuver flight turning conditions during the 13s-18s and 21s-25s time periods. The proposed control strategy enables the UAV to track continuous waypoints well under simulated strong external interference.

[0108] (2) Trajectory planning and tracking simulation verification

[0109] The external interference simulation in the experiment was random interference added to each channel. The designed quadrotor UAV trajectory tracking control technology based on adaptive neural network was used to track the desired trajectory. The corresponding control parameters were consistent with those in the control strategy simulation verification experiment. To verify the effectiveness of the input shaping load swing suppression strategy, a set of comparative experiments was proposed: First, both experiments generated the desired trajectory of the quadrotor UAV. Experiment 1 served as the control group. The desired trajectory signal of the load trajectory was obtained by mapping the differential flatness characteristic to obtain the desired trajectory signal of the UAV, which was directly used as the input of the controller to drive the UAV. Experiment 2 used the designed ZVDD shaper to shape the input signal of the desired load trajectory, and then used the processed desired signal to inversely calculate the desired trajectory signal of the UAV, which was used as the input of the controller to complete the simulation test of the experimental group. Both experiments used the designed control algorithm to track the generated desired trajectory. The specific parameters of the ZVDD shaper are shown in Table 1. The comparative experimental results are as follows: Figures 8 to 13 As shown.

[0110] Table 1. Specific parameters for the input shaping technology of the quadrotor UAV's payload system.

[0111]

[0112] from Figures 8 to 9It can be seen that, compared with the control group, the experimental group exhibited greater fluctuations in the expected trajectory along the three position channels during the 16s-24s and 24-33s time periods. However, the experimental group showed a better initial transient response than the control group, especially for the z-direction component corresponding to height. Simultaneously, all three position channels exhibited slight braking. This was due to the addition of input shaping technology, which divided the expected signal into multiple pulse signals. The filter caused a certain tilt in the curve, predicting the acceleration and deceleration of the system's motion at the beginning and end, allowing the pulse superposition to have an effect, thus suppressing the oscillation. From... Figures 10 to 11 It can be observed that input shaping significantly reduces the amplitude of system variables, except for the UAV's yaw angle, which is predefined as a flat output of a polynomial curve. Both the control and experimental methods exhibit low tracking errors for the desired trajectory, due to the good convergence and robustness of the control strategy designed based on this invention.

[0113] Depend on Figures 12 to 13 It can be seen that in the time periods of 8s-18s and 26s-36s, without the application of input shaping, the UAV performs large-amplitude maneuvers and exhibits violent oscillations in the middle of the path. However, the load swing angle convergence of the load swing suppression strategy using input shaping is higher than that of the control group without input shaping. Input shaping promotes a slight attenuation of transient effects, and the load swing is reduced by half. The load swing angle of both methods eventually tends to 0. The experimental results verify the effectiveness of the ZVDD shaper used in promoting load swing attenuation.

[0114] The above are merely preferred embodiments of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A method for flight control and load sway suppression of a quadcopter unmanned aerial vehicle (UAV) payload-carrying system, characterized in that, Includes the following steps: Step 1: Establish a dynamic model of the quadrotor UAV's payload-carrying system using the Newton-Euler method and the Euler-Lagrange energy method; Step 2: Employing a cascaded trajectory tracking control strategy, the system is decoupled into an outer-loop position control loop and an inner-loop attitude control loop. For each loop, an interference estimator is designed to address the external lumped disturbances present in both loops. To mitigate the significant interference estimation bias in the outer-loop position interference estimator, an adaptive multi-layer neural network output compensation strategy is designed to reduce the estimation error. The lumped disturbance in the position subsystem of the outer-loop position control loop is: (12)， The tracking error of the position subsystem filter is: , The parameters to be designed are: The tracking error of the position subsystem is: Expected position state , Design robust control laws for the position subsystem. , ; Further results were obtained: (13)， The general form of the position subsystem is expressed as follows: , , The control parameters to be designed; The lumped disturbance estimation error of the location subsystem is: (14)， in, To estimate the robustness term of the position subsystem against lumped disturbances, a neural network is designed consisting of an input layer, hidden layer 1, hidden layer 2, and an output layer. The input layer is selected as... It contains all current state information of the position subsystem; the mapping from the input layer to the first hidden layer is defined as follows: weight matrix The activation function is used to map to the second hidden layer and a bias is added, i.e. The activation function is expressed in the form of: , The output of the neural network is: (15)， Where the weight matrix Using estimated values and To replace ideal weights and ,assumed and There exists a non-negative upper bound. (16)， Accordingly, the hidden layer is determined by the estimated value. , replace and Select the network weight update law (17)， in, , These are the weight update rates for the two hidden layers, respectively. The interference estimator for the three channels representing the location estimates the error. , , For the parameters to be determined, by selecting appropriate parameters so that the output of the neural network effectively compensates for the lumped disturbance estimation error, the control law of the position subsystem is obtained as follows: (18)； Step 3: Using the load trajectory generation technology based on the principle of system differential flatness, the load position and the yaw angle information of the UAV are used as the flat output of the quadcopter UAV load-bearing system. The flat output and its higher-order derivatives are used to represent all states of the system, and a mapping relationship between the load motion trajectory and the UAV motion trajectory is established. Step 4: To address the load swaying issues during system flight and the residual load oscillations during drone hovering, a shaper is designed to suppress the residual oscillations, allowing the drone to drive the load along the desired trajectory.

2. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 1, characterized in that, The dynamic model of the quadcopter UAV load-bearing system established in step one is as follows: ， in, The mass of the quadcopter drone; For the mass of the load; Provides vertical lift for the four motors; This represents the lumped interference vector experienced by the quadcopter drone during flight. This represents the lumped disturbance vector experienced during payload flight. It is the force vector; gravitational vector , It is the acceleration due to gravity; This is the position vector of the UAV in the geodetic coordinate system; This is the position vector of the load in the geodetic coordinate system.

3. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 1 or 2, characterized in that, In step two, the lumped disturbance of the attitude subsystem in the inner loop attitude control loop is expressed as: (4)， in, , , These represent the inertial moments along the three directions of the body coordinate system, respectively. Represents the rotational inertia matrix; Represents the rotational torque along the three directions of the body axis; Represents lumped interference in the pitch channel. Represents the lumped interference of the roll angle channel. Represents lumped disturbance in the yaw angle channel; inner loop attitude state variables , , ; The attitude subsystem filter tracking error is selected as follows: (5)， Among them, the attitude subsystem tracking error , , ; These are non-negative design parameters; Differentiating formula (5) and combining it with formulas (3) and (4), we get (6)， Among them, the lumped interference of the attitude subsystem ; ; ; Design attitude robust control law Approximating the lumped disturbance of the attitude loop, substituting it into formula (6) yields: (7)， Design using feedback linearization ,in Since it is a non-negative control parameter, substituting it into (7) yields... (8)， As can be seen from formula (8), the lumped disturbances in the attitude subsystem can be estimated by the current state of the system and the expected input.

4. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 3, characterized in that, The estimation of the lumped disturbance of the attitude subsystem is expressed as: (9)， in, For a first-order low-pass filter, a robust control law This can be expressed as: (10)， in, Using the inverse Laplace transform operator, and combining formulas (6) to (10), the output of the attitude loop control law can be obtained as follows: ， in, It is a first-order low-pass filter.

5. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 3, characterized in that, Step three utilizes flat output to achieve effective dimensionality reduction of the system, including the instantaneous transition from one dynamic to another, so that the system can still maintain control over state changes in all degrees of freedom by controlling the rotational speeds of the four motors of the quadcopter drone. (19)， (20)， in, The length of the rope. Let the unit vector of the payload pointing to the UAV be further obtained. (21)， (22)， in, , Let the load swing angle be the load position when the load is in the swing angle. Given the given information, from formula (20) then It can be determined; combined with formula (21), it can be determined. Combining formula (19), the location of the UAV is... It can be determined that the mapping from the load motion state to the desired UAV motion state can be found by utilizing the system's differential flatness property.

6. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 1, characterized in that, The mathematical expression for the input shaper in step four is: (22)， in, Represents the system response. Representing the The pulse at time t, where s represents the Laplace transform of the pulse at the corresponding time.

7. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 6, characterized in that, The quadcopter UAV payload system is simplified into a second-order system, and its typical transfer function is obtained as follows: (23)， in, The system's inherent natural frequency, If the system is damped, then the system's unit impulse response is: (24)， in, The damped oscillation frequency of the system. The shaper contains a series of pulses with different amplitudes and time delays. The amplitude and time delay of the pulse at time step are respectively and The transient response caused by the pulse is (25)， in, Representing the current moment, Let $\mathbf$ represent the time intervals of each impulse response. The total system response is the sum of the responses of each component; the total system response is the sum of the responses of each component, expressed as: (26)， Percentage of residual oscillations Defined as (27)， in The purpose of the shaper is to divide the original signal into several segmented signals to excite the system, so that... =0, that is, it satisfies (29)， By making the percentage of residual vibration equal to zero, it is guaranteed that This constraint is met, and the position of the first fundamental pulse is set to... This determines the amplitude of the pulse and the excitation time of the second pulse. .

8. The flight control and load sway suppression method for a quadcopter UAV payload-carrying system according to claim 7, characterized in that, In order to eliminate the residual vibration of the system, In the time domain, it should be zero. Therefore, the selected impulse response amplitude and excitation time are expressed as follows: (30)， in .

Images