Multi-unmanned aerial vehicle distributed fault-tolerant control method, and multi-unmanned aerial vehicle control method and system

Abstract

The invention provides a multi-unmanned aerial vehicle distributed fault-tolerant control method, an unmanned aerial vehicle control method and an unmanned aerial vehicle control system, and belongs to the technical field of unmanned aerial vehicles. A position ring distributed fault-tolerant controller and an attitude ring distributed fault-tolerant controller are constructed in combination with a multi-unmanned-aerial-vehicle position ring virtual control law and a multi-unmanned-aerial-vehicle attitude ring virtual control law, and fault-tolerant processing is performed on the flight path of the fault unmanned aerial vehicle through the position ring distributed fault-tolerant controller; the fault unmanned aerial vehicle can keep tracking an expected trajectory; fault-tolerant processing is carried out on the attitude of the faulted unmanned aerial vehicle through the attitude ring distributed fault-tolerant controller, so that the faulted unmanned aerial vehicle can maintain a stable attitude angle, thereby ensuring that the whole unmanned aerial vehicle cluster system can operate safely and stably.

Description

Technical Field

[0001] The present invention belongs to the field of UAV technology and relates to the control technology of multiple UAVs, specifically a multi-UAV distributed fault-tolerant control method, a UAV control method and a system. Background Art

[0002] Drones, especially quadcopters, are widely used in both military and civilian applications due to their simple structure, ease of operation, and maneuverability. Multiple drones are increasingly used to coordinate missions. Compared to a single drone, multiple drones can cover a wider range of flight areas and perform more complex missions through collaboration. Therefore, safe formation control of multi-drone systems has become a hot topic of research.

[0003] When performing missions, multiple drones are susceptible to actuator failures, such as motor failure and blade damage, which can easily lead to instability or even collapse of the multi-drone formation. Currently, multi-drone formation control methods are mainly divided into centralized control and distributed control. Centralized control relies on a central controller to schedule and manage individual drones. Although its control structure is relatively simple, a single point of failure in the central controller can paralyze the entire centralized control system. Due to the shortcomings of centralized control, distributed control is becoming increasingly popular for multi-drone formation control.

[0004] However, existing distributed control cannot perform efficient fault tolerance under local communication and finite state information. Once a faulty drone appears in a multi-drone formation and is not dealt with in a timely manner, the faulty drone will be at risk of losing control, unable to maintain a stable attitude angle and unable to continue tracking the desired trajectory. It will collide with surrounding non-faulty drones, causing the entire multi-drone formation to lose control, affecting the safe and stable operation of the entire drone cluster system. Summary of the Invention

[0005] As described in the above background technology, the distributed control of multi-UAV formations used in the existing technology cannot perform efficient fault tolerance under local communication and finite state information, resulting in the technical problem that once a faulty UAV appears in the multi-UAV formation, the faulty UAV cannot maintain a stable attitude angle and cannot continue to track the desired trajectory, affecting the safe and stable operation of the entire UAV cluster system. To address this technical problem, the present invention proposes a multi-UAV distributed fault-tolerant control method, a UAV control method and a system.

[0006] The distributed fault-tolerant control method for multiple UAVs of the present invention performs fault-tolerant processing on the flight trajectory of a faulty UAV through a position loop distributed fault-tolerant controller, so that the faulty UAV can keep tracking the desired trajectory; and performs fault-tolerant processing on the attitude of the faulty UAV through an attitude loop distributed fault-tolerant controller, so that the faulty UAV can maintain a stable attitude angle, thereby ensuring that the entire UAV cluster system can operate safely and stably.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] The present invention provides a multi-UAV distributed fault-tolerant control method, comprising the following steps:

[0009] S1: Construct virtual control laws for multi-UAV position loops and multi-UAV attitude loops;

[0010] S2: Based on Nussbaum function, a distributed fault-tolerant controller for the position loop is constructed according to the virtual control law of the multi-UAV position loop;

[0011] Based on Nussbaum function, a distributed fault-tolerant controller for attitude loop is constructed according to the virtual control law of multi-unmanned attitude loop.

[0012] S3: Obtain the expected position parameters and expected attitude parameters of the faulty UAV in the multi-UAV formation;

[0013] S4: Substitute the expected position parameters of the faulty UAV into the position loop distributed fault-tolerant controller to perform fault-tolerant processing on the flight trajectory of the faulty UAV; substitute the expected attitude parameters of the faulty UAV into the attitude loop distributed fault-tolerant controller to perform fault-tolerant processing on the attitude of the faulty UAV; by performing fault-tolerant processing on the flight trajectory of the faulty UAV and the fault-tolerant processing on the attitude of the faulty UAV, distributed fault-tolerant control of multiple UAVs is performed.

[0014] It is further defined that the position loop distributed fault-tolerant controller is:

[0015]

[0016] Where, Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i , O iy =cosΦ i sinθ i sinΨ i-sinΦ i cosΨ i , O iz =cosΦ i sinθ i , Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T I ,-T I ,-T i ] T , unit: kg·m / s 2 ;T i is the total lift of the i-th faulty UAV, unit: N; Y i (ξ i )=diag{Y i (ξ I,1 ),Y i (ξ I,2 ),Y i (ξ I,3 )},ξ I,1 is the first element of the Nussbaum term of the position loop of the i-th faulty UAV; I,2 is the second element of the Nussbaum term of the position loop of the i-th faulty UAV; I,3 is the third element of the Nussbaum term of the position loop of the i-th faulty UAV; Y i (ξ I,1 ) is the first element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ I,2 ) is the second element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ I,3 ) is the third element of the Nussbaum gain of the position loop of the i-th faulty UAV; Unit: m 2 / s 3 ; is the derivative of ξi,1; is the derivative of ξi,2; is the derivative of ξi,3; is x i2d The derivative of, unit: m / s 2 ;x i2 =D i , unit: m / s; D iis the velocity vector of the i-th faulty UAV in the ground coordinate system, unit: m / s; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; X1=ω i1 +ω i2 L ii , unit: (s 2 ) -1 ;ω i1 is the weight coefficient of the position loop error of the i-th faulty UAV; ω i2 is the weight coefficient of the position loop error with the neighboring aircraft of the i-th faulty UAV; L ii is the element of the Laplace matrix of the i-th faulty UAV; Unit: m; e i1 is the position error of the i-th faulty UAV, unit: m / s; j is the serial number of the UAV in the UAV formation that can communicate with the i-th faulty UAV, i≠j; e j1 is the position error of the jth UAV, unit: m / s; L ij To divide L ii Elements in the Laplacian matrix other than N i is the set of neighbors of the i-th faulty UAV; K 12 is the positive definite symmetric matrix to be designed in the position loop distributed fault-tolerant controller, unit: s -1 ;e i2 is the speed tracking error of the i-th faulty UAV, unit: m / s; u i is the position loop controller of the i-th faulty UAV, unit: m / s 2 ;

[0017] The attitude ring distributed fault-tolerant controller is:

[0018]

[0019] Where, J i is the moment of inertia matrix of the i-th faulty UAV, unit: m / s 2 ;M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 ; is x i4d The derivative of , in rad / s 2 ;x i4 =Ω i , unit: rad / s; Ω i Angular velocity of the i-th faulty UAV, unit: rad / s; Unit: rad / s2 ;Ω i is the angular velocity of the i-th faulty UAV; e i4 is the angular velocity error of the i-th faulty UAV, unit: rad / s 2 ;K 22 is the positive definite symmetric matrix of the attitude loop distributed fault-tolerant controller to be designed, unit: s -1 ;H i is the attitude kinematic matrix of the i-th faulty UAV; X2=ω i3 +ω i4 L ii , unit: (s 2 ) -1 ;ω i3 is the weight coefficient of the attitude loop error of the i-th faulty UAV; ω i4 is the weight coefficient of the attitude loop error with the neighboring aircraft of the i-th faulty UAV; Unit: rad; e i3 is the attitude angle error of the i-th faulty UAV, unit: rad; e j3 is the attitude angle error of the j-th UAV, unit: rad; Y i (ζ i )=diag{Y i (ζ i,1 ),Y i (ζ i,2 ),Y i (ζ i,3 )},ζ i,1 is the first element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,2 is the second element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,3 is the third element of the Nussbaum term of the attitude loop of the i-th faulty UAV; Y i (ζ i,1 ) is the first element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,2 ) is the second element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,3 ) is the third element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Unit: rad 2 / s 3 ; For i,1 The derivative of For i,2 The derivative of For i,3 The derivative of uia is the attitude loop controller of the i-th faulty UAV, unit: rad / s 2 .

[0020] It is further defined that the expected position parameters of the faulty drone include the expected roll angle of the faulty drone; and the expected attitude parameters of the faulty drone include the pitch angle of the faulty drone and the total lift of the faulty drone.

[0021] The present invention provides a multi-UAV control method, comprising:

[0022] S1: Construct an auxiliary observer, form a dynamic representation of the residual based on the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation based on the dynamic representation of the residual;

[0023] S2: Set the detection threshold χ of the UAV actuator failure and determine whether the fault information of the faulty UAV is greater than the detection threshold χ;

[0024] If it is greater than, the above-mentioned multi-UAV distributed fault-tolerant control method is used to perform fault-tolerant processing on the flight trajectory and posture of the faulty UAV.

[0025] It is further defined that the auxiliary observer is:

[0026]

[0027] Where, is x i2 Derivative of the estimated value, unit: m / s 2 ;x i2 =D i , unit: m / s; D i is the velocity vector of the i-th faulty UAV in the ground coordinate system, Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θi is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ;T i is the total lift of the i-th faulty UAV, unit: N; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; is x i2 The estimated value of is in m / s; Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop, in s -1 ; is x i4 Derivative of the estimated value, in rad / s 2 ; Unit: rad / s 2 ;Ω i is the angular velocity of the i-th faulty UAV; J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ;M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 ;x i4 =Ω i , is x i4 The estimated value of is in rad / s; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-unmanned attitude loop, in s -1 .

[0028] It is further defined that the system residual of the auxiliary observer forms a dynamic representation of the residual for:

[0029]

[0030] Where A is the system matrix of the auxiliary observer; r i is the system residual of the auxiliary observer; k iis the fault factor matrix of the i-th faulty UAV; Q = diag{Q1,Q2}, Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV attitude loop; I6 is the identity matrix with dimension 6; U = [U i ,M i ] T , U i =[-T i ,-T i ,-T i ] T , T i is the total lift of the i-th faulty UAV, unit: N; M i is the three-axis torque of the i-th faulty UAV.

[0031] Further limiting,

[0032] The detection threshold χ is:

[0033]

[0034] Where, χ i is the detection threshold of the i-th faulty UAV; P is a positive definite matrix; λ b (P) is the maximum eigenvalue of the positive definite matrix; λ s (P) is the minimum eigenvalue of the positive definite matrix; t0 is the initial time; t is time; γ=PQ+Q T P-PAA T P, Q = diag{Q1, Q2}, Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV attitude loop; A is the system matrix of the auxiliary observer; λ s (γ) is the minimum eigenvalue of γ; x i (t0) is the initial value of the state of the i-th faulty UAV; is the estimated value of the initial state of the i-th faulty UAV.

[0035] The present invention provides a multi-UAV control system, comprising:

[0036] Fault detection module: used to build an auxiliary observer, form a dynamic representation of the residual based on the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation based on the dynamic representation of the residual;

[0037] And a fault-tolerant processing module: used to set a detection threshold χ for the failure of the UAV's actuator, and determine whether the fault information of the faulty UAV is greater than the detection threshold χ; if it is greater, the flight trajectory and posture of the faulty UAV are fault-tolerantly processed using the above-mentioned multi-UAV distributed fault-tolerant control method.

[0038] The present invention provides a memory storing a program file, wherein the program file is executed to implement program instructions formed by the above-mentioned multi-UAV control method.

[0039] The present invention provides an electronic device, comprising a processor and a memory coupled to each other, wherein:

[0040] The memory is used to store program instructions formed by the above-mentioned multi-UAV control method;

[0041] The processor is configured to execute program instructions stored in the memory.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The distributed fault-tolerant control method for multiple UAVs of the present invention introduces the Nussbaum function and respectively combines the virtual control law of the multi-UAV position loop and the virtual control law of the multi-UAV attitude loop to construct a position loop distributed fault-tolerant controller and an attitude loop distributed fault-tolerant controller. The position loop distributed fault-tolerant controller performs fault-tolerant processing on the flight trajectory of a faulty UAV so that the faulty UAV can keep tracking the desired trajectory; the attitude loop distributed fault-tolerant controller performs fault-tolerant processing on the attitude of the faulty UAV so that the faulty UAV can maintain a stable attitude angle, thereby ensuring the safe and stable operation of the entire UAV cluster system.

[0044] 2. The multi-UAV control method of the present invention enhances the multi-UAV formation's fault warning capabilities by constructing an auxiliary state observer and detecting faulty UAVs within the formation based on the observer's residuals. This method accurately detects anomalies in the actuators of faulty UAVs, improving the system's fault detection sensitivity and robustness. Combined with the aforementioned multi-UAV distributed fault-tolerant control method, this method ensures the stability of the multi-UAV formation's structure and enables the continuous completion of collaborative tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Schematic diagram of the distributed fault-tolerant control method for UAVs of the present invention;

[0046] Figure 2 Schematic diagram of the drone control method of the present invention;

[0047] Figure 3 Schematic diagram of the UAV control system of the present invention;

[0048] Figure 4 ω i1 、ω i2 、ω i3 and ω i4 An undirected graph with all values ​​1;

[0049] Figure 5 This is a schematic diagram of the fault detection results of a 4-UAV formation;

[0050] Figure 6 Schematic diagram of the three-dimensional desired trajectory tracking of a formation of four UAVs;

[0051] Figure 7 This is a top view of the three-dimensional desired trajectory tracking diagram of the four UAV formation;

[0052] Figure 8 This is a schematic diagram of the state error of a 4-UAV formation;

[0053] Figure 9 This is the controller curve change diagram of a 4-UAV formation;

[0054] Figure 10 This is a comparison diagram of the tracking errors of the present invention, the PID controller and the adaptive fault observer in the x, y and z directions. DETAILED DESCRIPTION

[0055] The technical solution of the present invention will be further explained below with reference to the accompanying drawings and embodiments, but the present invention is not limited to the embodiments described below.

[0056] The design idea of ​​the present invention is to establish a nonlinear model of a multi-UAV system including actuator failures, construct an auxiliary observer to identify the fault information of faulty UAVs in a multi-UAV formation, and construct a position loop distributed fault-tolerant controller and an attitude loop distributed fault-tolerant controller based on the Nussbaum function and combined with the multi-UAV position loop virtual control law and the multi-UAV attitude loop virtual control law respectively. Then, using the Lyapunov stability theory and consistency theory, it is proved that all error signals of the multi-UAV control method of the present invention are consistent and ultimately bounded.

[0057] According to the Newton-Euler equation, a nonlinear model of a multi-UAV system with actuator failure is established. The nonlinear model of the i-th faulty UAV in a multi-UAV formation can be expressed as:

[0058]

[0059] Where x i2 =D i , D i is the velocity vector of the i-th faulty UAV in the ground coordinate system, unit: m / s; is x i1 Derivative of, m / s; x i1 =S i , unit: m; S i is the position vector of the i-th faulty UAV in the ground coordinate system, unit: m; is x i2 Derivative, unit: m / s 2 ;k i is the failure factor matrix of the i-th faulty UAV; Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ;T i is the total lift of the i-th faulty UAV, unit: kg·m / s 2 ; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; is x i3 The derivative of x, in rad / s; i3 =Θ i , unit: rad; Θ i is the Euler angle of the i-th faulty UAV, unit: rad; H i is the attitude kinematic matrix of the i-th faulty UAV; x i4 =Ω i , unit: rad / s; is xi4 Estimated value of rad / s; Ω i is the angular velocity of the i-th faulty UAV; is x i4 The derivative of , in rad / s 2 ; Unit: rad / s 2 ; J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ;k ia is the fault factor matrix of the attitude loop of the i-th faulty UAV; M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 .

[0060] See also Figure 2 , the present invention proposes a multi-UAV control method, comprising:

[0061] S1: Construct an auxiliary observer, form a dynamic representation of the residual based on the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation based on the dynamic representation of the residual;

[0062] S2: Set the detection threshold χ of the UAV actuator failure and determine whether the fault information of the faulty UAV is greater than the detection threshold χ;

[0063] If it is greater than, the multi-UAV distributed fault-tolerant control method is used to perform fault-tolerant processing on the flight trajectory and attitude of the faulty UAV.

[0064] See also Figure 1 ,The multi-UAV distributed fault-tolerant control method includes the following steps:

[0065] S1: Construct virtual control laws for multi-UAV position loops and multi-UAV attitude loops;

[0066] S2: Based on Nussbaum function, a distributed fault-tolerant controller for the position loop is constructed according to the virtual control law of the multi-UAV position loop;

[0067] Based on Nussbaum function, a distributed fault-tolerant controller for attitude loop is constructed according to the virtual control law of multi-unmanned attitude loop.

[0068] S3: Obtain the expected position parameters and expected attitude parameters of the faulty UAV in the multi-UAV formation;

[0069] S4: Substitute the expected position parameters of the faulty UAV into the position loop distributed fault-tolerant controller to perform fault-tolerant processing on the flight trajectory of the faulty UAV; substitute the expected attitude parameters of the faulty UAV into the attitude loop distributed fault-tolerant controller to perform fault-tolerant processing on the attitude of the faulty UAV; by performing fault-tolerant processing on the flight trajectory of the faulty UAV and the fault-tolerant processing on the attitude of the faulty UAV, distributed fault-tolerant control of multiple UAVs is performed.

[0070] The auxiliary observer is:

[0071]

[0072] Where, is x i2 Derivative of the estimated value, unit: m / s 2 ;x i2 =D i , D i is the velocity vector of the i-th faulty UAV in the ground coordinate system; Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ;T i is the total lift of the i-th faulty UAV, unit: kg·m / s 2 ; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; is x i2 The estimated value of is in m / s; Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop, in s -1 ; is x i4 Derivative of the estimated value, in rad / s 2 ; Unit: rad / s 2 ;Ω i is the angular velocity of the i-th faulty UAV; J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ;M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 ;x i4 =Ω i , unit: rad / s; is x i4 The estimated value of is in rad / s; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-unmanned attitude loop, in s -1 .

[0073] The residual r of the auxiliary observer i2 and r i4 Expressed as:

[0074]

[0075] Where r i2 is the residual error of the auxiliary observer of the position loop, unit: m / s; r i4 is the residual of the auxiliary observer of the attitude loop, unit: rad / s.

[0076] The residual r of the auxiliary observer i2 and r i4 Taking the derivative we get and

[0077]

[0078] Where, For r i2 The derivative of, unit: m / s 2 ; For r i4 The derivative of , in rad / s 2 ; I3 is the identity matrix with dimension 3; Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop, unit: s -1; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-unmanned attitude loop, unit: s -1 .

[0079] Simplify the above and The formula can be used to obtain the dynamic representation of the residual for:

[0080]

[0081] A is the system matrix of the auxiliary observer, r i is the system residual of the auxiliary observer, r i =[r i2 ,r i4 ] T ; k is the failure factor matrix of the i-th faulty UAV, which is any value between (0,1), preferably, its value is 0.5; Q = diag{Q1,Q2}; I6 is the identity matrix with dimension 6; U = [U i ,M i ] T , U i =[-T i ,-T i ,-T i ] T , T i is the total lift of the i-th faulty UAV, unit: N; M i is the three-axis torque of the i-th faulty UAV.

[0082] Set the detection threshold χ of the UAV actuator failure, specifically:

[0083] Take the auxiliary observer Lyapunov function V0 as:

[0084] V0=r i T Pr i

[0085] P is a positive definite matrix.

[0086] After derivation, simplification and scaling of the auxiliary observer Lyapunov function V0, we can get

[0087]

[0088] In the formula, γ=PQ+Q T P-PAA T P,λ s (γ) is the smallest eigenvalue of γ; I6 is the identity matrix with dimension 6.

[0089] Based on the Reyleigh-Ritz theorem Converts to:

[0090]

[0091] right After solving the differential inequality, we get:

[0092]

[0093] After conversion and scaling, we get:

[0094]

[0095] Where t0 is the initial time; t is the time; λ b (P) is the maximum eigenvalue of the positive definite matrix; λ s (P) is the minimum eigenvalue of the positive definite matrix; r i (t0) is the residual of the auxiliary observer at the initial moment; τ is any moment between t0 and t; μ is the fault term. When the actuator failure occurs in the i-th UAV in the multi-UAV formation, μ≠0; r i (t0) is the system residual of the auxiliary observer at the initial moment.

[0096] The fault detection algorithm for faulty UAVs in a multi-UAV formation is obtained as follows:

[0097]

[0098] Among them, "0" means that no faulty drone is detected in the multi-UAV formation; 1 means that a faulty drone is detected in the multi-UAV formation.

[0099] The detection threshold χ is:

[0100]

[0101] Where, χ i is the detection threshold of the i-th faulty UAV; P is a positive definite matrix; λ b (P) is the maximum eigenvalue of the positive definite matrix; λ s (P) is the minimum eigenvalue of the positive definite matrix; t0 is the initial time; t is time; γ=PQ+Q T P-PAA T P, Q = diag{Q1, Q2}, Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV attitude loop; A is the system matrix of the auxiliary observer; λ s (γ) is the minimum eigenvalue of γ; x i(t0) is the initial value of the state of the i-th faulty UAV; is the estimated value of the initial state of the i-th faulty UAV.

[0102] The following describes the position loop distributed fault-tolerant controller:

[0103] The position loop equation is established as:

[0104]

[0105] x i2 =D i , unit: m / s; D i is the velocity vector of the i-th faulty UAV in the ground coordinate system, unit: m / s; is x i2 The derivative of, unit: m / s; is x i2 Derivative, unit: m / s 2 ;k i is the failure factor matrix of the i-th faulty UAV; Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ;T i is the total lift of the i-th faulty UAV, unit: kg·m / s 2 ; G a =[0,0,g] T, g is the acceleration due to gravity, unit: m / s 2 .

[0106] Among them, the position loop error of the i-th faulty UAV is:

[0107]

[0108] e i1 is the position error of the i-th faulty UAV, unit: m; e i2 is the speed tracking error of the i-th faulty UAV, unit: m / s; x i1 =S i , unit: m; S i is the position vector of the i-th faulty UAV in the ground coordinate system; x i1d is the expected position vector of the i-th faulty UAV, unit: m; x i2d is the expected velocity vector of the i-th faulty UAV, in m / s.

[0109] In order to achieve distributed fault-tolerant control of a multi-UAV formation, each UAV in the formation must not only track its own desired trajectory, but also consider the relative position error between it and its neighboring UAVs. Therefore, the calculation formula for the position synchronization formation tracking error is defined as:

[0110]

[0111] ω i1 is the weight coefficient of the position loop error of the i-th faulty UAV, which is generally set to 1; ω i2 is the weight coefficient of the position loop error with the neighboring aircraft of the i-th faulty UAV, which is greater than 0 and less than ω i1 Any value of E i1 is the position synchronization formation tracking error of the i-th faulty UAV, unit: m; N i is the set of neighboring drones of the i-th faulty drone; j is the serial number of the drone in the drone formation that can communicate with the i-th faulty drone, i≠j; c ij =-L ij , L ij To divide L ii Elements in the Laplacian matrix other than e j1 is the position error of the j-th UAV, unit: m.

[0112] The calculation formula of position synchronization formation tracking error is expanded and simplified to:

[0113]

[0114] Where X1 = ω i1 +ω i2 Lii , L ii is the element of the Laplace matrix of the i-th faulty UAV.

[0115] The derivative of the calculation formula of the position synchronization formation tracking error after expansion and simplification can be obtained

[0116]

[0117] Where, for e i1 The derivative of, unit: m / s; for e j1 The derivative of, unit: m / s; is x i1d The derivative of , unit: m / s.

[0118] The virtual control law x of the multi-UAV position loop can be obtained i2d for:

[0119]

[0120] Where K 11 is the positive definite symmetric matrix to be designed in the virtual control law of the position loop.

[0121] The virtual control law of the multi-UAV position loop x i2d Substitution The calculation formula can be obtained:

[0122]

[0123] The speed tracking error e of the i-th faulty UAV i2 Taking the derivative we get:

[0124]

[0125] Based on Nussbaum gain, the position loop distributed fault-tolerant controller is constructed according to the virtual control law of the multi-UAV position loop:

[0126]

[0127] Where Y i (ξ i )=diag{Y i (ξ i,1 ),Y i (ξ i,2 ),Y i (ξ i,3 )},ξ i,1 is the first element of the Nussbaum term of the position loop of the i-th faulty UAV; i,2is the second element of the Nussbaum term of the position loop of the i-th faulty UAV; i,3 is the third element of the Nussbaum term of the position loop of the i-th faulty UAV; Y i (ξ i,1 ) is the first element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ i,2 ) is the second element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ i,3 ) is the third element of the Nussbaum gain of the position loop of the i-th faulty UAV; Unit: m 2 / s 3 ; is the derivative of ξi,1; for ξ i,2 The derivative of for ξ i,3 The derivative of K 12 is the positive definite symmetric matrix to be designed in the position loop distributed fault-tolerant controller, unit: s -1 ; is x i2d The derivative of m / s 2 ;u i is the position loop controller of the i-th faulty UAV, m / s 2 .

[0128] Substitute the position loop distributed fault-tolerant controller into The calculation formula can be obtained:

[0129]

[0130] The Lyapunov function V1 of the position loop distributed fault-tolerant controller is selected as:

[0131]

[0132] After derivation, simplification, and scaling of the Lyapunov function V1 of the position loop distributed fault-tolerant controller, we can obtain:

[0133]

[0134] Where, κ1=min{2λ s (K 11 ),2λ s (K 12 )},λ s (K 11 ) is K 11 The minimum eigenvalue of s (K12 ) is K 12 The minimum eigenvalue of Y i (ξ i,n ) is the Nussbaum gain of the position loop of the i-th faulty UAV; is the Nussbaum term of the position loop of the i-th faulty UAV ξ i,n The derivative of .

[0135] In order to obtain the total lift of each UAV in the multi-UAV formation, R i U i =[U ix ,U iy ,U iz ] T , then the expected roll angle Φ of the i-th faulty UAV is id , the expected pitch angle θ of the i-th faulty UAV id And the total lift T of the i-th faulty UAV i It can be obtained by the following formula:

[0136]

[0137] Where, id is the expected yaw angle of the i-th faulty UAV, unit: rad; m is the mass of the i-th faulty UAV, unit: kg; U ix is the first element of the distributed fault-tolerant controller for the position loop of the i-th faulty UAV; U iy is the second element of the distributed fault-tolerant controller of the position loop of the i-th faulty UAV; U iz is the third element of the distributed fault-tolerant controller for the position loop of the i-th faulty UAV.

[0138] The desired position parameters of the faulty drone include the desired roll angle of the faulty drone, while the desired attitude parameters of the faulty drone include the pitch angle of the faulty drone and the total lift of the faulty drone. The desired trajectory and desired attitude are achieved through cascade control of the desired roll angle, pitch angle, and total lift of the faulty drone.

[0139] The attitude loop equation is established as:

[0140]

[0141] Where x i3 =Θ i , unit: rad; Θ i is the Euler angle of the i-th faulty UAV, unit: rad; is x i3 The derivative of H, unit: rad / s; iis the attitude kinematic matrix of the i-th faulty UAV; x i4 =Ω i , unit: rad / s; Ω i Angular velocity of the i-th faulty UAV, unit: rad / s; is x i4 The derivative of , in rad / s 2 ; Unit: rad / s 2 ;k ia is the fault factor matrix of the attitude loop of the i-th faulty UAV; J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ;M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 .

[0142] Among them, the attitude loop error of the i-th faulty UAV is:

[0143]

[0144] Where, e i3 is the attitude angle error of the i-th faulty UAV, unit: rad; e i4 is the angular velocity tracking error of the i-th faulty UAV, unit: rad / s; x i3d is the expected attitude angle of the i-th faulty UAV, unit: rad; x i4d is the expected angular velocity of the i-th faulty UAV, in rad / s.

[0145] In order to achieve distributed fault-tolerant control of multi-UAV formations, each UAV in the multi-UAV formation needs to track not only its own desired attitude angle, but also the relative attitude error between it and its neighboring UAVs. Therefore, the attitude synchronization formation tracking error E is defined as i3 The calculation formula is:

[0146]

[0147] Where, ω i3 is the weight coefficient of the attitude loop error of the i-th faulty UAV, which is generally set to 1; ω i4 is the weight coefficient of the attitude loop error with the neighboring aircraft of the i-th faulty UAV, and its value is greater than 0 and less than ω i3 Any value of N i is the set of neighboring drones of the i-th faulty drone; j is the serial number of the drone in the drone formation that can communicate with the i-th faulty drone, i≠j; C ij is the graph edge (vi ,v j ), which takes a value of 0 or 1. When the i-th faulty UAV is communicating with the j-th UAV, its value is 1, otherwise, its value is 0; v i is the i-th faulty drone; v j is the jth drone.

[0148] The calculation formula of attitude synchronization formation tracking error is expanded and simplified to:

[0149]

[0150] Where X2 = ω i3 +ω i4 L ij , unit: (s 2 ) -1 ;L ij To divide L ii The elements in the Laplace matrix other than ; j is the serial number of the drone in the drone formation that can communicate with the i-th faulty drone, i≠j; e i3 is the attitude angle error of the i-th faulty UAV, unit: rad; e j3 is the attitude angle error of the j-th UAV, unit: rad; N i is the set of neighbors of the i-th faulty UAV.

[0151] The derivative of the calculation formula of the attitude synchronization formation tracking error after expansion and simplification can be obtained

[0152] Where, for e i3 The derivative of , unit: rad / s; for e j3 The derivative of , unit: rad / s; is x i3d The derivative of , unit: rad / s.

[0153] The virtual control law of the multi-unmanned attitude loop can be obtained as:

[0154]

[0155] Where K 21 is the positive definite symmetric matrix to be designed in the virtual control law of the attitude loop.

[0156] The virtual control law of the multi-UAV position loop x i4d Substitution The calculation formula can be obtained:

[0157]

[0158] The angular velocity tracking error e of the i-th faulty UAV i4 Taking the derivative we get:

[0159]

[0160] Based on Nussbaum gain, the attitude loop distributed fault-tolerant controller is constructed according to the virtual control law of the multi-UAV position loop:

[0161]

[0162] Where K 22 The positive definite symmetric matrix of the attitude loop distributed fault-tolerant controller to be designed, unit: s -1 ;u ia Attitude loop controller of the i-th faulty UAV, unit: rad / s 2 ; Y i (ζ i )=diag{Y i (ζ i,1 ),Y i (ζ i,2 ),Y i (ζ i,3 )};ζ i,1 is the first element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,2 is the second element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,3 is the third element of the Nussbaum term of the attitude loop of the i-th faulty UAV; Y i (ζ i,1 ) is the first element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,2 ) is the second element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,3 ) is the third element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Unit: rad 2 / s 3 ; is the derivative of ζi,1; is the derivative of ζi,2; For i,3 The derivative of u ia is the attitude loop controller of the i-th faulty UAV.

[0163] Substitute the attitude loop distributed fault-tolerant controller into The calculation formula can be obtained:

[0164]

[0165] The Lyapunov function V2 of the attitude loop distributed fault-tolerant controller is selected as:

[0166]

[0167] After derivation, simplification, and scaling of the Lyapunov function V2 of the attitude loop distributed fault-tolerant controller, we can obtain:

[0168]

[0169] Where, κ2=min{2λ s (K 21 ),2λ s (K 22 )},λ s (K 21 ) is K 21 The minimum eigenvalue of s (K 22 ) is K 22 The minimum eigenvalue of ; is the Nussbaum term of the position loop of the i-th faulty UAV ζ i,n The derivative of .

[0170] See also Figure 3 The present invention also proposes a multi-UAV control system, including a fault detection module and a fault-tolerant processing module. The fault detection module is used to construct an auxiliary observer, form a dynamic representation of the residual according to the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation according to the dynamic representation of the residual; and the fault-tolerant processing module is used to set a detection threshold χ for the failure of the UAV actuator, and determine whether the fault information of the faulty UAV is greater than the detection threshold χ; if greater, the flight trajectory and posture of the faulty UAV are fault-tolerantly processed using the above-mentioned multi-UAV distributed fault-tolerant control method.

[0171] The multi-UAV control system of the present invention is completely corresponding to the above-mentioned multi-UAV control method. For the contents not specifically disclosed in the fault detection module and the fault-tolerant processing module, refer to the description of the above-mentioned multi-UAV control method.

[0172] The following uses Lyapunov stability theory and consistency theory to prove that all error signals of the multi-UAV control method of the present invention are consistently and ultimately bounded.

[0173] The position synchronization formation tracking error and attitude synchronization formation tracking error are:

[0174]

[0175] The position synchronization formation tracking error and the attitude position synchronization formation tracking error are simplified and combined to obtain the synchronization formation tracking error of the multi-UAV formation:

[0176]

[0177] Where, It refers to the set of synchronous formation tracking errors and attitude synchronous formation tracking errors of all faulty UAVs in a multi-UAV formation; is the weight coefficient matrix of the positions and attitudes of all faulty UAVs in a multi-UAV formation; is the weight matrix of the position and attitude of all neighboring faulty UAVs in the multi-UAV formation; I6 is the identity matrix with dimension 6; are the position and attitude errors of all faulty UAVs in the multi-UAV formation; L is the Laplace matrix.

[0178] After simplifying and scaling the synchronous formation tracking error of the multi-UAV formation, we can get:

[0179]

[0180] In the formula, η is the matrix The minimum singular value of 1 / 3 If it converges to a very small interval including 0, the synchronous formation tracking control target can be achieved, and each UAV can track its own reference signal.

[0181] For a multi-UAV formation with actuator failures, the position loop distributed fault-tolerant controller and attitude loop distributed fault-tolerant controller designed in the present invention can make the closed-loop system globally stable and the tracking error converge gradually.

[0182] Select Lyapunov function V3 as:

[0183] V3=V1+V2

[0184] Derivative and simplification of the Lyapunov function V3 yields:

[0185]

[0186] Where β is the coefficient in front of the Lyapunov function; α is the merged term.

[0187] According to the Nussbaum function lemma, V1, V2 and α are in [0,t f ] is bounded, then, the integral can be obtained:

[0188]

[0189] Where V3(0) is the value of V3 at time 0; t is time; t f is any time value greater than t.

[0190] As time goes by, V3 gradually converges, so all error signals of the closed-loop system are eventually uniformly bounded.

[0191] The present invention also provides a memory storing a program file that is executed to implement the program instructions formed by the multi-drone control method described above. The specific content of the multi-drone control method is described above and will not be repeated here.

[0192] The memory in the present invention may specifically include random access memory (RAM), internal memory, read-only memory (ROM), programmable ROM, erasable programmable ROM, register, hard disk, removable disk, or CD-ROM. It should be noted that those skilled in the art may select the form and type of storage medium based on actual use requirements, and the present invention does not further specify this.

[0193] The present invention also provides an electronic device comprising a processor and a memory coupled to each other, wherein the memory is configured to store program instructions generated by the aforementioned multi-drone control method, and the processor is configured to execute the program instructions stored in the memory. The details of the multi-drone control method are described above and will not be further elaborated herein.

[0194] The electronic device in the present invention includes any electronic device that can execute program instructions, such as a computer, a mobile terminal, a remote control device or a wearable device.

[0195] See also Figure 4 In the following, a simulation test of a multi-UAV control method of the present invention is conducted using a model formed by a 4-UAV formation (the 4 UAVs are UAV v1, UAV v2, UAV v3 and UAV v4). For specific simulation test results, please refer to the following description.

[0196] Drone v4 is designed as a fault-free drone, and drone v1, drone v2, and drone v3 are all designed as faulty drones. The faults of drone v1, drone v2, and drone v3 are:

[0197]

[0198]

[0199] Where k1 is the failure factor matrix of the first faulty UAV (UAV v1); k2 is the failure factor matrix of the second faulty UAV (UAV v2); and k3 is the failure factor matrix of the third faulty UAV (UAV v3).

[0200] As reference Figure 5 As shown, the horizontal axis represents the simulation time, and the vertical axis represents the residual norm of the auxiliary observer corresponding to the UAV || r i || and detection threshold χ i It is clearly observed that at 10s, 30s, and 40s, the vertical axes of UAV v1, UAV v2, and UAV v3 are greater than 0, which is the residual norm of the auxiliary observer of the corresponding UAV at this time || r i || Exceeds the detection threshold χ i , indicating that these three drones have actuator failures. However, the vertical axis for drone v4 remains less than 0 throughout the simulation time, indicating that drone v4 has no actuator failure. This demonstrates that the multi-drone control method of the present invention can promptly and accurately identify drone actuator failures and precisely locate the specific drone.

[0201] refer to Figure 6 and Figure 7 The figure shows the trajectory tracking results of the four UAV formation in the three-dimensional space x, y, and z, as well as a two-dimensional x and y top-down view. Start1, Start2, Start3, and Start4 represent the starting points of the four UAVs, respectively, and QUAV1, QUAV2, QUAV3, and QUAV4 represent the positions of the four UAVs at the end of the simulation. It is clearly observed that the distributed fault-tolerant control method for UAVs of the present invention allows the four UAVs to accurately track the desired trajectory even if an actuator failure occurs.

[0202] refer to Figure 8 The state error diagram of the 4-UAV formation is shown in the figure, where the horizontal axis represents the simulation time and the vertical axis represents the position tracking error of the 4 UAVs (X direction error e xi , Y direction error e yi , Z direction error e zi ) and attitude angle tracking error (roll angle error e Φi , pitch angle error e θi , yaw angle error e ψi It is clearly observed that even in the presence of an actuator failure, the position tracking errors and attitude angle tracking errors of the four drones eventually converge to zero. This demonstrates that, using the distributed fault-tolerant control method for drones presented herein, the four drones can continue to track the desired trajectory and maintain a stable attitude angle, even without knowing the specific details of the fault.

[0203] refer to Figure 9This is a controller curve change diagram for a formation of four UAVs. The horizontal axis represents the simulation time, and the vertical axis represents the changes in the four control inputs of the four UAVs. It can be clearly observed that when an actuator failure occurs in a UAV, the control input of the corresponding UAV will be adjusted in time to deal with the actuator failure in time.

[0204] refer to Figure 10 , the horizontal axis represents the simulation time, and the vertical axis represents the position tracking of UAV 1 (X direction error e xi , Y direction error e yi , Z direction error e zi ) error, compared with the PID controller, it is shown that the distributed fault-tolerant control method for multiple UAVs of the present invention can make the tracking error always converge to 0. The comparison results clearly show the superiority of the distributed fault-tolerant control method for multiple UAVs of the present invention.

[0205] It should be noted that, in the present invention, if the units of parameters are not specified, the units are dimensionless.

[0206] The above content is only used to illustrate the technical solution of the present invention, rather than to limit the present invention. Although the present invention has been described in detail with reference to the foregoing, it should be understood by those skilled in the art that the technical solution described above can still be modified, or some or all of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solution to deviate from the scope of the technical solution of the present invention.

Claims

1. A multi-UAV distributed fault-tolerant control method, characterized in that: The following steps are involved: S1: Construct virtual control laws for multi-UAV position loops and multi-UAV attitude loops; S2: Based on Nussbaum function, a distributed fault-tolerant controller for the position loop is constructed according to the virtual control law of the multi-UAV position loop; Based on Nussbaum function, a distributed fault-tolerant controller for attitude loop is constructed according to the virtual control law of multi-unmanned attitude loop. S3: Obtain the expected position parameters and expected attitude parameters of the faulty UAV in the multi-UAV formation; S4: Substitute the expected position parameters of the faulty UAV into the position loop distributed fault-tolerant controller to perform fault-tolerant processing on the flight trajectory of the faulty UAV; substitute the expected attitude parameters of the faulty UAV into the attitude loop distributed fault-tolerant controller to perform fault-tolerant processing on the attitude of the faulty UAV; by performing fault-tolerant processing on the flight trajectory of the faulty UAV and the fault-tolerant processing on the attitude of the faulty UAV, distributed fault-tolerant control of multiple UAVs is performed.

2. The multi-UAV distributed fault-tolerant control method according to claim 1 is characterized in that: The position loop distributed fault-tolerant controller is: Where, Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ; T i is the total lift of the i-th faulty UAV, unit: kg·m / s 2 ; Y i (ξ i )=diag{Y i (ξ i,1 ),Y i (ξ i,2 ),Y i (ξ i,3 )},ξ i,1 is the first element of the Nussbaum term of the position loop of the i-th faulty UAV; i,2 is the second element of the Nussbaum term of the position loop of the i-th faulty UAV; i,3 is the third element of the Nussbaum term of the position loop of the i-th faulty UAV; Y i (ξ i,1 ) is the first element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ i,2 ) is the second element of the Nussbaum gain of the position loop of the i-th faulty UAV; Y i (ξ i,3 ) is the third element of the Nussbaum gain of the position loop of the i-th faulty UAV; Unit: m 2 / s 3 ; for ξ i,1 The derivative of for ξ i,2 The derivative of for ξ i,3 The derivative of is x i2d The derivative of, unit: m / s 2 ;x i2 =D i , unit: m / s; D i is the velocity vector of the i-th faulty UAV in the ground coordinate system, unit: m / s; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; X1=ω i1 +ω i2 L ii ;ω i1 is the weight coefficient of the position loop error of the i-th faulty UAV; ω i2 is the weight coefficient of the position loop error with the neighboring aircraft of the i-th faulty UAV; L ii is the element of the Laplace matrix of the i-th faulty UAV; Unit: m; e i1 is the position error of the i-th faulty UAV, unit: m / s; j is the serial number of the UAV in the UAV formation that can communicate with the i-th faulty UAV, i≠j; e j1 is the position error of the jth UAV, unit: m / s; L ij To divide L ii Elements in the Laplacian matrix other than N i is the set of neighbors of the i-th faulty UAV; K 12 is the positive definite symmetric matrix to be designed in the position loop distributed fault-tolerant controller, unit: s -1 ; e i2 is the speed tracking error of the i-th faulty UAV, unit: m / s; u i is the position loop controller of the i-th faulty UAV, unit: m / s 2 ; The attitude ring distributed fault-tolerant controller is: Where, J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ; M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 ; is x i4d The derivative of , in rad / s 2 ;x i4 =Ω i , unit: rad / s; Ω i Angular velocity of the i-th faulty UAV, unit: rad / s; Unit: rad / s 2 ; Ω i is the angular velocity of the i-th faulty UAV; e i4 is the angular velocity error of the i-th faulty UAV, unit: rad / s 2 ; K 22 is the positive definite symmetric matrix of the attitude loop distributed fault-tolerant controller to be designed, unit: s -1 ; H i is the attitude kinematic matrix of the i-th faulty UAV; X2=ω i3 +ω i4 L ii ;ω i3 is the weight coefficient of the attitude loop error of the i-th faulty UAV; ω i4 is the weight coefficient of the attitude loop error with the neighboring aircraft of the i-th faulty UAV; Unit: rad; e i3 is the attitude angle error of the i-th faulty UAV, unit: rad; e j3 is the attitude angle error of the j-th UAV, unit: rad; Y i (ζ i )=diag{Y i (ζ i,1 ),Y i (ζ i,2 ),Y i (ζ i,3 )},ζ i,1 is the first element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,2 is the second element of the Nussbaum term of the attitude loop of the i-th faulty UAV; i,3 is the third element of the Nussbaum term of the attitude loop of the i-th faulty UAV; Y i (ζ i,1 ) is the first element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,2 ) is the second element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Y i (ζ i,3 ) is the third element of the Nussbaum gain of the attitude loop of the i-th faulty UAV; Unit: rad 2 / s 3 ; For i,1 The derivative of For i,2 The derivative of For i,3 The derivative of u ia is the attitude loop controller of the i-th faulty UAV, unit: rad / s 2 .

3. The multi-UAV distributed fault-tolerant control method according to claim 1, characterized in that: The expected position parameters of the faulty UAV include the expected roll angle of the faulty UAV; The expected attitude parameters of the faulty UAV include the pitch angle of the faulty UAV and the total lift of the faulty UAV.

4. A multi-UAV control method, characterized in that: include: S1: Construct an auxiliary observer, form a dynamic representation of the residual based on the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation based on the dynamic representation of the residual; S2: Set the detection threshold χ of the UAV actuator failure and determine whether the fault information of the faulty UAV is greater than the detection threshold χ; If it is greater than, the multi-UAV distributed fault-tolerant control method according to any one of claims 1 to 3 is used to perform fault-tolerant processing on the flight trajectory and posture of the faulty UAV.

5. The multi-UAV control method according to claim 4, characterized in that: The auxiliary observer is: Where, is x i2 Derivative of the estimated value, unit: m / s 2 ;x i2 =D i , unit: m / s; D i is the velocity vector of the i-th faulty UAV in the ground coordinate system; Unit: kg -1 ; m is the mass of the i-th faulty UAV, unit: kg; O ix =cosΦ i sinθ i cosΨ i +sinΦ i sinΨ i ;O iy =cosΦ i sinθ i sinΨ i -sinΦ i cosΨ i ;O iz =cosΦ i sinθ i Φ i is the roll angle of the i-th faulty UAV, unit: rad; θ i is the pitch angle of the i-th faulty UAV, unit: rad; Ψ i is the yaw angle of the i-th faulty UAV, unit: rad; U i =[-T i ,-T i ,-T i ] T , unit: kg·m / s 2 ; T i is the total lift of the i-th faulty UAV, unit: N; G a =[0,0,g] T , g is the acceleration due to gravity, unit: m / s 2 ; is x i2 The estimated value of is in m / s; Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop, in s -1 ; is x i4 Derivative of the estimated value, unit: m / s 2 ; Unit: rad / s 2 ; Ω i is the angular velocity of the i-th faulty UAV; J i is the moment of inertia matrix of the i-th faulty UAV, unit: (kg·m 2 ) -1 ; M i is the three-axis torque of the i-th faulty UAV, unit: kg·m 2 / s 2 ; x i4 =Ω i , unit: rad / s; is x i4 The estimated value of is in rad / s; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-unmanned attitude loop, in s -1 .

6. The multi-UAV control method according to claim 5, characterized in that: The system residual of the auxiliary observer forms a dynamic representation of the residual for: Where A is the system matrix of the auxiliary observer; r i is the system residual of the auxiliary observer; k i is the fault factor matrix of the i-th faulty UAV; Q = diag{Q1,Q2}, Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV attitude loop; I6 is the identity matrix with dimension 6; U = [U i ,M i ] T , U i =[-T i ,-T i ,-T i ] T , T i is the total lift of the i-th faulty UAV, unit: N; M i is the three-axis torque of the i-th faulty UAV.

7. The multi-UAV control method according to claim 4, characterized in that: The detection threshold χ is: Where, χ i is the detection threshold of the i-th faulty UAV; P is a positive definite matrix; λ b (P) is the largest eigenvalue of a positive definite matrix; λ s (P) is the minimum eigenvalue of the positive definite matrix; t0 is the initial time; t is time; γ=PQ+Q T P-PAA T P, Q = diag{Q1, Q2}, Q1 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV position loop; Q2 is the positive definite symmetric matrix to be designed in the virtual control law of the multi-UAV attitude loop; A is the system matrix of the auxiliary observer; λ s (γ) is the minimum eigenvalue of γ; x i (t0) is the initial value of the state of the i-th faulty UAV; is the estimated value of the initial state of the i-th faulty UAV.

8. A multi-UAV control system, characterized in that: include: Fault detection module: used to build an auxiliary observer, form a dynamic representation of the residual based on the system residual of the auxiliary observer, and detect the fault information of the faulty UAV in the multi-UAV formation based on the dynamic representation of the residual; And a fault-tolerant processing module: used to set a detection threshold χ for a failure of an actuator of a UAV, and determine whether the fault information of the faulty UAV is greater than the detection threshold χ; if greater, using the multi-UAV distributed fault-tolerant control method described in any one of claims 1 to 3 to perform fault-tolerant processing on the flight trajectory and posture of the faulty UAV.

9. A memory, characterized in that: A program file is stored, and the program file is executed to implement program instructions formed by the multi-UAV control method according to any one of claims 4 to 7.

10. An electronic device, characterized in that: The system comprises a processor and a memory coupled to each other, wherein: The memory is used to store program instructions formed by the multi-UAV control method according to any one of claims 4 to 7; The processor is configured to execute program instructions stored in the memory.

Images