CN117008637A - Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle - Google Patents
Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle Download PDFInfo
- Publication number
- CN117008637A CN117008637A CN202311108504.0A CN202311108504A CN117008637A CN 117008637 A CN117008637 A CN 117008637A CN 202311108504 A CN202311108504 A CN 202311108504A CN 117008637 A CN117008637 A CN 117008637A
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- fault
- error
- radial basis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000006243 chemical reaction Methods 0.000 claims abstract description 52
- 238000013528 artificial neural network Methods 0.000 claims abstract description 47
- 240000007049 Juglans regia Species 0.000 claims abstract description 18
- 230000006870 function Effects 0.000 claims description 124
- 239000011159 matrix material Substances 0.000 claims description 22
- 230000007774 longterm Effects 0.000 claims description 18
- 230000015572 biosynthetic process Effects 0.000 claims description 16
- 230000002787 reinforcement Effects 0.000 claims description 14
- 230000001133 acceleration Effects 0.000 claims description 12
- 238000013461 design Methods 0.000 claims description 12
- 238000009499 grossing Methods 0.000 claims description 5
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 230000005484 gravity Effects 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 2
- -1 related to dynamics Chemical class 0.000 claims description 2
- 230000000452 restraining effect Effects 0.000 abstract description 2
- 230000006854 communication Effects 0.000 description 12
- 238000004891 communication Methods 0.000 description 12
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 230000007175 bidirectional communication Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Abstract
The application discloses a fault-tolerant control method for a fault unmanned aerial vehicle in a swarm unmanned aerial vehicle, and relates to the technical field of aircrafts. The method comprises the following steps: constructing a dynamics model of the ith unmanned aerial vehicle considering the fault of the actuator; constructing a conversion error according to the distributed tracking error of the ith unmanned aerial vehicle and a preset performance function, wherein the preset performance function is used for restraining the distributed tracking error; constructing a fractional order sliding mode error based on the conversion error; estimating an unknown nonlinear term by using a radial basis function neural network to obtain a nonlinear term estimated value, and estimating an unknown gain by using a Nussbaum function to obtain a gain estimated value; combining a dynamics model considering the fault of the actuator, a fractional order sliding mode error, a nonlinear item estimated value and a gain estimated value to construct a fault-tolerant control law; and determining a fault-tolerant control signal of the ith unmanned aerial vehicle based on the fault-tolerant control law and controlling the ith unmanned aerial vehicle according to the fault-tolerant control signal. The method can realize the fine fault-tolerant control of the swarm unmanned aerial vehicle.
Description
Technical Field
The application relates to the technical field of unmanned aerial vehicles, in particular to a fault-tolerant control method for a fault unmanned aerial vehicle in a swarm unmanned aerial vehicle.
Background
Along with the rapid development of the unmanned aerial vehicle industry, the networked unmanned aerial vehicle is widely applied in the fields of collaborative search and rescue, forest fire monitoring, wireless communication relay and the like. Compared with a rotor unmanned aerial vehicle, the fixed-wing unmanned aerial vehicle has long flight time and high flight speed, and can provide a large-scale monitoring/observation.
However, the rapid flight speed and complex aerodynamic characteristics present significant challenges for formation control design. To solve this problem, there are some effective formation control architectures in the related art, such as pilot-following, virtual structure frames. In addition, slip-mode control, adaptive control, backstepping control and intelligent control methods are typically integrated into the control architecture described above to facilitate control design. However, when the swarm unmanned aerial vehicle executes a task, a certain formation unmanned aerial vehicle is not processed for a long time after encountering a fault, the phenomenon of out of control is very easy to be caused, and even the unmanned aerial vehicle collides with other aircrafts when serious, and the unmanned aerial vehicle swarm flight task is difficult to estimate. Furthermore, to ensure formation flight safety in the event of a fault, user-specified formation control requirements in the event of a fault need to be explicitly considered.
Disclosure of Invention
Aiming at the problems and the technical requirements, the inventor provides a fault-tolerant control method for a fault unmanned aerial vehicle in a swarm unmanned aerial vehicle, and the technical scheme of the application is as follows:
in one aspect, a fault-tolerant control method for a fault unmanned aerial vehicle in a swarm unmanned aerial vehicle is provided, including the following steps:
constructing a dynamics model of an ith unmanned aerial vehicle, which considers the failure of an actuator and comprises an unknown nonlinear term and an unknown gain, wherein i is a positive integer;
constructing a conversion error according to the distributed tracking error of the ith unmanned aerial vehicle and a preset performance function, wherein the preset performance function is used for constraining the distributed tracking error, the conversion error is an error for converting the distributed tracking error into a preset performance deviation, and the distributed tracking error is a relative distance error between the ith unmanned aerial vehicle and other unmanned aerial vehicles in the swarm unmanned aerial vehicle;
constructing a fractional order sliding mode error based on the conversion error, wherein the fractional order sliding mode error converges in a limited time;
estimating the unknown nonlinear term by using a radial basis function neural network to obtain a nonlinear term estimation value, and estimating the unknown gain by using a Nussbaum function to obtain a gain estimation value;
combining the dynamics model considering the actuator fault, the fractional order sliding mode error, the nonlinear item estimated value and the gain estimated value to construct a fault-tolerant control law;
and under the condition that the ith unmanned aerial vehicle fails, determining a fault-tolerant control signal of the ith unmanned aerial vehicle based on the fault-tolerant control law, wherein the fault-tolerant control signal is used for controlling the ith unmanned aerial vehicle to meet formation control requirements.
Wherein, further scheme is:
the distributed tracking error of the ith unmanned aerial vehicle is as follows:
wherein e i ＝[e i1 ，e i2 ，e i3 ] T Is the distributed tracking error of the ith unmanned aerial vehicle, a ij Is a weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle, b i Is the weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle, N i Q is the total number of following unmanned aerial vehicles in the bee colony unmanned aerial vehicle i ＝[x i ，y i ，h i ] T Is the position vector, q of the ith unmanned aerial vehicle 0 ＝[x 0 ，y 0 ，h 0 ] T Is the position vector, delta of the leading unmanned aerial vehicle ij Is the expected relative distance, delta, between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle i Is the ith unmanned aerial vehicle and theA desired relative distance between the lead unmanned aerial vehicles;
the preset performance function is as follows:
wherein, xi v0 With xi vT Is positive parameter, a T Is a time constant.
The constructing a conversion error according to the distributed tracking error and a preset performance function of the ith unmanned aerial vehicle comprises:
and applying error constraint to the distributed tracking error in the following manner:
wherein k is v And (3) withV=1, 2,3, ζ for design parameters v The preset performance function is set;
and carrying out equation conversion on the error constraint to obtain:
e iv ＝ξ v φ iv (η iv )
wherein phi is iv (η iv ) Is defined as:
inverting the equation obtained by the conversion to obtain a conversion value:
based on the conversion value, the conversion error is constructed as follows:
wherein E is iv Is the conversion error.
The fractional order sliding mode error constructed based on the conversion error is as follows:
wherein lambda is 11 ，λ 12 Lambda of 13 Is a positive parameter of design, D * For fractional calculus symbols, a is a fractional calculus operator.
The dynamic model considering the failure of the actuator is thatWherein, lambda i Is an unknown gain matrix g i Is a known gain matrix, f i Is a compound item related to dynamics, b if Is the actuation deviation vector, u i0 ＝[u i01 ，u i02 ，u i03 ] T Is the control input vector corresponding to the control signal;
the combining the dynamics model considering the actuator fault, the fractional order sliding mode error, the nonlinear term estimation value and the gain estimation value to construct a fault-tolerant control law comprises the following steps:
combining the dynamics model considering the actuator fault and the conversion error, determining the second derivative of the conversion error as follows:
wherein, xi i ＝diag{Ξ i1 ，Ξ i2 ，Ξ i3 And (3)
And combining the second derivative of the conversion error to obtain a differential equation of the fractional order sliding mode error, wherein the differential equation is as follows:
wherein F is i For the unknown non-linear term,
and compensating the unknown nonlinear term and the unknown gain matrix in the fractional order sliding mode error differential equation by using the nonlinear term estimation value and the gain estimation value, and constructing the fault-tolerant control law based on the differential equation of the fractional order sliding mode error, wherein the fault-tolerant control law is as follows:
wherein,is said Λ i Corresponding said gain estimates, < >>Is said F i Corresponding to the nonlinear term estimation value, K 2 And > 0 is a positive diagonal matrix.
The radial basis function neural network is obtained based on comment-action structure reinforcement learning, and the method further comprises the following steps:
constructing a reviewer radial basis function network of the ith unmanned aerial vehicle and determining a weight learning rule of the reviewer radial basis function network, wherein the reviewer radial basis function network is used for approximating a long-term performance index of the fractional order sliding mode error, and the weight learning rule is used for iteratively obtaining the network weight of the radial basis function network;
constructing an actor radial basis function network of the ith unmanned aerial vehicle and determining a weight learning rule of the actor radial basis function network, wherein the network weight of the actor radial basis function network is obtained by combining the evaluator radial basis function network, and the actor radial basis function network is used for approaching the unknown nonlinear item;
updating the network weights of the evaluator radial basis function network and the actor radial basis function network by using a reinforcement learning mode and combining the weight learning rule of the evaluator radial basis function network and the actor radial basis function network to obtain the evaluator radial basis function network and the actor radial basis function network after reinforcement learning;
the method for estimating the unknown nonlinear term by using the radial basis function neural network to obtain a nonlinear term estimation value comprises the following steps:
and estimating the unknown nonlinear term by using the commentator radial basis function network and the actor radial basis function network after reinforcement learning to obtain the nonlinear term estimated value.
The construction of the commentator radial basis function network of the ith unmanned aerial vehicle and the determination of the weight learning rule of the commentator radial basis function network comprise the following steps:
defining a long term performance index of the fractional order slip plane:
wherein T is a positive time constant, 0 < sigma i < 1 as discount factor ζ i (S i (τ)＝[ζ i1 (S i1 (τ))，ζ i2 (S i2 (τ))，ζ i3 (S i3 (τ))] T Is a performance function;
and is also provided with
Wherein,a preset threshold value;
determining a historical moment performance index based on the long-term performance index to obtain:
wherein,
constructing the commentator radial basis function neural network for approximating the long-term performance index of the fractional order sliding mode error as follows:
wherein,is a bounded ideal weight matrix of the radial basis function neural network of the evaluator, < >>Is the minimum approximation error vector of the radial basis function neural network of the evaluator,/and>is a bounded Gaussian basis function vector;
and the weight learning rule of the radial basis function neural network of the evaluator is as follows:
wherein,for said->Estimate of (K), K 11 ＞0，κ 12 ＞0；
The constructing the actor radial basis function network of the ith unmanned aerial vehicle and determining the weight learning rule of the actor radial basis function network comprise the following steps:
wherein,is a bounded ideal weight matrix of an actor radial basis function neural network, epsilon ia Is the minimum approximation error and,is a bounded gaussian basis function vector;
and determining a weight learning rule of the actor radial basis neural network based on the commentator radial basis neural network, wherein the weight learning rule is as follows:
wherein,namely, the estimated value kappa of the long-term performance index estimated by the radial basis function neural network of the evaluator 21 ＞0，κ 22 ＞0。
The estimating the unknown gain by using the Nussbaum function to obtain a gain estimation value includes:
constructing a Nussbaum function to obtain:
wherein θ i ＝[θ i1 ，θ i2 ，θ i3 ] T ；
Constructing a smooth function θ i The smoothing function theta i The differential amounts of (2) are:
smoothing the function theta i Bringing the Nussbaum function to obtain the gain estimate.
The constructing a kinetic model of the ith unmanned aerial vehicle considering faults comprises the following steps:
constructing a basic dynamics model of the ith unmanned aerial vehicle;
constructing an actuator fault model u of the ith unmanned aerial vehicle i ＝Λ i u i0 +b if Wherein, lambda i ＝diag{Λ i1 ，Λ i2 ，Λ i3 The matrix of unknown gains, b if ＝[b if1 ，b if2 ，b if3 ] T Is an actuation bias vector;
and introducing the actuator fault model into the basic dynamics model to obtain the dynamics model considering the fault.
The basic dynamics model of the ith unmanned aerial vehicle is constructed in the following manner:
wherein, (x) i ，y i ，h i ) Representing displacement distances of the ith unmanned aerial vehicle in three directions in three-dimensional space, V i Representing the flight rate, gamma i Represent the flying course angle, χ i Representing the flying pitch angle;
wherein g is gravity acceleration, u i ＝[u i1 ，u i2 ，u i3 ] T Respectively represents forward acceleration, yaw acceleration and pitch acceleration, d i ＝[d i1 ，d i2 ，d i3 ] T Is a disturbance vector;
the step of introducing the actuator fault model into the basic dynamics model to obtain the dynamics model considering the fault comprises the following steps:
transforming the basic dynamics model to obtain:
wherein,
carrying the actuator fault model into the transformed basic dynamics model to obtain the dynamics model considering the actuator fault
The beneficial technical effects of the application are as follows:
in the embodiment of the application, a preset performance function is utilized to apply specified constraint to the distributed tracking error of the unmanned aerial vehicle under the condition of the fault of the actuator, a radial basis function neural network and a Nussbaum function are utilized to estimate an unknown item caused by the fault of the actuator of the unmanned aerial vehicle, and the unknown item is compensated to establish a fault-tolerant control law under the fault of the actuator. The distributed tracking error can be restrained through the preset performance function, the position control of the networked fixed wing unmanned aerial vehicle is realized, and the fine adjustment of the fault tolerance performance can be realized through a fractional order control mode, so that the position of the fault unmanned aerial vehicle in the swarm unmanned aerial vehicle meets the formation control requirement.
Drawings
FIG. 1 is a block diagram of fault tolerant control provided by an exemplary row embodiment of the present application;
fig. 2 is a schematic communication topology of a swarm drone according to an example embodiment of the present application;
FIG. 3 is a schematic illustration of a flight trajectory of each fixed wing drone provided by one example row embodiment of the present application;
fig. 4 is a graph of synchronous tracking error for each fixed wing drone provided by one example row embodiment of the present application.
Fig. 5 is a schematic diagram illustrating a single fixed wing drone for characterizing a flight status of a system according to one embodiment of the present application.
FIG. 6 is a schematic diagram of various fixed wing drones for characterizing application control input signals provided by one example row embodiment of the present application.
Detailed Description
The following describes the embodiments of the present application further with reference to the drawings.
In order to solve the above problems, the embodiment of the present application provides a fault-tolerant control method for a failed unmanned aerial vehicle in a swarm unmanned aerial vehicle, where the method is described by taking computer equipment as an example. The method comprises the following steps:
step S1, a dynamics model of the ith unmanned aerial vehicle, which considers the failure of the actuator, is constructed, wherein the dynamics model of the failure of the actuator comprises an unknown nonlinear term and an unknown gain, and i is a positive integer.
In order to realize fault-tolerant control of the ith unmanned aerial vehicle with the fault of the executor in the swarm unmanned aerial vehicle, the application designs a fault-tolerant control method, so that the movement of the ith unmanned aerial vehicle with the fault of the executor can meet the formation control requirement. And the fault-tolerant control is realized by constructing a fault-tolerant control law, and in the process, the relationship between the position of the unmanned aerial vehicle and the control signal and the relationship between the unmanned aerial vehicle and the faults are required to be determined. The relation among the three can be determined according to a dynamics model of the unmanned aerial vehicle in consideration of the failure of the actuator, so that a dynamics model of the unmanned aerial vehicle in consideration of the failure of the actuator is firstly constructed, and the model can be obtained by introducing the failure model of the actuator into a basic dynamics model of the unmanned aerial vehicle in the ith frame.
Step S2, constructing a conversion error according to the distributed tracking error of the ith unmanned aerial vehicle and a preset performance function, wherein the preset performance function is used for restraining the distributed tracking error, the conversion error is an error for converting the distributed tracking error into a preset performance deviation, and the distributed tracking error is a relative distance error between the ith unmanned aerial vehicle and other unmanned aerial vehicles in the bee colony unmanned aerial vehicle.
The swarm unmanned aerial vehicle comprises a leading unmanned aerial vehicle and a following unmanned aerial vehicle, and the distributed tracking error comprises a relative distance error between the unmanned aerial vehicle and the leading unmanned aerial vehicle and/or the following unmanned aerial vehicle. Under the condition that the ith unmanned aerial vehicle can access the leading unmanned aerial vehicle (namely, communication connection is established between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle), the distributed tracking errors of the ith unmanned aerial vehicle respectively comprise relative distance errors between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle and between the ith unmanned aerial vehicle and the following unmanned aerial vehicle. It should be noted that the ith unmanned aerial vehicle is a following unmanned aerial vehicle.
In order to realize the constraint of the unmanned aerial vehicle distributed tracking error, namely, in the error range appointed by a user, a preset performance function is introduced. In order to facilitate the design of fault-tolerant control, error constraint corresponding to the distributed tracking error is converted to obtain conversion error, and the problem of finite time error constraint control when the unmanned aerial vehicle actuator fails is converted into a problem of consistent and bounded control for the failure.
And step S3, constructing a fractional sliding mode error based on the conversion error, wherein the fractional sliding mode error converges in a limited time.
In one possible embodiment, the conversion error is used to construct a fractional slip-mode error, i.e., a fractional slip-mode surface. According to the application, the design of fractional order sliding mode errors enables the movement position of the ith unmanned aerial vehicle to still maintain formation control requirements.
The fractional order sliding mode error can enable the conversion error to be converged in a limited time, so that the corresponding distributed tracking error meets the control requirement.
And S4, estimating an unknown nonlinear term by using a radial basis function neural network to obtain a nonlinear term estimated value, and estimating an unknown gain by using a Nussbaum function to obtain a gain estimated value.
In order to process an unknown nonlinear term and an unknown gain caused by unmanned aerial vehicle actuator faults, in the embodiment of the application, the unknown nonlinear term is estimated by utilizing the approximation characteristic of a radial basis function neural network so as to compensate the unknown nonlinear term; and the unknown gain is estimated by using a Nussbaum function, so that the compensation of the unknown gain is realized. When the fault-tolerant control law is constructed later, the unknown items caused by faults can be compensated by using the nonlinear estimated value obtained by estimating the radial basis function neural network and the gain estimated value obtained by estimating the Nussbaum function, so that the fault-tolerant control under the fault condition is realized. And the radial basis neural network is trained based on the comment-action reinforcement learning structure.
And S5, constructing a fault-tolerant control law by combining a dynamics model considering the fault of the actuator, a fractional order sliding mode error, a nonlinear item estimated value and a gain estimated value.
When the fault-tolerant control law is constructed, the relation between the position of the ith unmanned aerial vehicle and the fault and the control signal can be established by combining a dynamic model considering the fault of the actuator, and the fault-tolerant control law is obtained by combining fractional order sliding mode error reasoning according to the established relation. And in the process, the nonlinear estimation value and the gain estimation value are utilized to compensate the unknown nonlinear item and the unknown gain. Under the condition that the error-tolerant control law enables the fractional order sliding mode error to be converged in a limited time, the control signals corresponding to the error-tolerant control law can control the actual movement position of the unmanned aerial vehicle to meet the formation requirement.
And S6, determining a fault-tolerant control signal of the ith unmanned aerial vehicle based on the fault-tolerant control law, and controlling the ith unmanned aerial vehicle according to the fault-tolerant control signal, wherein the fault-tolerant control signal is used for controlling the ith unmanned aerial vehicle to meet formation control requirements.
In the flight process of the actual swarm unmanned aerial vehicle, the control system can determine a fault-tolerant control signal for the ith unmanned aerial vehicle according to a fault-tolerant control law, so that the position of the ith unmanned aerial vehicle still maintains the formation requirement under the condition of fault.
In the embodiment of the application, a preset performance function is utilized to apply specified constraint to the distributed tracking error of the unmanned aerial vehicle under the condition of the fault of the actuator, a radial basis function neural network and a Nussbaum function are utilized to estimate an unknown item caused by the fault of the actuator of the unmanned aerial vehicle, and the unknown item is compensated to establish a fault-tolerant control law under the fault of the actuator. The distributed tracking error can be restrained through the preset performance function, the position control of the networked fixed wing unmanned aerial vehicle is realized, and the fine adjustment of the fault tolerance performance can be realized through a fractional order control mode, so that the position of the fault unmanned aerial vehicle in the swarm unmanned aerial vehicle meets the formation control requirement.
Firstly, a dynamic model of the ith unmanned aerial vehicle considering faults is established, and the process comprises the following steps:
(1) And constructing a basic dynamics model of the ith unmanned aerial vehicle.
The method for constructing the basic dynamics model of the ith unmanned aerial vehicle is as follows:
wherein i=1,..n represents the i-th unmanned aerial vehicle, (x) i ，y i ，h i ) Representing displacement distances of the ith unmanned aerial vehicle in three directions in three-dimensional space, V i Representing the flight rate, gamma i Represent the flying course angle, χ i Representing the flying pitch angle;
and is also provided with
Wherein g is gravity acceleration, u i ＝[u i1 ，u i2 ，u i3 ] T Respectively represents forward acceleration, yaw acceleration and pitch acceleration, d i ＝[d i1 ，d i2 ，d i3 ] T Is a disturbance vector.
(2) And constructing an actuator fault model of the ith unmanned aerial vehicle.
The method for designing the unmanned aerial vehicle actuator efficiency loss and bias fault actuator fault model comprises the following steps:
u i ＝Λ i u i0 +b if
wherein, lambda i ＝diag{Λ i1 ，Λ i2 ，Λ i3 The matrix of unknown gains, Λ iv ∈(0，1]，b if ＝[b if1 ，b if2 ，b if3 ] T Is an actuation bias matrix, u i0 ＝[u i01 ，u i02 ，u i03 ] T Is the control input vector corresponding to the control signal.
(3) And introducing the actuator fault model into the basic dynamics model to obtain a dynamics model considering faults.
Firstly, transforming the basic dynamics model to obtain:
wherein q i ＝[x i ，y i ，h i ] T Is the position vector of the ith unmanned aerial vehicle, and
carrying the actuator fault model into the transformed basic dynamics model to obtain a dynamics model considering the actuator fault
In the above embodiment, in order to constrain the distributed tracking error of the following unmanned aerial vehicle relative to the adjacent unmanned aerial vehicle, a finite time preset performance function is introduced, and the distributed tracking error is converted into a conversion error by combining the transient and steady state requirements specified by a user, so as to establish a fractional order sliding mode surface, and the process of establishing the fractional order sliding mode surface comprises the following steps:
(1) A distributed tracking error is defined.
The distributed tracking error of the ith unmanned aerial vehicle is as follows:
wherein e i ＝[e i1 ，e i2 ，e i3 ] T Is the distributed tracking error of the ith unmanned aerial vehicle, a ij Is a weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle, b i Is the weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle, N i Q is the total number of following unmanned aerial vehicles in the bee colony unmanned aerial vehicle i ＝[x i ，y i ，h i ] T Is the position vector, q of the ith unmanned aerial vehicle 0 ＝[x 0 ，y 0 ，h 0 ] T Is the position vector, delta of the leading unmanned aerial vehicle ij ＝[δ ij1 ，δ ij2 ，δ ij3 ] T ＝δ i -δ j Is the expected relative distance, delta, between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle i ＝[δ i1 ，δ i2 ，δ i3 ] T Is a desired relative distance between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle; and a ij ＝a ji Is an undirected information flow, b when the ith unmanned aerial vehicle can access the leading unmanned aerial vehicle i If yes, b) if the leading unmanned aerial vehicle cannot be accessed i =0. And i e {0, 1..N }, j e {0, 1..N }.
(2) And designing a preset performance function.
The preset performance function is as follows:
wherein, xi v0 With xi vT Is positive parameter, a T Is a time constant. The preset performance function can be more than or equal to 0 and less than or equal to t < a T The period converges and t is equal to or greater than a T During which a minimum value ζ is reached vT Thereby ensuring distributed tracking error e i Is provided for the limited time convergence of (a). And the preset performance function is continuous over the interval 0, ++).
(3) And constructing a conversion error according to the distributed tracking error and a preset performance function of the ith unmanned aerial vehicle.
First, an error constraint is applied to the distributed tracking error in the following manner:
wherein k is v And (3) withV=1, 2,3, ζ for design parameters v Is a preset performance function;
to facilitate control design, the error constraint is subjected to equality conversion to obtain:
e iv ＝ξ v φ iv (η iv )
wherein phi is iv (η iv ) Is defined as:
inverting the equation obtained by the conversion to obtain a conversion value:
based on the conversion value, a conversion error is constructed as follows:
wherein E is iv Is the conversion error. By defining the conversion error, the finite time error constraint control in the presence of faults can be askedThe problem translates into a consistent bounded control problem for the fault.
For conversion error E iv The derivation can be obtained:
wherein, xi i ＝diag{Ξ i1 ，Ξ i2 ，Ξ i3 And (3)
And deriving to obtainCan be written as +.>Conversion error E iv The second derivative of (2) is as follows:
wherein E is i ＝[E i1 ，E i2 ，E i3 ] T ，ξ i ＝diag{ξ i1 ，ξ i2 ，ξ i3 }，/>
(4) The fractional order sliding mode error constructed based on the conversion error is as follows:
wherein lambda is 11 ，λ 12 Lambda of 13 Is a positive parameter of design, D * For fractional calculus symbol, aE (0, 1) is fractional calculus operator, S i ＝[s i1 ，s i2 ，s i3 ] T And sig a (·)＝|·| a sign(·)。
Then, a fault-tolerant control law can be constructed by using a dynamics model considering the fault of the actuator, fractional order sliding mode errors, nonlinear item estimated values and gain estimated values, and the fault-tolerant control law comprises the following processes:
the second derivative of the distributed tracking error is:
dynamics model in combination with consideration of actuator faultsAnd a conversion error, determining a second derivative of the conversion error as:
and combining the second derivative of the conversion error to obtain a differential equation of the fractional order sliding mode error, wherein the differential equation is as follows:
can be abbreviated as:
wherein F is i Is the unknown nonlinear term.
After a differential equation of fractional order sliding mode errors is obtained, a fault-tolerant control law can be obtained by back-pushing.
First, the nonlinear term estimation value and the gain estimation value are utilized to calculate the unknown nonlinear term F in the fractional order sliding mode error differential equation i Unknown gain matrix Λ i And (3) compensating, wherein a fault-tolerant control law can be constructed based on a differential equation of the compensated fractional order sliding mode error, and the fault-tolerant control law is obtained as follows:
wherein,is lambda i The corresponding gain estimate value is used to determine,is F i Corresponding nonlinear term estimate, K 2 And > 0 is a positive diagonal matrix.
The unknown nonlinear term F i Based on radial basis function neural network estimation. Wherein, radial basis neural network is based on comment-action structure reinforcement learning, and the process includes:
(1) Constructing a reviewer radial basis function neural network of the ith unmanned aerial vehicle, determining a weight learning rule of the reviewer radial basis function neural network, wherein the reviewer radial basis function neural network is used for approximating a long-term performance index of a fractional order sliding mode error, and the weight learning rule is used for iteratively obtaining the network weight of the radial basis function neural network.
First, defining a long-term performance index of fractional order sliding mode error:
wherein, gamma i (t)＝[γ i1 (t)，γ i2 (t)，γ i3 (t)] T T is a positive time constant, 0 < sigma i < 1 as discount factor ζ i (S i (τ))＝[ζ i1 (S i1 (τ))，ζ i2 (S i2 (τ))，ζ i3 (S i3 (τ))] T Is a performance function;
and is also provided with
Wherein,a preset threshold value;
determining a historical moment performance index based on the long-term performance index to obtain:
wherein,
because the long-term performance index contains future information, a reviewer radial basis neural network for approximating the long-term performance index of the fractional order sliding mode error is constructed as follows:
wherein,is a bounded ideal weight matrix of a reviewer radial basis function neural network,>is the minimum approximation error vector of the commentator radial basis function neural network, < >>Is a bounded Gaussian basis functionA number vector;
and the weight learning rule of the radial basis function neural network of the commentator is as follows:
wherein,is->Estimate of (K), K 11 ＞0，κ 12 ＞0。
(2) Constructing an actor radial basis function network of the ith unmanned aerial vehicle, determining a weight learning rule of the actor radial basis function network, and combining the network weight of the actor radial basis function network with a reviewer radial basis function network, wherein the actor radial basis function network is used for approaching an unknown nonlinear item.
An actor radial basis neural network for approximating an unknown nonlinear term is constructed as follows:
wherein,is a bounded ideal weight matrix of an actor radial basis function neural network, epsilon ia Is the minimum approximation error and,is a bounded gaussian basis function vector;
based on the reviewer radial basis function neural network, determining a weight learning rule of the actor radial basis function neural network as follows:
wherein,namely, the estimated value of the long-term performance index estimated by the radial basis function neural network of the commentator, kappa 21 ＞0，κ 22 ＞0。
(3) And updating the network weights of the commentator radial basis function network and the actor radial basis function network by combining the weight learning rule of the commentator radial basis function network and the actor radial basis function network in a reinforcement learning mode to obtain the commentator radial basis function network and the actor radial basis function network after reinforcement learning.
After reinforcement learning, the unknown nonlinear term can be estimated by using the commentator radial basis function network and the actor radial basis function network after reinforcement learning to obtain a nonlinear term estimated value.
In one possible implementation, the actor radial basis function network estimate may be utilized to derive a non-linear estimate
In this embodiment, estimating the unknown gain by using Nussbaum function to obtain a gain estimation value includes:
constructing a Nussbaum function to obtain:
wherein θ i ＝[θ i1 ，θ i2 ，θ i3 ] T 。
Constructing a smooth function θ i The smoothing function theta i The differential amounts of (2) are:
when the following condition is satisfied, the following is satisfied in the interval [0, t 1 ) The smooth function L (t) > 0, θ (t) and N (θ) are consistently bounded:
wherein h is 1 >0,h 2 >0,∈ i >0。
Will smooth the function theta i The gain estimate is obtained by taking into account the Nussbaum function.
Schematically, as shown in fig. 1, a control block diagram for an ith drone is shown. The communication network can determine the distributed tracking error of the ith unmanned aerial vehicle according to the position information of the ith unmanned aerial vehicle, the position information of the following unmanned aerial vehicle adjacent to the ith unmanned aerial vehicle and the position information of the leading unmanned aerial vehicle, and then the distributed tracking error is combined with a preset performance function to obtain a conversion error, so that a fractional order sliding mode surface can be constructed. Based on the critic radial basis function neural network, the long-term performance index can be estimated and obtained, and the actor radial basis function neural network is combined with the long-term performance index to perform unknown nonlinear item estimation, so that a nonlinear item estimated value is obtained; and performing unknown gain estimation by using a Nussbaum function to obtain a gain estimation value. And constructing a fault-tolerant control law based on the fractional order sliding mode surface, the nonlinear item estimated value and the gain estimated value, and controlling the ith unmanned aerial vehicle under the fault of the actuator.
In order to verify the effectiveness of the application on the fault-tolerant control of the swarm unmanned aerial vehicle, a simulation example is constructed. The method is applicable to the swarm unmanned aerial vehicle, the swarm unmanned aerial vehicle comprises a leader unmanned aerial vehicle #0 and N fixed wing unmanned aerial vehicles which fly along with the leader unmanned aerial vehicle #0, and in the method, in fig. 1, for example, 4 fixed wing unmanned aerial vehicles #1,2,3 and 4 are included, a communication topology is established between unmanned aerial vehicles in the swarm unmanned aerial vehicle, and the method comprises the following steps: communication connection is established between leader unmanned aerial vehicle #0 and the fixed wing unmanned aerial vehicle to establish communication connection between the fixed wing unmanned aerial vehicle. Each drone may establish a communication connection with all other drones, or only some of the other drones. In one illustrative example, a two-way communication connection is established between the fixed wing drones, while a one-way communication connection is established between the leader drone #0 and the fixed wing drone to transfer information from the leader drone to the fixed wing drone. As shown in the communication topology diagram of fig. 1, a leader unmanned aerial vehicle #0, a fixed wing unmanned aerial vehicle #1 and a fixed wing unmanned aerial vehicle #4 respectively establish unidirectional communication connection, a fixed wing unmanned aerial vehicle #3 respectively establishes bidirectional communication connection with the fixed wing unmanned aerial vehicles #1,2 and 4, and the fixed wing unmanned aerial vehicle #1 also establishes bidirectional communication connection with the fixed wing unmanned aerial vehicles #2 and 4.
When the swarm drones fly in formation under the proposed controller, the corresponding weighted adjacency matrix is designed as:
wherein,correspond to a ij ，/>Correspond to b i 。
The parameters are selected as follows: a=0.3, k 2 ＝diag{60，60，13.5}，κ 11 ＝96.2，κ 12 ＝2.3，κ 21 ＝96.7，κ 22 ＝0.7，λ 11 ＝2，λ 12 ＝3.5，λ 13 ＝1，ξ 10 ＝0.6，ξ 20 ＝0.6，ξ 30 ＝0.6，ξ 1T ＝0.2，ξ 2T ＝0.2，ξ 3T ＝0.2，v＝1，2，3；T＝0.6，C ζ1 ＝0.3，C ζ2 ＝0.3，C ζ3 ＝0.3，a T ＝10。
The initial position of the unmanned aerial vehicle is q 0 (0)＝[30t+200，10sin(0.2t)，15sin(0.3t)+1000] T ，q 1 (0)＝[-2，298，998] T ，q 2 (0)＝[-298，402，1203] T ，q 3 (0)＝[-302，-402，902] T ，q 4 (0)＝[0，-300，1000] T . The initial state is set as: v (V) i (0)＝30，χ i (0)＝0，γ i (0) =0. The offset of the unmanned aerial vehicle is delta 1 ＝[-200，300，0] T ，δ 4 ＝[-200，-300，0] T ，δ 12 ＝[300，-100，-200] T ，δ 13 ＝[300，700，100] T ，δ 14 ＝[0，600，0] T ，δ 21 ＝[-300，100，200] T ，δ 23 ＝[0，800，300] T ，δ 31 ＝[-300，-700，-100] T ，δ 32 ＝[0，-800，-300] T ，δ 34 ＝[-300，-100，-100] T ，δ 41 ＝[0，-600，0] T ，δ 43 ＝[300，100，100] T 。
Assuming that unmanned aerial vehicle #1 encounters speed, course angle and course angle channel faults at 10s, unmanned aerial vehicle #2 encounters speed channel faults at 20s, unmanned aerial vehicle #3 encounters course angle channel faults at 30s, and the fault signal is selected as Λ 11 ＝Λ 12 ＝Λ 13 ＝1，b 1f1 ＝b 1f2 ＝b 1f3 ＝0，t＜10；Λ 11 ＝0.3e -2(t-10) +0.7；Λ 12 ＝0.2e -2(t-10) +0.8；Λ 13 ＝0.25e -2(t-10) +0.75；b 1f1 ＝6e -2(t-10) -6；b 1f2 ＝2e -2(t-10) -2；b 1f3 ＝5e -2(t-10) -5；t≥10；Λ 21 ＝Λ 22 ＝Λ 23 ＝1，b 2f1 ＝b 2f2 ＝b 2f3 ＝0，t＜20；Λ 21 ＝0.25e -2(t-20) +0.75，Λ 22 ＝1，Λ 23 ＝1，b 2f1 ＝5e -2(t-20) -5，b 2f2 ＝0，t≥20；Λ 31 ＝Λ 32 ＝Λ 33 ＝1；b 3f1 ＝b 3f2 ＝b 3f3 ＝0，t＜30；Λ 31 ＝1，Λ 32 ＝0.2e -2(t-30) +0.8，Λ 33 ＝1，b 3f1 ＝0，b 3f1 ＝3e -2(t-30) -3，b 3f1 =0, t+.30. Furthermore, all following unmanned aerial vehicles are suddenly subjected to d at 5s t1 ＝1.5，d i2 ＝1.2，d i3 External interference of=2. Fig. 3 shows the flight trajectory of the drone, it can be seen that even if the disturbance is injected at 5s, the drone #1-3 encounters a fault at 10s,20s,30s, respectively, following the drone #1-4, the leading drone can still be tracked with a pre-designed offset.
Fig. 4 shows the distributed tracking errors of four following drones. To simulate the actual situation in engineering, the initial distribution tracking error is selected to be a non-zero value, including: e, e 11 (0)＝-6m，e 21 (0)＝6.8m，e 31 (0)＝-4.2m，e 41 (0)＝1.8m，e 12 (0)＝-6m，e 22 (0)＝6.8m，e 32 (0)＝-4.2m，e 42 (0)＝1.8m，e 13 (0)＝-10.1m，e 23 (0)＝5.3m，e 33 (0)＝3.4m，e 43 (0) -0.2m. It can be observed from the figure that if the following unmanned aerial vehicle #1-3 suffers from 10s,20s and 30s failure, respectively, if the disturbance is injected at 5s, a slight performance degradation and variation is caused. However, under the fault-tolerant control method provided by the application, the system change tends to be stable, and the error is strictly limited to the range designated by the user within the limited time of 10sAnd (3) inner part.
Fig. 5 shows the speed, heading angle and pitch angle, respectively. Tracking errors due to non-zero initial distribution are shown in fig. 4. V (V) i ，x i And gamma i I=1, 2,3,4, there will be a large change in the beginning to reduce the distributed tracking error. All states are stable and bounded even if the following drone encounters disturbances and faults.
The control inputs are shown in fig. 6. To cope with disturbances at 5s, the fault tolerant controller adjusts rapidlyControl input signal u i01 ，u i02 And u i03 And the stability of unmanned aerial vehicle formation is guaranteed. When following unmanned aerial vehicle #1 breaks down at 10s, u is changed rapidly 101 ，u 102 And u 103 To stabilize the failed drone #1 and the neighbor drone. Furthermore, as can be seen from FIG. 6, u 402 Also mutated at 10s, due to the information exchange between unmanned aerial vehicle #1 and unmanned aerial vehicle # 4. When the unmanned aerial vehicle #2 speed channel fails at 20s, the control input signal u is dynamically adjusted 201 Guarantee e i1 Is a convergence of (a). In addition, since the three heading control channels of each fixed-wing unmanned aerial vehicle are highly coupled, the heading control channels input the signal u 203 Slightly varied at 20 s. A similar phenomenon can also be found when unmanned aerial vehicle #3 fails at 30 s.
By adopting the control response mode, the distributed tracking error converges to interference and faults, and all following unmanned aerial vehicles can track the leading unmanned aerial vehicle by a pre-designed offset.
The above is only a preferred embodiment of the present application, and the present application is not limited to the above examples. It is to be understood that other modifications and variations which may be directly derived or contemplated by those skilled in the art without departing from the spirit and concepts of the present application are deemed to be included within the scope of the present application.
Claims (10)
1. A fault tolerant control method for a failed drone in a swarm drone, the method comprising:
constructing a dynamics model of an ith unmanned aerial vehicle, which considers the failure of an actuator and comprises an unknown nonlinear term and an unknown gain, wherein i is a positive integer;
constructing a conversion error according to the distributed tracking error of the ith unmanned aerial vehicle and a preset performance function, wherein the preset performance function is used for constraining the distributed tracking error, the conversion error is an error for converting the distributed tracking error into a preset performance deviation, and the distributed tracking error is a relative distance error between the ith unmanned aerial vehicle and other unmanned aerial vehicles in the swarm unmanned aerial vehicle;
constructing a fractional order sliding mode error based on the conversion error, wherein the fractional order sliding mode error converges in a limited time;
estimating the unknown nonlinear term by using a radial basis function neural network to obtain a nonlinear term estimation value, and estimating the unknown gain by using a Nussbaum function to obtain a gain estimation value;
combining the dynamics model considering the actuator fault, the fractional order sliding mode error, the nonlinear item estimated value and the gain estimated value to construct a fault-tolerant control law;
and determining a fault-tolerant control signal of the ith unmanned aerial vehicle based on the fault-tolerant control law and controlling the ith unmanned aerial vehicle according to the fault-tolerant control signal, wherein the fault-tolerant control signal is used for controlling the ith unmanned aerial vehicle to meet formation control requirements.
2. The method of claim 1, wherein the distributed tracking error of the ith drone is as follows:
wherein e i ＝[e i1 ，e i2 ，e i3 ] T Is the distributed tracking error of the ith unmanned aerial vehicle, a ij Is a weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle, b i Is the weight coefficient corresponding to the distance error between the ith unmanned aerial vehicle and the leading unmanned aerial vehicle, N i Q is the total number of following unmanned aerial vehicles in the bee colony unmanned aerial vehicle j ＝[x i ，y i ，h i ] T Is the position vector, q of the ith unmanned aerial vehicle 0 ＝|x 0 ，y 0 ，h 0 ] T Is the position vector, delta of the leading unmanned aerial vehicle ij Is the expected relative distance, delta, between the ith unmanned aerial vehicle and the jth following unmanned aerial vehicle i Is saidA desired relative distance between an ith unmanned aerial vehicle and the lead unmanned aerial vehicle;
the preset performance function is as follows:
wherein, xi v0 With xi vT Is positive parameter, a T Is a time constant.
3. The method of claim 2, wherein constructing the conversion error based on the distributed tracking error of the ith drone and the preset performance function comprises:
and applying error constraint to the distributed tracking error in the following manner:
wherein,and->V=1, 2,3, ζ for design parameters v The preset performance function is set;
and carrying out equation conversion on the error constraint to obtain:
e iv ＝ξ v φ iv (η iv )
wherein phi is iv (η iv ) Is defined as:
inverting the equation obtained by the conversion to obtain a conversion value:
based on the conversion value, the conversion error is constructed as follows:
wherein E is iv Is the conversion error.
4. A method according to claim 3, wherein the fractional order sliding mode error constructed based on the conversion error is:
wherein lambda is 11 ，λ 12 Lambda of 13 Is a positive parameter of design, D * For fractional calculus symbols, a is the fractional calculus operator and a e (0, 1).
5. The method of claim 4, wherein the dynamic model that accounts for actuator failure isWherein, lambda i Is an unknown gain matrix g i Is a known gain matrix, f i Is a compound item related to dynamics, b if Is the actuation deviation vector, u i0 ＝[u i01 ，u i02 ，u i03 ] T Is the control input vector corresponding to the control signal;
the combining the dynamics model considering the actuator fault, the fractional order sliding mode error, the nonlinear term estimation value and the gain estimation value to construct a fault-tolerant control law comprises the following steps:
combining the dynamics model considering the actuator fault and the conversion error, determining the second derivative of the conversion error as follows:
wherein, xi i ＝diag{Ξ i1 ，Ξ i2 ，Ξ i3 And (3)
And combining the second derivative of the conversion error to obtain a differential equation of the fractional order sliding mode error, wherein the differential equation is as follows:
wherein F is i For the unknown non-linear term,
and compensating the unknown nonlinear term and the unknown gain matrix in the fractional order sliding mode error differential equation by using the nonlinear term estimation value and the gain estimation value, and constructing the fault-tolerant control law based on the differential equation of the fractional order sliding mode error, wherein the fault-tolerant control law is as follows:
wherein,is said Λ i Corresponding said gain estimates, < >>Is said F i Corresponding to the nonlinear term estimation value, K 2 And > 0 is a positive diagonal matrix.
6. The method of claim 1, wherein the radial basis function neural network is derived based on comment-action structure reinforcement learning, the method further comprising:
constructing a reviewer radial basis function network of the ith unmanned aerial vehicle and determining a weight learning rule of the reviewer radial basis function network, wherein the reviewer radial basis function network is used for approximating a long-term performance index of the fractional order sliding mode error, and the weight learning rule is used for iteratively obtaining the network weight of the radial basis function network;
constructing an actor radial basis function network of the ith unmanned aerial vehicle and determining a weight learning rule of the actor radial basis function network, wherein the network weight of the actor radial basis function network is obtained by combining the evaluator radial basis function network, and the actor radial basis function network is used for approaching the unknown nonlinear item;
updating the network weights of the evaluator radial basis function network and the actor radial basis function network by using a reinforcement learning mode and combining the weight learning rule of the evaluator radial basis function network and the actor radial basis function network to obtain the evaluator radial basis function network and the actor radial basis function network after reinforcement learning;
the method for estimating the unknown nonlinear term by using the radial basis function neural network to obtain a nonlinear term estimation value comprises the following steps:
and estimating the unknown nonlinear term by using the commentator radial basis function network and the actor radial basis function network after reinforcement learning to obtain the nonlinear term estimated value.
7. The method of claim 6, wherein the constructing the reviewer radial basis function network of the ith unmanned aerial vehicle and determining the weight learning rule of the reviewer radial basis function network comprise:
defining a long-term performance index of the fractional order sliding mode error:
wherein T is a positive time constant, 0 < sigma i < 1 as discount factor ζ i (S i (τ))＝[ζ i1 (S i1 (τ))，ζ i2 (S i2 (τ))，ζ i3 (S i3 (τ))] T Is a performance function;
and is also provided with
Wherein,a preset threshold value;
determining a historical moment performance index based on the long-term performance index to obtain:
wherein,
constructing the commentator radial basis function neural network for approximating the long-term performance index of the fractional order sliding mode error as follows:
wherein,is a bounded ideal weight matrix of the radial basis function neural network of the evaluator, < >>Is the minimum approximation error vector of the radial basis function neural network of the evaluator,/and>is a bounded Gaussian basis function vector;
and the weight learning rule of the radial basis function neural network of the evaluator is as follows:
wherein,for said->Estimate of (K), K 11 ＞0，κ 12 ＞0；
The constructing the actor radial basis function network of the ith unmanned aerial vehicle and determining the weight learning rule of the actor radial basis function network comprise the following steps:
constructing the actor radial basis neural network for approximating the unknown nonlinear term as follows:
wherein,is a bounded ideal weight matrix of an actor radial basis function neural network, epsilon ia Is the minimum approximation error vector and,is a bounded gaussian basis function vector;
and determining a weight learning rule of the actor radial basis neural network based on the commentator radial basis neural network, wherein the weight learning rule is as follows:
wherein,namely, the estimated value kappa of the long-term performance index estimated by the radial basis function neural network of the evaluator 21 ＞0，κ 22 ＞0。
8. The method of claim 5, wherein estimating the unknown gain using a Nussbaum function results in a gain estimate, comprising:
constructing a Nussbaum function to obtain:
wherein θ i ＝[θ i1 ，θ i2 ，θ i3 ] T ；
Constructing a smooth function θ i The smoothing function theta i The differential amounts of (2) are:
smoothing the function theta i Carrying into the Nussbaum functionTo the gain estimate.
9. The method according to any one of claims 1 to 8, wherein said constructing a failure-considered kinetic model of the ith unmanned aerial vehicle comprises:
constructing a basic dynamics model of the ith unmanned aerial vehicle;
constructing an actuator fault model u of the ith unmanned aerial vehicle i ＝Λ i u i0 +b if Wherein, lambda i ＝diag{Λ i1 ，Λ i2 ，Λ i3 The matrix of unknown gains, b if ＝[b if1 ，b if2 ，b if3 ] T Is an actuation bias matrix;
and introducing the actuator fault model into the basic dynamics model to obtain the dynamics model considering the fault.
10. The method of claim 9, wherein the basic kinetic model of the ith unmanned aerial vehicle is constructed in the following manner:
wherein, (x) i ，y i ，h i ) Representing displacement distances of the ith unmanned aerial vehicle in three directions in three-dimensional space, V i Representing the flight rate, gamma i Represent the flying course angle, χ i Representing the flying pitch angle;
wherein g is gravity acceleration, u i ＝[u i1 ，u i2 ，u i3 ] T Respectively represents forward acceleration, yaw acceleration and pitch acceleration, d i ＝[d i1 ，d i2 ，d i3 ] T Is a disturbance vector;
the step of introducing the actuator fault model into the basic dynamics model to obtain the dynamics model considering the fault comprises the following steps:
transforming the basic dynamics model to obtain:
wherein,
carrying the actuator fault model into the transformed basic dynamics model to obtain the dynamics model considering the actuator fault
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311108504.0A CN117008637A (en) | 2023-08-30 | 2023-08-30 | Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311108504.0A CN117008637A (en) | 2023-08-30 | 2023-08-30 | Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117008637A true CN117008637A (en) | 2023-11-07 |
Family
ID=88572817
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311108504.0A Pending CN117008637A (en) | 2023-08-30 | 2023-08-30 | Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117008637A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113268084A (en) * | 2021-06-29 | 2021-08-17 | 南京工业大学 | Intelligent fault-tolerant control method for unmanned aerial vehicle formation |
| CN115903908A (en) * | 2022-12-14 | 2023-04-04 | 南京航空航天大学 | Fault-tolerant cooperative control method for swarm unmanned aerial vehicle based on rapid terminal sliding mode |
-
2023
- 2023-08-30 CN CN202311108504.0A patent/CN117008637A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113268084A (en) * | 2021-06-29 | 2021-08-17 | 南京工业大学 | Intelligent fault-tolerant control method for unmanned aerial vehicle formation |
| CN115903908A (en) * | 2022-12-14 | 2023-04-04 | 南京航空航天大学 | Fault-tolerant cooperative control method for swarm unmanned aerial vehicle based on rapid terminal sliding mode |
Non-Patent Citations (3)
| Title |
|---|
| ZIQUAN YU 等: "Distributed Adaptive Fault-Tolerant Time-Varying Formation Control of Unmanned Airships With Limited Communication Ranges Against Input Saturation for Smart City Observation", IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, vol. 33, no. 5, pages 1891 - 1904 * |
| ZIQUAN YU 等: "Reinforcement Learning-Based Fractional-Order Adaptive Fault-Tolerant Formation Control of Networked Fixed-Wing UAVs With Prescribed Performance", IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, pages 1 - 15 * |
| 徐斌彦;齐瑞云;姚雪莲;: "高超声速飞行器舵面故障Nussbaum增益自适应容错控制", 战术导弹技术, no. 04, pages 109 - 118 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yu et al. | Decentralized fractional-order backstepping fault-tolerant control of multi-UAVs against actuator faults and wind effects | |
| Yu et al. | Fault‐tolerant formation control of multiple UAVs in the presence of actuator faults | |
| Wu et al. | Disturbance observer based robust mixed H2/H∞ fuzzy tracking control for hypersonic vehicles | |
| CN110597061A (en) | Multi-agent fully-distributed active-disturbance-rejection time-varying formation control method | |
| CN113342037B (en) | Multi-rotor unmanned aerial vehicle time-varying formation control method and system with input saturation | |
| Yu et al. | Distributed adaptive fault-tolerant close formation flight control of multiple trailing fixed-wing UAVs | |
| CN113268064B (en) | Multi-mobile-robot cooperative formation control method considering communication time delay | |
| CN112305918A (en) | Multi-agent system sliding mode fault-tolerant consistency control algorithm under supercoiled observer | |
| Wu et al. | Event-triggered-based adaptive NN cooperative control of six-rotor UAVs with finite-time prescribed performance | |
| Wang et al. | Disturbance observer-based backstepping formation control of multiple quadrotors with asymmetric output error constraints | |
| CN113848887A (en) | Under-actuated unmanned ship trajectory tracking control method based on MLP method | |
| Xu et al. | Fuzzy adaptive nonlinear information fusion model predictive attitude control of unmanned rotorcrafts | |
| Xingling et al. | Prescribed fast tracking control for flexible air-breathing hypersonic vehicles: An event-triggered case | |
| Liu et al. | Antisaturation fixed-time attitude tracking control based low-computation learning for uncertain quadrotor UAVs with external disturbances | |
| Shao et al. | Robust discrete‐time fractional‐order control for an unmanned aerial vehicle based on disturbance observer | |
| Qian et al. | Sliding mode control‐based distributed fault tolerant tracking control for multiple unmanned aerial vehicles with input constraints and actuator faults | |
| Wang et al. | Adaptive fractional‐order nonsingular fast terminal sliding mode formation control of multiple quadrotor UAVs‐based distributed estimator | |
| Li et al. | Learning-observer-based adaptive tracking control of multiagent systems using compensation mechanism | |
| Cui et al. | Fixed‐time backstepping distributed cooperative control for multiple unmanned aerial vehicles | |
| CN116107339B (en) | Fault-tolerant cooperative control method for bee colony unmanned aerial vehicle under thrust loss fault | |
| CN117008637A (en) | Fault-tolerant control method for fault unmanned aerial vehicle in bee colony unmanned aerial vehicle | |
| CN110109357A (en) | For half global self-adaptation control method of non-standard non-linear aircraft | |
| Wang et al. | Fixed-time event-triggered sliding mode cooperative path-following control with prescribed performance for USVs based on lumped disturbance observer | |
| Abdessameud et al. | Velocity-free consensus algorithms for double-integrator dynamics with input saturations constraints | |
| Wang et al. | Observer‐based adaptive neural fault‐tolerant control for nonlinear systems with prescribed performance and input dead‐zone |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |