Fault-tolerant clustering emergence control manned-unmanned aerial vehicle cluster fault direction identification method

Abstract

The invention relates to a manned-unmanned aerial vehicle cluster fault direction identification method based on fault-tolerant clustering emergence control, and belongs to the technical field of fault tolerance of unmanned aerial vehicle clusters. The method comprises the following steps: establishing an attitude system model of a manned-unmanned aerial vehicle cluster under a reverse fault; establishing a distributed preset performance observer, and performing multi-target clustering emergence control; a buffeting-free preset performance fault-tolerant controller is established, target information estimated by the distributed preset performance observer is tracked, fault tolerance of reverse faults is achieved, and the buffeting phenomenon is restrained; and calculating a fault identification index by using the parameters of the preset performance fault-tolerant controller, and identifying a fault direction. According to the method, attitude angle distributed buffeting-free fault-tolerant multi-target clustering emergence of a manned-unmanned aerial vehicle cluster attitude system can be realized, and identification and early warning are carried out on the generated reverse fault, so that high performance and safety of manned-unmanned aerial vehicle cluster operation are ensured; the identification of the fault direction provides a fault positioning basis for solving the reverse fault from the physical level.

Description

Technical Field

[0001] The present invention relates to the technical field of fault tolerance of unmanned aerial vehicle clusters, and in particular to a method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control. Background Art

[0002] In recent years, drones have been widely used in emergency rescue, aerial photography, plant protection, power inspection, geological exploration, film and television shooting and other fields. Compared with a single drone, a drone cluster completes various tasks through communication and coordination between drones, and has higher application and research value. For the control of drone clusters, manned-drone cluster control is an effective control method, which guides the follower drones to achieve the control target by controlling the leader drone by the operator. This manned-drone cluster control algorithm has strong robustness and can effectively avoid accidents. On the other hand, emergent control is an effective method for dealing with complex systems by adjusting local behaviors to achieve the expected overall behavior. The clustered multi-target emergence of manned-drone clusters is to divide the drone cluster into different clusters, and use the drone controlled by the operator to tow some drones to achieve the synchronization of the postures of drones in the same cluster and different postures of drones in different clusters.

[0003] In addition, the rotor of the drone may experience physical failure, which manifests as excessive or insufficient torque input to the drone. In addition, a special type of failure is in the yaw angle channel of the attitude system where the torque is determined by the rotor differential. A rotor failure may cause a reverse torque failure in the yaw angle channel, which may cause the performance of a single drone to degrade, or even cause the drone to crash or even endanger the safe operation of the entire drone cluster. Therefore, achieving fault-tolerant control of the manned-drone cluster under reverse failure is crucial to the safe and stable operation of the drone cluster.

[0004] For reverse faults, existing fault-tolerant methods are based on Nussbaum functions to achieve fault-tolerant control. However, the existing reverse fault-tolerant methods based on Nussbaum functions have inevitable defects. That is, under this method, since the Nussbaum function will produce a sign jump when looking for the correct control direction after a reverse fault occurs, it will inevitably cause the system state to vibrate. This defect is caused by the characteristics of the Nussbaum function method itself, so it is impossible to eliminate the occurrence of the jitter phenomenon from the root. However, the jitter phenomenon cannot be ignored in the actual application of drones, so it is necessary to design a jitter-free controller to suppress the occurrence of the jitter phenomenon.

[0005] On the other hand, compared with multiplicative faults, reverse faults are more harmful and require timely warning and troubleshooting. Although the existing fault-tolerant control method has achieved performance-guaranteed control under reverse faults. However, it is impossible to identify and warn the reverse fault, resulting in the inability to timely detect and eliminate the reverse fault. Existing fault direction identification methods are all concentrated in the power grid, which identifies the fault direction through the phase angle between voltage and current. This method cannot be used in the UAV attitude system because there is no corresponding reference signal in the attitude system. Therefore, in order to ensure the safe, reliable and stable operation of manned-UAV clusters under reverse faults, the present invention proposes a non-jittering fault-tolerant cluster synchronization control method and a fault direction identification algorithm to solve the above problems. Summary of the invention

[0006] In view of the above analysis, the present invention aims to disclose a method for identifying the fault direction of a manned-UAV cluster with fault-tolerant cluster emergence control; to solve the jitter problem existing in the current fault-tolerant method based on the Nussbaum function mentioned in the background technology, to achieve multi-target cluster emergence without jitter in the attitude angle of the manned-UAV cluster attitude system and to identify and warn of the reverse fault that occurs, thereby ensuring the high performance and safety of the manned-UAV cluster. The identification of the fault direction provides a fault location basis for solving the reverse fault from a physical level.

[0007] The present invention discloses a method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control, comprising:

[0008] S1. Establish the dynamic model of the attitude system of manned-unmanned aerial vehicle cluster under reverse fault;

[0009] S2. Establish a distributed preset performance observer, obtain the leader UAV information in the cluster according to the traction node in each cluster, and use the conversion error information of the cluster synchronization error interacted by the UAV communication network to estimate the leader UAV information for the follower UAVs in each cluster that cannot directly obtain the leader UAV information, as their target information, and perform multi-target cluster emergence control;

[0010] Each cluster of a manned-UAV swarm includes a leader UAV controlled by a leader and multiple follower UAVs;

[0011] S3, establish a chatter-free preset performance fault-tolerant controller, track the target information estimated by the distributed preset performance observer, achieve fault tolerance for reverse faults and suppress chattering;

[0012] S4. Calculate the fault identification index using the parameters of the preset performance fault-tolerant controller to identify the fault direction.

[0013] Furthermore, under reverse fault, the dynamic model of the attitude system of the i-th following rotor UAV is:

[0014]

[0015] Among them, x i,11 、x i,21 、x i,31 Respectively represent the roll angle, pitch angle and yaw angle of the i-th following rotor UAV; x i,12 、x i,22 、x i,32 Respectively represent the angular velocities of the roll, pitch and yaw channels of the i-th following rotor UAV; d i,p2 Indicates interference; i,p Represents the output signal; f i,p represents a nonlinear term; represents the input under fault conditions;

[0016] The dynamic model of the attitude system of the leader UAV is:

[0017]

[0018] in, and Respectively represent the angular position and angular velocity of the leader UAV in the roll, pitch and yaw channels; l i The code that represents the cluster where the drone is located; represents the output of the leader drone; It represents the operational input of the leader; represents a nonlinear term;

[0019] The output of the leader drone does not have mutations; it has constants make Established.

[0020] Furthermore, under fault conditions the input for:

[0021]

[0022] Among them, u i,p Represents the input signal of the attitude system design of the following rotor UAV; ρ i,p represents the multiplicative factor of the fault, which satisfies ρ in the roll channel and the pitch channel i,p > 0, satisfying ρ in the yaw channel i,p ≠0; in the yaw channel, there is a reverse fault. At this time, ρ i,p <0; r i,p Indicates a bias fault.

[0023] Furthermore, the distributed preset performance observer is expressed as:

[0024]

[0025] Among them, u i ∈R 3 With ξ i ∈R 3 are the input and output signals of the observer respectively; N is the number of following drones in the cluster;

[0026] Input signal u at time t i The calculation expression is:

[0027]

[0028] Wherein, c2 and c3 are positive constant gains; ij The matrix Ξ=(L+B) -1 The elements of , L and B are the Laplace matrix and traction matrix of the manned-unmanned aerial vehicle cluster communication topology respectively; μ 1j and μ 2j is the auxiliary variable matrix; is the kth jth following drone j kth j +1 trigger moment, is the kth jth following drone j The conversion error of the cluster synchronization error at the triggering time, δ j satisfy is δ j The estimated value of .

[0029] Furthermore, in the observer input of the i-th following rotorcraft, the conversion error i,p ,p=1,2,3The calculation formula is:

[0030]

[0031] Among them, a ij 、a im is the element of the adjacency matrix that follows the UAV cluster communication topology; is the output vector of the leader drone; Represents cluster l i The set of nodes that send information to the following drones in other clusters, Represents cluster l i The set of nodes that follow the drones in other clusters and receive information;

[0032] When the i-th following rotorcraft UAV meets: 1) no neighbor nodes or 2) or 3) When , the i-th following rotor UAV serves as the traction node;

[0033] b i∈{0,1} is the traction gain, which constitutes the traction matrix B=diag{b1,…,b N}, when the i-th following rotor UAV is the towing node, b i =1, otherwise, b i =0; is the set of neighbor nodes of drone i; ∩ represents the intersection of two sets;

[0034]

[0035] Among them, γ i (t) = [γ i,1 ,γ i,2 ,γ i,3 ] T Used to characterize the boundary distance, h(γ i,p (t)) is the error conversion function, c1 is the constant gain, 13 is the three-dimensional column vector of all 1s, and the symbol It means that the elements of two vectors are multiplied respectively, a1 is the set safety constant, and β is the set safety margin;

[0036] Auxiliary variables μ in the auxiliary variable matrix 1j,p and μ 2j,p , the calculation formulas for p=1,2,3 are:

[0037]

[0038] in, for The p-th element of .

[0039] Furthermore, the event triggering conditions for the communication between drones are:

[0040]

[0041] in, is the difference between the conversion error at the current moment and the conversion error at the most recent trigger moment, η j is a positive constant trigger threshold, sup{} represents the supremum of the variable, min{} represents the minimum value of the set, and λ max () represents the maximum eigenvalue of the matrix.

[0042] Furthermore, the chatter-free preset performance fault-tolerant controller adopts the Nussbaum function, and suppresses the chattering phenomenon that occurs when the Nussbaum function method solves the reverse fault through the preset performance method. The chatter-free preset performance fault-tolerant controller is expressed as:

[0043]

[0044] Among them,i,p is an adaptive variable; is an auxiliary variable; is the Nussbaum function used; p = 1, 2, 3.

[0045] Furthermore, the Nussbaum function It is expressed as:

[0046]

[0047] In the formula, σ1 and σ2 are parameters for adjusting the amplitude and frequency of the Nussbaum function respectively;

[0048] In the controller, the adaptive variable υ i,p and auxiliary variables The calculation formula is as follows:

[0049]

[0050] In the formula, q i,p is a positive constant gain, The output signal of the drone angular velocity and the first-order filter The error between is the intermediate variable, z i,p1 is the conversion error, h(κ i,p (t)) is the conversion function, κ i,p is the boundary distance for adjusting the error, l i is a constant gain, a 2i,p is the positive safety constant of the design;

[0051] is the adaptive gain of the neural network, Z i,p =[x i,11 ,x i,12 ,x i,21 ,x i,22 ,x i,31 ,x i,32 ] T represents the input of the neural network, is the radial basis function vector.

[0052] Furthermore, the calculation formula of the radial basis function vector is:

[0053]

[0054] Among them, M1 is the number of neurons, o m and π m represent the center and width of the neuron, respectively;

[0055] The conversion error z i,p1, the boundary distance of adjustment error κ i,p , conversion function h(κ i,p The calculation formulas for (t) are:

[0056]

[0057] Among them, e i,p The adjustment error of the UAV output signal;

[0058] The first-order filter is expressed as:

[0059]

[0060] Among them, ν i,p is a positive time constant, α i,p (t) is the virtual control law as the filter input, and its calculation formula is:

[0061]

[0062] In the formula, Ω i,p is an auxiliary variable, and its calculation formula is:

[0063]

[0064] in, is an intermediate variable.

[0065] Adaptive Gain The update law is:

[0066]

[0067] Among them, r i,p1 and r i,p2 is a positive constant gain.

[0068] Furthermore, the S4 specifically includes the following contents:

[0069] S4.1, select the sampling time interval h and Moments for the Nussbaum function Take samples;

[0070] S4.2. Settings And store the last two sampling values ​​of the Nussbaum function and

[0071] S4.3. Initialization of identification index J i,p (0) = 1;

[0072] S4.4. Calculate the fault direction identification index:

[0073]

[0074] S4.5 If J i,p (t) = 1, indicating no reverse fault, jump to S4.9, otherwise jump to S4.6;

[0075] S4.6, if J i,p (t) = -1, indicating that there is a reverse fault and an alarm is issued, jumping to S4.9, otherwise jumping to S4.7;

[0076] S4.7 If Indicates that the current moment cannot be judged, and exits this loop, otherwise jumps to S4.8;

[0077] S4.8. Calculate the fault direction identification index:

[0078] Jump to S4.5;

[0079] S4.9, end the judgment;

[0080] S4.10, when J i,p When (Qh)=-1, output t=Qh.

[0081] The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control disclosed in the present invention can achieve the following beneficial effects:

[0082] 1) The control strategy designed in the present invention does not require prior knowledge of the initial conditions of the drone, and the designed observer is a first-order integral structure, which simplifies the design difficulty and computational complexity of the observer. In addition, the control strategy of the present invention can autonomously set the system convergence time and convergence accuracy as needed, thereby improving the performance of the manned-drone cluster under fault conditions.

[0083] 2) This method can realize distributed non-jittering fault-tolerant multi-target clustering emergence control of manned-unmanned aerial vehicle clusters. The leading drone affects the traction node to achieve multi-target clustering emergence of the entire drone cluster. The present invention uses the preset performance method to suppress the state jitter phenomenon that occurs when the Nussbaum function method solves the reverse fault, ensures that the output signal of the follower drone is a smooth signal, and improves the performance of the system.

[0084] 3) The fault direction identification algorithm in the present invention does not rely on a specific reference signal, but only identifies the fault direction based on the parameters in the controller. Therefore, the method of the present invention not only ensures the safe and stable operation of the manned-unmanned aerial vehicle cluster under reverse fault conditions, but also provides a fault location basis for further solving reverse faults from a physical aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0086] Figure 1 It is a schematic flow chart of a method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control in an embodiment of the present invention;

[0087] Figure 2 This is a control principle diagram of a method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control in an embodiment of the present invention. DETAILED DESCRIPTION

[0088] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used to illustrate the principles of the present invention together with the embodiments of the present invention.

[0089] An embodiment of the present invention discloses a method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control, such as Figure 1 As shown, the following steps are included:

[0090] S1. Establish the dynamic model of the attitude system of manned-unmanned aerial vehicle cluster under reverse fault;

[0091] S2. Establish a distributed preset performance observer, obtain the leader UAV information in the cluster according to the traction node in each cluster, and use the conversion error information of the cluster synchronization error interacted by the UAV communication network to estimate the leader UAV information for the follower UAVs in each cluster that cannot directly obtain the leader UAV information, as their target information, and perform multi-target cluster emergence control;

[0092] Each cluster of a manned-UAV swarm includes a leader UAV controlled by a leader and multiple follower UAVs;

[0093] S3, establish a chatter-free preset performance fault-tolerant controller, track the target information estimated by the distributed preset performance observer, achieve fault tolerance for reverse faults and suppress chattering;

[0094] S4. Calculate the fault identification index using the parameters of the preset performance fault-tolerant controller to identify the fault direction.

[0095] Combination Figure 1 and Figure 2 The specific implementation process of the method of the present invention includes the following contents:

[0096] Specifically, in S1, under reverse fault, the dynamic model of the attitude system of the i-th following rotor UAV is:

[0097]

[0098] Among them, x i,11 、x i,21 、x i,31 Respectively represent the roll angle, pitch angle and yaw angle of the i-th following rotor UAV; x i,12 、x i,22 、x i,32 Respectively represent the angular velocities of the roll, pitch and yaw channels of the i-th following rotor UAV; d i,p2 Indicates interference; i,p Represents the output signal; f i,p represents a nonlinear term; represents the input under fault conditions;

[0099] More specifically, under fault conditions the input for:

[0100]

[0101] Among them, u i,p Represents the input signal of the attitude system design of the following rotor UAV; ρ i,p represents the multiplicative factor of the fault, which satisfies ρ in the roll channel and the pitch channel i,p > 0, satisfying ρ in the yaw channel i,p ≠0; in the yaw channel, there is a reverse fault. At this time, ρ i,p <0; r i,p Indicates a bias fault.

[0102] Nonlinear term f i,p It is expressed as:

[0103]

[0104] Among them, J x , J y , J z Represents the moment of inertia in different directions, J r is the rotor inertia, I1, I2, I3 are the aerodynamic damping coefficients, Ω r is the residual angular velocity of all motors of the rotorcraft. In addition, the topological graph of the N following drones contains a directed spanning tree.

[0105] The dynamic model of the attitude system of the leader UAV is:

[0106]

[0107] in, and Respectively represent the angular position and angular velocity of the leader UAV in the roll, pitch and yaw channels; l iThe code that represents the cluster where the drone belongs; represents the output of the leader drone; It represents the operational input of the leader; represents the nonlinear term; its expression is the same as that of the following drone.

[0108] The output of the leader drone does not have mutations; it has constants make Established.

[0109] Specifically, in S2, the constructed distributed preset performance observer is expressed as:

[0110]

[0111] Among them, u i ∈R 3 With ξ i ∈R 3 are the input and output signals of the observer respectively; N is the number of following drones in the cluster;

[0112] Input signal u at time t i The calculation expression is:

[0113]

[0114] Wherein, c2 and c3 are positive constant gains; ij The matrix Ξ=(L+B) -1 The elements of , L and B are the Laplace matrix and traction matrix of the manned-unmanned aerial vehicle cluster communication topology respectively; μ 1j and μ 2j is the auxiliary variable matrix μ 1j =diag{μ 1j,1 ,μ 1j,2 ,μ 1j,3} and μ 2j =[μ 2j,1 ,μ 2j,2 ,μ 2j,3 ] T is the auxiliary variable matrix; is the kth jth following drone j kth j +1 trigger moment, is the kth jth following drone j The conversion error of the cluster synchronization error at the triggering time, ∈ j =[∈ j,1 ,∈ j,2 ,∈ j,3 ] T , δ j =[δ j,1 ,δ j,2,δ j,3 ] T satisfy is δ j The subscripts 1, 2, and 3 in the formula represent the roll, pitch, and yaw channels, respectively.

[0115] Among them, in the observer input of the i-th following rotor UAV, the conversion error i,p ,p=1,2,3The calculation formula is:

[0116]

[0117] Among them, a ij 、a im is the element of the adjacency matrix that follows the UAV cluster communication topology; is the output vector of the leader drone; Represents cluster l i The set of nodes that send information to the following drones in other clusters, Represents cluster l i The set of nodes that follow the drones in other clusters and receive information;

[0118] When the i-th following rotorcraft UAV satisfies: 1) there is no neighbor node or 2) or 3) When , the i-th following rotor UAV serves as the traction node;

[0119] The node without neighbor refers to a drone that has no communication with all the following drones.

[0120] b i ∈{0,1} is the traction gain, which constitutes the traction matrix B=diag{b1,…,b N}, when the i-th following rotor UAV is the towing node, b i =1, otherwise, b i =0; is the set of neighbor nodes of drone i; ∩ represents the intersection of two sets. In addition:

[0121]

[0122] Among them, γ i (t) = [γ i,1 ,γ i,2 ,γ i,3 ] T Used to characterize the boundary distance, h(γ i,p (t)) is the error conversion function, c1 is the constant gain, 13 is the three-dimensional column vector of all 1s, and the symbol It means that the elements of two vectors are multiplied respectively, a1 is the set safety constant, and β is the set safety margin;

[0123] The calculation formula of the safety margin is as follows:

[0124]

[0125] Among them, T is the designed convergence time, and ε is the designed convergence accuracy.

[0126] Auxiliary variables μ in the auxiliary variable matrix 1j,p and μ 2j,p , the calculation formulas for p=1,2,3 are:

[0127]

[0128] in, for The p-th element of .

[0129] The event trigger conditions for communication between drones are:

[0130]

[0131] in, is the difference between the conversion error at the current moment and the conversion error at the most recent trigger moment, η j is a positive constant trigger threshold, sup{} represents the supremum of the variable, min{} represents the minimum value of the set, and λ max () represents the maximum eigenvalue of the matrix.

[0132] Define observation error:

[0133]

[0134] Taking the derivative of the observation error, we can obtain:

[0135]

[0136] Furthermore, the compact form of N following drones can be expressed as:

[0137]

[0138] in, ∈=col{∈1,...,∈ N},∈ i =[∈ i,1 ,∈ i,2 ,∈ i,3 ] T ,μ1=diag{μ 11 ,...,μ 1N},μ 1i =diag{μ 1i,1 ,μ 1i,2 ,μ 1i,3},μ2=diag{μ 21 ,...,μ 2N},μ 2i =[μ 2i,1 ,μ 2i,2 ,μ 2i,3 ] T ; Ψ=col{Ψ1,…,Ψ N},Ψ i =[Ψ i,1 ,Ψ i,2 ,Ψ i,3 ] T .

[0139] Based on (9) and (16), we can get:

[0140]

[0141] in,

[0142] Taking the derivative of (8), its compact form can be expressed as:

[0143]

[0144] Similarly, the compact form of (7) is:

[0145]

[0146] in,

[0147] Consider a Lyapunov function of the following form:

[0148]

[0149] Taking the derivative of (22) we can get:

[0150]

[0151] Where δ=col{δ1,…,δ N} and c3δ T δ / 2 is a constant.

[0152] Considering the event triggering condition (15), (23) can be expressed as:

[0153]

[0154] in, Therefore, Vo (t) i,p as well as The designed observer (5) can estimate the output of the leader UAV and the observation error converges to the preset accuracy within the preset time.

[0155] Specifically, in S3, the chatter-free preset performance fault-tolerant controller is constructed using the Nussbaum function, and the chattering phenomenon that occurs when the Nussbaum function method solves the reverse fault is suppressed by the preset performance method, thereby realizing the tracking of the observer output signal. The chatter-free preset performance fault-tolerant controller is expressed as:

[0156]

[0157] Among them, i,p is an adaptive variable; is an auxiliary variable; is the Nussbaum function used; p = 1, 2, 3.

[0158] More specifically, the Nussbaum function It is expressed as:

[0159]

[0160] Where σ1 and σ2 are parameters for adjusting the amplitude and frequency of the Nussbaum function, respectively.

[0161] In the controller, the adaptive variable υ i,p and auxiliary variables The calculation formula is as follows:

[0162]

[0163] In the formula, q i,p is a positive constant gain, The output signal of the drone angular velocity and the first-order filter The error between is the intermediate variable, z i,p1 is the conversion error, h(κ i,p (t)) is the conversion function, κ i,p is the boundary distance for adjusting the error, l i is a constant gain, a 2i,p is the positive safety constant of the design, is the adaptive gain of the neural network, Z i,p =[x i,11 ,x i,12 ,x i,21 ,x i,22 ,x i,31 ,x i,32 ]T represents the input of the neural network, is the radial basis function vector, and its calculation formula is:

[0164]

[0165] Among them, M1 is the number of neurons, o m and π m Represent the center and width of the neuron respectively.

[0166] Conversion error z i,p1 , the boundary distance of adjustment error κ i,p , conversion function h(κ i,p The calculation formula of (t)) is:

[0167]

[0168] e i,p =y i,p -ξ i,p (31)

[0169]

[0170] Among them, e i,p is the adjustment error of the drone output signal.

[0171] The first-order filter is expressed as:

[0172]

[0173] Among them, ν i,p is a positive time constant, α i,p (t) is the virtual control law as the filter input, and its calculation formula is:

[0174]

[0175] In the formula, Ω i,p is an auxiliary variable, and its calculation formula is:

[0176]

[0177] in, is an intermediate variable.

[0178] Adaptive Gain The update law is

[0179]

[0180] Among them, r i,p1 and r i,p2 is a positive constant gain.

[0181] The above controller is designed based on the backstepping method. First, consider the following form of Lyapunov function:

[0182]

[0183] Combining (1), (30)-(33), (35), taking the derivative of (39) and using Young's inequality, we can obtain:

[0184]

[0185] in,

[0186] Further consider the following Lyapunov function:

[0187]

[0188] in, yes Estimates.

[0189] Combining (25)-(28), the derivative of (41) is expressed as:

[0190]

[0191] Substituting (38) into (42), we obtain the following inequality:

[0192]

[0193] in, and is a constant.

[0194] Integrating both sides of (43), we obtain:

[0195]

[0196] in, is a constant. Then V i,p (t),υ i,p (t) and Therefore, the adjustment error of the UAV converges to the preset accuracy within the preset time, thus achieving fault-tolerant control under reverse faults.

[0197] Specifically, in S4, based on the proposed chatter-free fault-tolerant controller, the fault direction identification index is calculated and whether a reverse fault occurs is determined based on the index. The fault direction identification process of the drone cluster includes:

[0198] S4.1, select the sampling time interval h and Moments for the Nussbaum function Take samples;

[0199] S4.2. Settings And store the last two sampling values ​​of the Nussbaum function and

[0200] S4.3. Initialization of identification index J i,p (0) = 1;

[0201] S4.4. Calculate the fault direction identification index:

[0202]

[0203] S4.5 If J i,p (t) = 1, indicating no reverse fault, jump to S4.9, otherwise jump to S4.6;

[0204] S4.6, if J i,p (t) = -1, indicating that there is a reverse fault and an alarm is issued, jumping to S4.9, otherwise jumping to S4.7;

[0205] S4.7 If Indicates that the current moment cannot be judged, and exits this loop, otherwise jumps to S4.8;

[0206] S4.8. Calculate the fault direction identification index:

[0207] Jump to S4.5;

[0208] S4.9, end the judgment;

[0209] S4.10, when J i,p When (Qh)=-1, output t=Qh.

[0210] In order to verify the reliability of the above algorithm, the Lyapunov function (41) is considered in the case of no fault, and its derivative can be obtained:

[0211]

[0212] At this point, the control law of the drone is designed as:

[0213]

[0214] Substituting (46) into (45), we can obtain:

[0215]

[0216] In the formula, and The definition of is the same as (43). Therefore, the control law (46) under fault-free conditions can ensure the stability of the system.

[0217] Comparing (25)-(27), we get Combined with (25) It can be seen that it is reliable to identify the fault direction using the sign of the Nussbaum function.

[0218] In this embodiment, stability analysis shows the effectiveness of the designed manned-UAV cluster's jitter-free preset performance fault-tolerant control method and fault direction identification method, which can effectively solve the jitter problem caused by the Nussbaum function method under reverse faults and accurately identify the time of the reverse fault and the channel where the fault is located, thereby improving the robustness of the manned-UAV cluster and ensuring the safe and stable operation of the system.

[0219] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant cluster emergence control, characterized in that: include: S1. Establish the dynamic model of the attitude system of manned-unmanned aerial vehicle cluster under reverse fault; S2. Establish a distributed preset performance observer, obtain the leader UAV information in the cluster according to the traction node in each cluster, and use the conversion error information of the cluster synchronization error interacted by the UAV communication network to estimate the leader UAV information for the follower UAVs in each cluster that cannot directly obtain the leader UAV information, as their target information, and perform multi-target cluster emergence control; Each cluster of a manned-UAV swarm includes a leader UAV controlled by a leader and multiple follower UAVs; S3, establish a chatter-free preset performance fault-tolerant controller, track the target information estimated by the distributed preset performance observer, achieve fault tolerance for reverse faults and suppress chattering; S4. Calculate the fault identification index using the parameters of the preset performance fault-tolerant controller to identify the fault direction.

2. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 1 is characterized in that: Under reverse fault, the dynamic model of the attitude system of the i-th following rotor UAV is: Among them, x i,11 、x i,21 、x i,31 Respectively represent the roll angle, pitch angle and yaw angle of the i-th following rotor UAV; x i,12 、x i,22 、x i,32 Respectively represent the angular velocities of the roll, pitch and yaw channels of the i-th following rotor UAV; d i,p2 Indicates interference; i,p Represents the output signal; f i,p represents a nonlinear term; represents the input under fault conditions; The dynamic model of the attitude system of the leader UAV is: in, and Respectively represent the angular position and angular velocity of the leader UAV in the roll, pitch and yaw channels; l i The code that represents the cluster where the drone belongs; represents the output of the leader drone; It represents the operational input of the leader; represents a nonlinear term; The output of the leader drone does not have mutations; it has constants make Established.

3. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 2 is characterized in that: During a fault condition the input for: Among them, u i,p Represents the input signal of the attitude system design of the following rotor UAV; ρ i,p represents the multiplicative factor of the fault, which satisfies ρ in the roll channel and the pitch channel i,p > 0, satisfying ρ in the yaw channel i,p ≠0; in the yaw channel, there is a reverse fault. At this time, ρ i,p <0; r i,p Indicates a bias fault.

4. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 2 is characterized in that: The distributed pre-configured performance observer is expressed as: Among them, u i ∈R 3 With ξ i ∈R 3 are the input and output signals of the observer respectively; N is the number of following drones in the cluster; Input signal u at time t i The calculation expression is: Wherein, c2 and c3 are positive constant gains; ij The matrix Ξ=(L+B) -1 The elements of , L and B are the Laplace matrix and traction matrix of the manned-unmanned aerial vehicle cluster communication topology respectively; μ 1j and μ 2j is the auxiliary variable matrix; is the kth jth following drone j kth j +1 trigger moment, is the kth jth following drone j The conversion error of the cluster synchronization error at the triggering time, δ j satisfy is δ j The estimated value of .

5. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 4 is characterized in that: In the observer input of the i-th following rotorcraft, the conversion error ∈ i,p ,p=1,2,3The calculation formula is: Among them, a ij 、a im is the element of the adjacency matrix that follows the UAV cluster communication topology; is the output vector of the leader drone; Represents cluster l i The set of nodes that send information to the following drones in other clusters, Represents cluster l i The set of nodes that follow the drones in other clusters and receive information; When the i-th following rotorcraft UAV satisfies: 1) there is no neighbor node or 2) or 3) When , the i-th following rotor UAV serves as the traction node; b i ∈{0,1} is the traction gain, which constitutes the traction matrix B=diag{b1,…,b N }, when the i-th following rotor UAV is the towing node, b i =1, otherwise, b i =0; is the set of neighbor nodes of drone i; ∩ represents the intersection of two sets; Among them, γ i (t) = [γ i,1 ,γ i,2 ,γ i,3 ] T Used to characterize the boundary distance, h(γ i,p (t)) is the error conversion function, c1 is the constant gain, 13 is the three-dimensional column vector of all 1s, and the symbol It means that the elements of two vectors are multiplied respectively, a1 is the set safety constant, and β is the set safety margin; Auxiliary variables μ in the auxiliary variable matrix 1j,p and μ 2j,p , the calculation formulas for p=1,2,3 are: in, for The p-th element of .

6. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 5 is characterized in that: The event trigger conditions for communication between drones are: in, is the difference between the conversion error at the current moment and the conversion error at the most recent trigger moment, η j is a positive constant trigger threshold, sup{} represents the supremum of the variable, min{} represents the minimum value of the set, and λ max () represents the maximum eigenvalue of the matrix.

7. The method for identifying the fault direction of manned-unmanned aerial vehicle clusters with fault-tolerant clustering emergent control according to claim 2 is characterized in that: The chatter-free preset performance fault-tolerant controller adopts the Nussbaum function and suppresses the chattering phenomenon that occurs when the Nussbaum function method solves the reverse fault through the preset performance method. The chatter-free preset performance fault-tolerant controller is expressed as: Among them, i,p is an adaptive variable; is an auxiliary variable; is the Nussbaum function used; p = 1, 2, 3.

8. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 7 is characterized in that: Nussbaum function It is expressed as: In the formula, σ1 and σ2 are parameters for adjusting the amplitude and frequency of the Nussbaum function respectively; In the controller, the adaptive variable υ i,p and auxiliary variables The calculation formula is as follows: In the formula, q i,p is a positive constant gain, The output signal of the drone angular velocity and the first-order filter The error between is the intermediate variable, z i,p1 is the conversion error, h(κ i,p (t)) is the conversion function, κ i,p To adjust the error boundary distance, is a constant gain, a 2i,p is the positive safety constant of the design; is the adaptive gain of the neural network, Z i,p =[x i,11 ,x i,12 ,x i,21 ,x i,22 ,x i,31 ,x i,32 ] T represents the input of the neural network, is the radial basis function vector.

9. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 8 is characterized in that: The calculation formula of the radial basis function vector is: Among them, M1 is the number of neurons, o m and π m represent the center and width of the neuron, respectively; The conversion error z i,p1 , the boundary distance of adjustment error κ i,p , conversion function h(κ i,p The calculation formulas for (t) are: Among them, e i,p The adjustment error of the output signal of the UAV; The first-order filter is expressed as: Among them, ν i,p is a positive time constant, α i,p (t) is the virtual control law as the filter input, and its calculation formula is: In the formula, Ω i,p is an auxiliary variable, and its calculation formula is: in, is an intermediate variable; Adaptive Gain The update law is: Among them, r i,p1 and r i,p2 is a positive constant gain.

10. The method for identifying the fault direction of a manned-unmanned aerial vehicle cluster with fault-tolerant clustering emergent control according to claim 8, characterized in that: The S4 specifically includes the following contents: S4.1, select the sampling time interval h and at t = Q1h, Moments for the Nussbaum function Take samples; S4.

2. Settings And store the last two sampling values ​​of the Nussbaum function and S4.

3. Initialization of identification index J i,p (0) = 1; S4.

4. Calculate the fault direction identification index: S4.5 If J i,p (t) = 1, indicating no reverse fault, jump to S4.9, otherwise jump to S4.6; S4.6, if J i,p (t) = -1, indicating that there is a reverse fault and an alarm is issued, jumping to S4.9, otherwise jumping to S4.7; S4.7 If Indicates that the current moment cannot be judged, and exits this loop, otherwise jumps to S4.8; S4.

8. Calculate the fault direction identification index: Jump to S4.5; S4.9, end the judgment; S4.10, when J i,p When (Qh)=-1, output t=Qh.

Images