Full-state limited control method for quad-rotor unmanned aerial vehicle under uncertainty of kinetic parameters and time-varying load

Abstract

The invention relates to a full-state limited control method for a quad-rotor unmanned aerial vehicle under the conditions of time-varying load, uncertain external disturbance and unknown kinetic parameters, and the method comprises the steps: estimating the unknown kinetic parameters, the unknown time-varying load and the external disturbance of a system through a self-adaptive method; a logarithmic obstacle function is utilized to limit the full state of the quad-rotor unmanned aerial vehicle, a parameter exponential attenuation item is introduced to quickly and accurately converge a final trajectory error, a dynamic surface technology is utilized to remarkably optimize a back-stepping recursion design process, and a complex step of deriving virtual control input is skillfully avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle control, and in particular to a full-state limited control method for a quad-rotor unmanned aerial vehicle under uncertain dynamic parameters and time-varying loads. Background Art

[0002] In recent years, quadrotor drones have attracted considerable attention due to their small size and easy maintenance, as well as their significant advantages such as vertical lift, transportation, strong maneuverability, and easy operation. They are used in forest inspection, agricultural automation, transportation, and logistics. Quadrotor drones have the characteristics of nonlinearity, strong coupling, and underactuation, which makes the control technology of quadrotor drones receive extensive attention and research. However, considering the safety and efficiency of quadrotor drone systems, there are still major challenges in the adaptive trajectory tracking control of quadrotor drones.

[0003] In practical applications, especially when used in outdoor environments, complex and changeable wind speed, wind direction and precipitation can have a significant impact on the flight of the quadrotor drone, causing it to deviate from the preset trajectory. This requires consideration of time-varying loads, dynamic parameters and external disturbance uncertainties.

[0004] At present, most studies only consider the problem of unknown external disturbances. Therefore, it is more practical to consider the situation of time-varying loads and unknown dynamic parameters at the same time. Summary of the invention

[0005] The purpose of the present invention is to provide a full-state constrained control method for a quadrotor unmanned aerial vehicle under uncertain dynamic parameters and time-varying loads, which improves transient performance, keeps the tracking error within the trajectory tracking constraint boundary, and improves the trajectory tracking error convergence speed.

[0006] The present invention adopts the following technical solution:

[0007] A full-state constrained control method for a quadrotor UAV under uncertain dynamic parameters and time-varying loads is proposed. Based on filters, backstepping control and obstacle Lyapunov functions, an adaptive method is adopted and a parameter exponential decay term is introduced to control the position subsystem and attitude subsystem.

[0008] Further, the position subsystem is shown as follows:

[0009]

[0010] The attitude subsystem is shown as follows:

[0011]

[0012] Among them, [x, y, z] represents the position state, and [φ, θ, ψ] represents the attitude state. x =1 / m x , τ y =1 / m y , τ z =1 / m z , τ φ =1 / I x , τ θ =1 / I y , τ ψ =1 / I z ;u x =(cosφsinθcosψ+sinφsinψ)·u1,u y =(cosφsinθsinψ-sinφcosψ)·u1,u z =cosφcosθ·u1;u1,u φ ,u θ ,u ψ Represent the control input of the quadrotor drone, m x ,m y ,m z represents the mass of the quadcopter and they are equal to each other, d i , i = x, y, z, φ, θ, ψ represent the external disturbance of the body, g represents the acceleration of gravity, σ represents the length of the wing, I x ,I y ,I z represents the moment of inertia around the x-axis, y-axis, and z-axis, G i , i=1,3,5,7,9,11 are the air resistance coefficients.

[0013] Furthermore, the tracking error variable of the x-position subsystem is defined as follows:

[0014] Z x1 (t) = x1-X 1d ,

[0015] Among them, X 1d is the reference signal given, is the derivative of x1, α1 is the virtual control input, It is an ideal virtual control input.

[0016] x-position subsystem adaptive controller: u x and

[0017]

[0018] Among them, the design parameter ε1=e -tand c x2 >0; is an estimate of the unknown parameters of D1, yes Estimation of unknown parameters, is the derivative of α1;

[0019] The adaptive law is:

[0020]

[0021] Among them, r x1 >0, r x2 >0 is a known constant; δ x1 =e -t and δ x2 =e -t is a term that decays over time.

[0022] Furthermore, the tracking error variable of the y-position subsystem is defined as follows:

[0023] Z y1 (t) = x3-X 3d ,

[0024] Among them, X 3d is the reference signal given, is the derivative of x3, α3 is the virtual control input, It is an ideal virtual control input.

[0025] y position subsystem adaptive controller: u y and

[0026]

[0027] Among them, the design parameter ε3 = e -t , c y2 >0, is an estimate of the unknown parameters of D3, yes Estimation of unknown parameters, is the derivative of α3;

[0028] The adaptive law is:

[0029]

[0030] Among them, r y1 >0, r y2 >0 is a known constant; δ y1 =e-t and δ y2 =e -t is a term that decays over time.

[0031] Furthermore, the tracking error variable of the z position subsystem is defined as follows:

[0032] Z z1 (t) = x5 - X 5d ,

[0033] Among them, X 5d is the reference signal given, is the derivative of x5, α5 is the virtual control input, It is an ideal virtual control input.

[0034] z position subsystem adaptive controller: u z and

[0035]

[0036] Among them, the design parameter ε5 = e -t , c z2 >0; is the estimate of the unknown parameter of D5,

[0037] yes Estimation of unknown parameters, is the derivative of α5;

[0038] The adaptive law is:

[0039]

[0040] Among them, r z1 >0, r z2 >0 is a known constant; δ z1 =e -t and δ z2 =e -t is a term that decays over time.

[0041] Furthermore, the tracking error variable of the φ attitude subsystem is defined as follows:

[0042] Z φ1 =x7-X 7d ,

[0043] Among them, X 7d is the reference signal given, is the derivative of x7, α7 is the virtual control input, It is an ideal virtual control input.

[0044] φAttitude subsystem adaptive controller: u φ and

[0045]

[0046] Among them, the design parameter ε7 = e -t and c φ2 >0; yes estimates of unknown parameters,

[0047] yes Estimation of unknown parameters, is the derivative of α7;

[0048] The adaptive law is:

[0049]

[0050] Among them, r φ1 >0 and r φ2 >0 is a known constant; d φ1 =e -t , d φ2 =e -t is a term that decays over time.

[0051] Furthermore, the tracking error variable of the θ attitude subsystem is defined as follows:

[0052] Z θ1 =x9-X 9d ,

[0053] Among them, X 9d is the reference signal given, is the derivative of x9, α9 is the virtual control input, It is an ideal virtual control input.

[0054] θ attitude subsystem adaptive controller: u θ and

[0055]

[0056] Among them, the design parameter ε9 = e -t and c θ2 >0; yes estimates of unknown parameters,

[0057] yes Estimation of unknown parameters, is the derivative of α9;

[0058] The adaptive law is:

[0059]

[0060] Among them, r θ1 >0, r θ2 >0 is a known constant; d θ1 =e -t , d θ2 =e -t is a term that decays over time.

[0061] Furthermore, the tracking error variable of the ψ attitude subsystem is defined as follows:

[0062] Z ψ1 =x 11 -X 11d ,

[0063] Among them, X 11d is the reference signal given, For x 11 The derivative of 11 is the virtual control input, It is an ideal virtual control input.

[0064] ψAttitude subsystem adaptive controller: u ψ and

[0065]

[0066] Among them, the design parameter ε 11 =e -t and c ψ2 >0; yes estimates of unknown parameters, yes Estimation of unknown parameters, is α 11 The derivative of

[0067] The adaptive law is:

[0068]

[0069] Among them, r ψ1 >0, r ψ2 >0 is a known constant; dψ1 =e -t , d ψ2 =e -t is a term that decays over time.

[0070] The beneficial effects of the present invention are:

[0071] (1) The present invention improves transient performance by combining a logarithmic barrier Lyapunov function with a parameter exponential decay term, keeps the tracking error within the trajectory tracking constraint boundary, and improves the trajectory tracking error convergence speed.

[0072] (2) The present invention overcomes the influence of external disturbances, unknown dynamic parameters and time-varying loads on quadrotor drones, and successfully solves the limitations caused by relying on known system uncertainty boundaries.

[0073] (3) The present invention introduces dynamic surface technology, significantly optimizes the backstepping recursive design process, cleverly avoids the complex operation of directly deriving the virtual control input in the design, and improves the design efficiency and performance of the control system. Finally, the simulation results prove that the control method can better track and control the quadcopter UAV. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a three-dimensional stereogram and projection diagram of the quadrotor tracking trajectory.

[0075] Figure 2 This is a position tracking graph.

[0076] Figure 3 It is the posture tracking curve graph.

[0077] Figure 4 is the position tracking error curve.

[0078] Figure 5 is the posture tracking error curve. DETAILED DESCRIPTION

[0079] The following is combined with Figures 1 to 5 The following examples are only used to illustrate and explain the present invention, and do not constitute a limitation on the technical solution of the present invention.

[0080] 1. Mathematical model and system description of uncertain dynamics

[0081] The unknown dynamic parameters, unknown time-varying loads and external disturbances of the system are estimated by constructing a quadrotor drone model and using an adaptive method. The model adopts a modular approach and is divided into a position subsystem and an attitude subsystem.

[0082]

[0083] The following parameters are defined: [x, y, z] represents the position state, and [φ, θ, ψ] represents the attitude state. x =1 / m x , τ y =1 / m y , τ z =1 / m z , τ φ =1 / I x , τ θ =1 / I y , τ ψ =1 / I z ;u x =(cosφsinθcosψ+sinφsinψ)·u1,u y =(cosφsinθsinψ-sinφcosψ)·u1,u z =cosφcosθ·u1;u1,u φ ,u θ ,u ψ Represent the control input of the quadrotor drone, m x ,m y ,m z represents the mass of the quadcopter and they are equal to each other, d i , i = x, y, z, φ, θ, ψ represent the external disturbance of the body, g represents the acceleration of gravity, σ represents the length of the wing, I x ,I y ,I z represents the moment of inertia around the x-axis, y-axis, and z-axis, G i , i=1,3,5,7,9,11 are the air resistance coefficients.

[0084] Lemma 1. For any |Z η1 |<k η ,inequality

[0085]

[0086] Established.

[0087] Assumption 1: Consider the existence of time-varying payload, so the parameter τ x ,τ y ,τ z ,τ φ ,τ θ and τ mψ The upper and lower bounds of are unknown. Define τ x ,τ y ,τ z ,τ φ ,τθ and τ ψ The lower bound of and is unknown and greater than 0. From the above formula, it is reasonable to assume that the parameter T φ =(I y -I z ) / I x , T θ =(I z -I x ) / I y , T ψ =(I x -I y ) / I z The upper bound is unknown.

[0088] Assumption 2: Considering the existence of disturbances such as unknown airflow, the parameter d x ,d y ,d z ,d φ ,d θ and d ψ The upper and lower bounds of are unknown. Define |τ x d x |,|τ y d y |,|τ z d z |,max{|τ φ d φ |,T φ},max{|τ θ d θ |,T θ} and max{|τ ψ d ψ |,T ψ} are less than or equal to D x ,D y ,D z , and Assuming parameter D x ,D y ,D z , and is unknown.

[0089] Assumption 3: Consider a continuous function ε on [0,∞) i (t) satisfies ε i (t)>0,i=1,3,5,7,9,11 and limit |Z η2 (t)|<k η ,η=x,y,z,φ,θ,ψ, where k η >0 is a given constant.

[0090] 2. Position controller design

[0091] Define the error variable:

[0092] Z x1 (t) = x1(t) - X 1d (t) (2)

[0093] In the formula, X 1d (t) represents the expectation in the x-axis direction. Derivative of equation (2) yields:

[0094]

[0095] The virtual control error is defined as:

[0096]

[0097] Among them, X 1d is the given reference signal, α1 is the virtual control input, It is an ideal virtual control input.

[0098] In order to ensure that the trajectory error is constrained within the pre-designed boundaries, the logarithmic barrier Lyapunov function is selected:

[0099] Derivative it and substitute it into equations (3) and (4), and through Young's inequality we get:

[0100]

[0101] The ideal virtual control input is designed by equation (6): for:

[0102]

[0103] Among them, c x1 >0 as a design constant.

[0104] The ideal virtual control input Substituting into (6), we get:

[0105]

[0106] Then, by setting The first-order filter is:

[0107]

[0108] Among them, λ1>0 is a time constant.

[0109] Through Young's inequality we get:

[0110]

[0111] Where γ1 = e -t ,

[0112] Further, choose the Lyapunov function:

[0113]

[0114] in, is an estimate of the unknown parameters of D1, yes Estimation of unknown parameters, r x1 >0 and r y2 >0 is a known constant. x2 The derivative is:

[0115]

[0116] Design the x-position subsystem controller: u x and

[0117]

[0118] Among them, the design parameter ε1=e -t and c x2 > 0. is an estimate of the unknown parameters of D1, yes Estimation of unknown parameters, is the derivative of α1;

[0119] Substitute u into equation (13) x , so that V x2 The time derivative of can be rewritten as follows:

[0120]

[0121] Bring in have to:

[0122]

[0123] In which, define ε1=e -t and c x2 >0.

[0124] The designed adaptive law is:

[0125]

[0126] Among them, r x1 >0, r x2 >0 is a known constant. x1 =e -t and δ x2 =e -t is a term that decays over time.

[0127] Therefore, V x2 The time derivative of can be rewritten as follows:

[0128]

[0129] Among them, μ x =min{2c x1 ,2c x2 ,δ x1 r x1 ,δ x2 r x2}, δ x1 =e -t , δ x2 =e -t ,γ1=e -t and ε1=e -t .

[0130] From the above equation we can see

[0131] The following relationship holds:

[0132]

[0133] Depend on Available

[0134] From the above formula, we can get:

[0135]

[0136] At this point, the control design and stability verification are completed.

[0137] Similarly, the controller u y 、u z , and They are designed as follows:

[0138]

[0139] Among them, the design parameter ε3 = e -t , ε5=e -t , cy2 >0 and c z2 >0; is an estimate of the unknown parameters of D3, is the estimate of the unknown parameter of D5, yes Estimation of unknown parameters, yes Estimation of unknown parameters, is the derivative of α3, is the derivative of α5, the ideal virtual controller input and for:

[0140]

[0141] Among them, X 3d ,X 5d is the given reference signal, and the design parameter c y1 >0 and c z1 >0.

[0142] The adaptive law is designed as follows:

[0143]

[0144] Among them, r y1 >0, r y2 >0, r z1 >0, r z2 >0 is a known constant. y1 =e -t , δ y2 =e -t , δ z1 =e -t , δ z2 =e -t is a term that decays over time.

[0145] Based on the above analysis, we can get Theorem 1 as follows:

[0146] Theorem 1. For the outer loop position subsystem of a quadrotor drone with unknown dynamic parameters, time-varying loads and unknown external disturbances, adaptive controllers (13) and (19) as well as adaptive laws (16) and (21) are designed based on filters, backstepping control and obstacle Lyapunov functions to ensure that the trajectory tracking error of the position subsystem is stable. Finally, we get ζ=x,y,z。

[0147] 3. Attitude controller design

[0148] For the rotation subsystem, the inner loop attitude subsystem controller is designed through the attitude extraction algorithm. The required roll angle subsystem, pitch angle subsystem and yaw angle subsystem can be derived from the translation subsystem. The attitude extraction algorithm is as follows:

[0149]

[0150] For the inner loop attitude subsystem controller, the controller design includes three steps: (1) Constructing the obstacle Lyapunov function based on backstepping control to ensure that the trajectory error is within the pre-set boundary. (2) Using a filter to estimate the derivative of the ideal virtual input to simplify the calculation. (3) Introducing an exponential function to make the final trajectory error converge to 0. The details are as follows.

[0151] First, for the roll angle subsystem:

[0152]

[0153] The roll angle tracking error is defined as follows:

[0154] Z φ1 =x7-X 7d (twenty three)

[0155] Where X 7d is the given reference signal, based on formula (23), φ1 Taking the derivative, we get:

[0156] α7 is a virtual control input.

[0157] The virtual control error is defined as:

[0158]

[0159] It is an ideal virtual control input.

[0160] As with the position subsystem, choose the logarithmic barrier Lyapunov function:

[0161]

[0162] Its time derivative is:

[0163]

[0164] If you select where c φ1 >0 as the design constant. By substituting Lemma 1, we get:

[0165]

[0166] Then, by setting The first-order filter is:

[0167]

[0168] Among them, λ7>0 is a time constant.

[0169] Through Young's inequality we get:

[0170]

[0171] Where γ7 = e -t ,

[0172] Further select the Lyapunov function:

[0173]

[0174] in, yes estimates of unknown parameters, yes Estimation of unknown parameters, r φ1 >0 and r φ2 >0 is a known constant.

[0175] V φ2 The time derivative of is:

[0176]

[0177] Continuous Adaptive Controller

[0178]

[0179] Therefore, V φ2 The time derivative of can be rewritten as follows:

[0180]

[0181] Among them, the design parameter ε7 = e -t and c φ2 >0.

[0182] Adaptive Law

[0183]

[0184] where r φ1 >0 and r φ2 >0 is a known constant, d φ1 =e -t , d φ2=e -t is a term that decays over time.

[0185] V φ2 The time derivative of can be rewritten as follows:

[0186]

[0187] in, For simplicity, choose d φ1 =e -t , d φ2 =e -t ,γ7=e -t and ε7=e -t . Not difficult to obtain

[0188] The following relationship holds:

[0189]

[0190] From the above formula, it is not difficult to obtain

[0191]

[0192] At this point, the control design and stability verification are completed.

[0193] Similarly, the controller u θ 、u ψ , and They are designed as follows:

[0194]

[0195] Among them, the design parameter ε9 = e -t , ε 11 =e -t , c θ2 >0 and c ψ2 >0; yes estimates of unknown parameters; yes estimates of unknown parameters; yes Estimation of unknown parameters, is the derivative of α9, is α 11 The derivative of

[0196] Ideal virtual controller input and for

[0197]

[0198] Among them, X 9d ,X 11d is the given reference signal, and c θ2 >0 and c ψ2 >0 as a design constant.

[0199] The adaptive law is designed as:

[0200]

[0201] Among them, r θ1 >0, r θ2 >0, r ψ1 >0, r ψ2 >0 is a known constant. θ1 =e -t , d θ2 =e -t , d ψ1 =e -t , d ψ2 =e -t is a term that decays over time.

[0202] Based on the above analysis, the designed controller can obtain Theorem 2 as follows:

[0203] Theorem 2: For the inner loop attitude subsystem of a quadrotor drone with unknown dynamic parameters, time-varying loads and unknown external disturbances, the adaptive trajectory tracking controllers (33) and (38) and adaptive laws (35) and (40) are designed according to Theorem 1 to ensure that the trajectory tracking error of the attitude subsystem is stable, and finally we get χ=φ,θ,ψ.

[0204] 4. Simulation results

[0205] In order to more accurately prove the effectiveness of the above adaptive control method, numerical simulation verification is carried out. The parameters of the UAV system are selected as follows:

[0206] Initial quadcopter position and attitude settings: [0.01, 0.001, 0.001] T and [0.01,0.1,0.01] T .

[0207] UAV system parameter design: p1 = 0.1sin(10t), τ mx =τ my =τ mz =0.67+p1(kg),τ mφ =τ mθ =33+5p1,τ mψ =25+5p1(kg.m2 ).

[0208] Gravitational acceleration: g = 9.8 m / s 2 .

[0209] The air flow disturbance is designed as:

[0210] The reference signal is as follows: [x d ,y d ,z d ,ψ d ] T =[0.5sin(t),0.5cos(t),0.1t,0] T .

[0211] Controller parameters: k x =k y =k z =0.08sin(5t)(kg / s), k φ =k θ =k ψ =0.08sin(5t)(kg / rad), σ=0.2, c1=20, c3=1, c5=9, c7=10, c9=12, c 11 =15, c2=20, c4=0.01, c6=9, c8=6, c 10 =10, c 12 =10. r1=0.1, r2=0.1, r3=0.3, r4=0.2, r5=0.1, r6=0.2, r7=1, r8=2, r9=0.5, r 10 =1, r 11 =1, r 12 =0.7. G1=0.1,G3=0.1,G5=0.1,G7=12,G9=12,G 11 =12.

[0212] The simulation results of this paper are as follows Figures 1 to 5 As shown. Figure 1 It is a three-dimensional stereogram and three-dimensional projection diagram of the quadrotor tracking trajectory, and the drone quickly tracks the desired curve. Figures 2-3 They are the tracking curve graphs of position variables and posture variables respectively. It can be clearly concluded that the obstacle function converges faster than the one without the obstacle function. Figure 4 It is the position tracking error curve. It can be seen from the figure that Z x , Z y and Z zStabilization is achieved quickly, and the tracking error remains within the pre-set boundaries (shown by the upper and lower boundaries in the figure), while without considering the barrier function, it can be seen that the tracking error does not remain within the pre-set boundaries and takes longer to converge. Figure 5 This is the posture tracking error curve. It can be seen in the figure Z θ and Z ψ Stabilization is achieved quickly, and the tracking error remains within the pre-set boundaries (shown by the upper and lower boundaries in the figure), while the case without considering the barrier function also exceeds the pre-set boundaries and takes longer to converge.

[0213] It can be seen from the simulation results that in the adaptive barrier Lyapunov progressive tracking control scheme, unknown parameters and external disturbances can obtain good tracking performance under uncertainties and have good robustness to disturbances.

Claims

1. A full-state constrained control method for a quadrotor drone under uncertain dynamic parameters and time-varying loads, characterized in that: Based on filters, backstepping control and obstacle Lyapunov function, an adaptive method is adopted and parameter exponential decay terms are introduced to control the position subsystem and attitude subsystem.

2. According to claim 1, a method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads, characterized in that: The position subsystem is shown as follows: The attitude subsystem is shown as follows: Among them, [x, y, z] represents the position state, and [φ, θ, ψ] represents the attitude state. x =1 / m x , τ y =1 / m y , τ z =1 / m z , τ φ =1 / I x , τ θ =1 / I y , τ ψ =1 / I z ;u x =(cosφsinθcosψ+sinφsinψ)·u1,u y =(cosφsinθsinψ-sinφcosψ)·u1,u z =cosφcosθ·u1;u1,u φ ,u θ ,u ψ Represent the control input of the quadrotor drone, m x ,m y ,m z represents the mass of the quadcopter and they are equal to each other, d i , i = x, y, z, φ, θ, ψ represent the external disturbance of the body, g represents the acceleration of gravity, σ represents the length of the wing, I x ,I y ,I z represents the moment of inertia around the x-axis, y-axis, and z-axis, G i , i=1,3,5,7,9,11 are the air resistance coefficients.

3. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 2, characterized in that: Define the tracking error variable as follows: Among them, X 1d is the reference signal given, is the derivative of x1, α1 is the virtual control input, It is an ideal virtual control input; x-position subsystem adaptive controller: u x and Among them, the design parameter ε1=e -t and c x2 >0; is an estimate of the unknown parameters of D1, yes Estimation of unknown parameters, is the derivative of α1; The adaptive law is: Among them, r x1 >0, r x2 >0 is a known constant; δ x1 =e -t and δ x2 =e -t is a term that decays over time.

4. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 3, characterized in that: Define the tracking error variable as follows: Among them, X 3d is the reference signal given, is the derivative of x3, α3 is the virtual control input, It is an ideal virtual control input; y position subsystem adaptive controller: u y and Among them, the design parameter ε3 = e -t , c y2 >0, is an estimate of the unknown parameters of D3, yes Estimation of unknown parameters, is the derivative of α3; The adaptive law is: Among them, r y1 >0, r y2 >0 is a known constant; δ y1 =e -t and δ y2 =e -t is a term that decays over time.

5. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 4, characterized in that: Define the tracking error variable as follows: Among them, X 5d is the reference signal given, is the derivative of x5, α5 is the virtual control input, It is an ideal virtual control input; z position subsystem adaptive controller: u z and Among them, the design parameter ε5 = e -t , c z2 >0; is the estimate of the unknown parameter of D5, yes Estimation of unknown parameters, is the derivative of α5; The adaptive law is: Among them, r z1 >0, r z2 >0 is a known constant; δ z1 =e -t and δ z2 =e -t is a term that decays over time.

6. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 5, characterized in that: Define the tracking error variable as follows: Among them, X 7d is the reference signal given, is the derivative of x7, α7 is the virtual control input, It is an ideal virtual control input; φAttitude subsystem adaptive controller: u φ and Among them, the design parameter ε7 = e -t and c φ2 >0; yes estimates of unknown parameters, yes Estimation of unknown parameters, is the derivative of α7; The adaptive law is: Among them, r φ1 >0 and r φ2 >0 is a known constant; d φ1 =e -t , d φ2 =e -t is a term that decays over time.

7. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 6, characterized in that: Define the tracking error variable as follows: Among them, X 9d is the reference signal given, is the derivative of x9, α9 is the virtual control input, It is an ideal virtual control input; θ attitude subsystem adaptive controller: u θ and Among them, the design parameter ε9 = e -t and c θ2 >0; yes estimates of unknown parameters, yes Estimation of unknown parameters, is the derivative of α9; The adaptive law is: Among them, r θ1 >0, r θ2 >0 is a known constant; d θ1 =e -t , d θ2 =e -t is a term that decays over time.

8. A method for full-state constrained control of a quadrotor drone under uncertain dynamic parameters and time-varying loads according to claim 7, characterized in that: Define the tracking error variable as follows: Among them, X 11d is the reference signal given, For x 11 The derivative of 11 is the virtual control input, It is an ideal virtual control input; ψAttitude subsystem adaptive controller: u ψ and Among them, the design parameter ε 11 =e -t and c ψ2 >0; yes estimates of unknown parameters, yes Estimation of unknown parameters, is α 11 The derivative of The adaptive law is: Among them, r ψ1 >0, r ψ2 >0 is a known constant; d ψ1 =e -t , d ψ2 =e -t is a term that decays over time.

Images