CN-121979243-A - Four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack

CN121979243ACN 121979243 ACN121979243 ACN 121979243ACN-121979243-A

Abstract

The invention discloses a four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack, which comprises the steps of defining a state error of an unmanned aerial vehicle formation system according to an unmanned aerial vehicle formation model, constructing a distributed controller of the formation system according to the state error, acquiring an error state equation according to the distributed controller of the formation system to construct a formation error state vector considering the influence of asynchronous packet loss and random UDP attack, constructing a double-threshold dynamic event trigger mechanism based on historical memory fusion to design the unmanned aerial vehicle formation control law, constructing a closed loop equation of the four-rotor unmanned aerial vehicle formation system according to the established unmanned aerial vehicle formation control law, the error state equation and the formation error state vector, and combining the distributed controller of the formation system to realize four-rotor unmanned aerial vehicle formation control suitable for asynchronous packet loss and random UDP attack. The method solves the problem that the existing method is difficult to realize the distributed formation control of the four-rotor unmanned aerial vehicle under the conditions of asynchronous packet loss and random UDP attack.

Inventors

YUE WEI
WANG BOYU
LIANG XIAO
ZHANG XIAOYONG

Assignees

大连海事大学

Dates

Publication Date: 20260505
Application Date: 20260128

Claims (6)

1. A four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack is characterized by comprising the following steps: s1, constructing a follower unmanned aerial vehicle model and a leader unmanned aerial vehicle model to establish an unmanned aerial vehicle formation model; S2, defining a state error of an unmanned aerial vehicle formation system according to an unmanned aerial vehicle formation model, constructing a formation system distributed controller according to the state error, and acquiring an error state equation according to the formation system distributed controller; S3, constructing a formation error state vector under the influence of asynchronous packet loss and random UDP attack based on the state error of the unmanned aerial vehicle formation system; s4, constructing a double-threshold dynamic event trigger mechanism based on history memory fusion, and designing an unmanned aerial vehicle formation control law according to the double-threshold dynamic event trigger mechanism; s5, constructing a closed loop equation of a four-rotor unmanned aerial vehicle formation system according to the established unmanned aerial vehicle formation control law, the error state equation and the formation error state vector; and according to a closed loop equation of the four-rotor unmanned aerial vehicle formation system, a distributed controller of the formation system is combined, so that the four-rotor unmanned aerial vehicle formation control suitable for asynchronous packet loss and random UDP attack is realized.
2. The four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack according to claim 1, wherein the unmanned aerial vehicle formation model established in S1 is: (1) wherein: representing a coefficient matrix and Indicating that unmanned aerial vehicle is in Damping coefficients in the triaxial directions; represent the first A position vector of each follower unmanned plane, and ; Respectively represent the first The position abscissa, ordinate and vertical coordinate of the individual follower unmanned aerial vehicle; represent the first A speed vector of each follower unmanned plane, and ; Respectively represent the first Along with individual follower drones Velocity components in the three axis directions; represent the first Personal follower unmanned plane is in External environment disturbance in triaxial directions and ; Represent the first Personal follower unmanned plane is in Control input signal in three axes ; Representing a transpose; Representation of Is a first order derivative of (a); Representing that a leader unmanned aerial vehicle is in Position vector in three axis directions and ; Representing a velocity vector of the leader drone and , Representing the damping coefficient of the leader unmanned aerial vehicle ; Indicating the number of follower drones.
3. The four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack according to claim 2, wherein the S2 specifically comprises the following steps: s21, defining a state error of an unmanned aerial vehicle formation system according to an unmanned aerial vehicle formation model as follows: (2) wherein: respectively representing a position error and a speed error; represent the first A desired distance between the individual follower drones and the leader drone; s22, constructing a formation system distributed controller according to the state error, wherein the formation system distributed controller comprises the following components: (3) wherein: Follower unmanned aerial vehicle capable of mutually exchanging state information Is a part of the unmanned aerial vehicle of the neighborhood, Representing a neighbor unmanned aerial vehicle set; Respectively represent the first The communication link weights between the individual follower unmanned aerial vehicle and its neighbor unmanned aerial vehicle and the leader unmanned aerial vehicle; Representing the controller gain; A control signal representing a distributed controller of the formation system; s23, deriving a state error derivative of the unmanned aerial vehicle formation system to obtain the state error derivative, wherein the state error derivative is as follows: (4) by defining error vectors The error state equation obtained according to the formation system distributed controller is as follows: (5) (6) wherein: Representing error vectors Is a first order derivative of (a); Representing a parameter matrix; representing the identity matrix.
4. The four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack according to claim 3, wherein the S3 specifically comprises the following steps: s31, defining a sampling period sequence for sampling unmanned aerial vehicle formation state information And define event trigger time sequence as And (2) and ; S32 when the first step Personal follower Unmanned Aerial Vehicle (UAV) and first Normal communication between unmanned aerial vehicle of individual follower, obtain the first based on unmanned aerial vehicle formation system's state error Input signal in unmanned aerial vehicle controller design And is also provided with The method comprises the following steps: (7) wherein: Representing the current sampling time; represent the first A desired distance vector of the individual follower unmanned aerial vehicle and the leader unmanned aerial vehicle; Representing the current sampling instant Corresponds to the first Position vectors and velocity vectors of the individual follower drones; Representing the current sampling instant A position vector and a speed vector corresponding to the leader unmanned aerial vehicle; S33, based on the step S32, respectively acquiring input signals under the influence of asynchronous packet loss and random UDP attack, wherein the method comprises the following steps: if the UDP attack is suffered, the method and the device are that No. H of personal follower unmanned aerial vehicle communication The state information of the individual follower unmanned aerial vehicle, namely the neighbor unmanned aerial vehicle is added with disturbance to obtain: (8) wherein: , respectively represent the input to the first at the current sampling time Interfered first of individual follower unmanned aerial vehicle Position vectors and velocity vectors of the individual follower unmanned aerial vehicles; Representing an attack intensity matrix and , Representing the last event trigger sampling time; representing event-triggered sampling moments The corresponding position vector and velocity vector; If packet loss occurs in network communication, the network communication is connected with the first node No. H of personal follower unmanned aerial vehicle communication The state information of the individual follower unmanned aerial vehicle, namely the neighbor unmanned aerial vehicle, is: (9) wherein: , representing the last sampling instant Is the first of (2) The state values of the follower unmanned aerial vehicle are position vectors and speed vectors; s34, defining two independent parameter variables obeying Bernoulli distribution And The method comprises the following steps: (10) wherein: Representing the probability of communication failure and the probability of UDP attack, respectively ; And according to the parameter variables And The method for confirming whether the communication is successful or not and whether the communication is lost or is attacked by UDP specifically comprises the following steps: When (when) When it indicates abnormal communication The time indicates that the communication is normal; if communication is abnormal and Indicating that unmanned aerial vehicle formation communication is attacked by UDP; if communication is abnormal and Indicating that the unmanned aerial vehicle formation communication is lost; and the expression of whether the communication is successful and packet loss and whether the communication is attacked by UDP is as follows: (11) S35, acquiring a system state vector of unmanned aerial vehicle formation based on S34 The method comprises the following steps: (12) wherein: Representation and the first No. H of personal follower unmanned aerial vehicle communication The state of the individual follower unmanned aerial vehicle is under the expression form after UDP attack; Neighbor unmanned aerial vehicle representing last-time communication, namely the first The state of the individual follower unmanned aerial vehicle; Representing the state of the neighbor unmanned aerial vehicle at the last event triggering sampling moment; S36, constructing a first algorithm based on the formula (8), the formula (9) and the formula (12) under the influence of asynchronous packet loss and random UDP attack No. H of personal follower unmanned aerial vehicle communication Formation error state vector for individual follower unmanned aerial vehicle The method comprises the following steps: (13) wherein: represent the first Corresponding parameter variable obeying Bernoulli distribution of individual follower unmanned aerial vehicle , ; Indicating the last time The following unmanned aerial vehicle inputs to the first Input signals to the unmanned aerial vehicle controller; represents the last event-triggered sampling time, the first The following unmanned aerial vehicle inputs to the first Input signals to the unmanned aerial vehicle controller; Representation of Manifestation after UDP attack.
5. The four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack according to claim 4, wherein the S4 specifically comprises the following steps: s41, constructing a dual-threshold dynamic event trigger mechanism based on history memory fusion, wherein the dual-threshold dynamic event trigger mechanism comprises the following steps: (14) (15) (16) wherein: representing the current trigger time of the dual-threshold dynamic event; Representing the last event trigger sampling time; Representing the number of sampling steps between adjacent trigger moments and , Representing the maximum number of sampling steps; Representing a sampling period of the system; Representing an error weight matrix; representing a design constant; , Representing a dynamic threshold function and , The basic parameter values representing the dynamic threshold function, An upper bound representing a dynamic threshold; A data difference representing error status information to be transmitted currently and error status information that has been transmitted last time; a data difference representing the error status information currently to be transmitted and an average value of the plurality of historical error status information; , , respectively representing the current state information to be transmitted by the sender unmanned aerial vehicle, the last transmitted state information and the historically transmitted information data; Representing design parameters; s42, defining time-varying delay according to a dual-threshold dynamic event trigger mechanism as follows: (17) (18) wherein: Indicating the last trigger time; representing the current sampling step size; Representing the starting time from the current trigger period To the current time Offset of (2); Respectively representing a network delay lower bound, a maximum network delay and a communication maximum delay; S43. according to formulae (14) to (16), it is possible to obtain: (19) wherein: representing global intermediate variables and ; ; And formula (19) is rewritten as follows according to formula (17): (20) wherein: ; S44, acquiring the first based on the formula (20) combined with the dual-threshold dynamic event trigger mechanism Unmanned aerial vehicle formation control law of individual follower unmanned aerial vehicle The method comprises the following steps: (21) wherein: representing intermediate parameters and , Representation Time input to the first Interfered first of individual follower unmanned aerial vehicle Position vectors and velocity vectors of the individual follower unmanned aerial vehicles; representing the control gain.
6. The four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack according to claim 5, wherein the step S5 specifically comprises the following steps: S51, aiming at the established unmanned aerial vehicle formation control law and an error state equation, the unmanned aerial vehicle formation system is subjected to the first step The state error system of the unmanned aerial vehicle based on the dual-threshold dynamic event trigger mechanism is expressed as: (22) s52, defining global intermediate variables as: , ,(23) And the closed loop equation of the four-rotor unmanned aerial vehicle formation system based on the double-threshold dynamic event triggering mechanism is obtained by combining the formula (5), the formula (13), the formula (22) and the formula (23): (24) wherein: representing intermediate parameters and , Representing a matrix of weight parameters of a communication link and Representing a Laplace matrix ; Representing a parameter matrix and Representing intermediate variables and Representing an adjacency matrix and Representing a real number; Representing the identity matrix; Representation of The disturbed manifestation; And S53, according to a closed loop equation of the four-rotor unmanned aerial vehicle formation system and a distributed controller of the formation system, the four-rotor unmanned aerial vehicle formation control suitable for asynchronous packet loss and random UDP attack is realized.

Description

Four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack Technical Field The invention relates to the technical field of unmanned aerial vehicle formation control, in particular to a four-rotor unmanned aerial vehicle formation control method suitable for asynchronous packet loss and random UDP attack. Background In recent years, formation control of multi-frame four-rotor unmanned aerial vehicle capable of realizing various complex tasks has become a future development trend. Different from the centralized control [1] provided with the central control unit and the distributed control [2] which is in one-to-one correspondence with the unmanned aerial vehicles, the distributed control has no control core, and each unmanned aerial vehicle in the formation needs to interact with adjacent nodes, so that the control of the whole formation is realized as in the document [3]. A new distributed reinforcement learning behavior control is presented, document [4], that can reduce the cumulative and instantaneous costs in the distributed formation and obstacle avoidance process. Document [5] devised a new distributed formation control method, selecting additional states in Model Predictive Contour Control (MPCC) as co-parameters, requiring QUAV to communicate co-parameters and location to neighbors. However, existing formation control schemes lack comprehensive coping mechanisms for hybrid communication anomalies (e.g., network attacks and non-aggressive faults). The existing research mostly solves the problems of malicious attack or random packet loss in isolation, and fails to fully reflect the complexity of asynchronous occurrence of abnormal problems in the actual communication environment, so that the stable operation of the system under the real and unreliable communication conditions is difficult to ensure. The network information security of unmanned aerial vehicle formation is particularly important whether the information exchange between adjacent unmanned aerial vehicles or the communication between the unmanned aerial vehicles and the ground station is carried out. User datagram protocol (User)Protocol, UDP) is a Protocol that works at the transport layer in the open system interconnect model, and is mainly characterized by connectionless, not guaranteed to reliably transmit and face the message [6]; The network security method and the network security device draw attention to QUAV networks, especially network security problems, design a new intrusion detection and response scheme capable of detecting network threat anomalies [7], and provide a method based on regional or parallel polyhedrons ) The novel detection method of the unmanned aerial vehicle comprises the steps of (8) taking the flight process of the unmanned aerial vehicle into consideration, establishing a system model of the unmanned aerial vehicle by utilizing quaternions, effectively detecting an attack of the unmanned aerial vehicle system (9), analyzing the stability influence of random packet loss on the same type of problems in a literature [10], taking a single fixed-wing aircraft as an example in a literature [11] and taking control of UDP attack and packet loss asynchronization as an example, and providing a prediction compensation method in a literature [12] and designing a controller by using a predicted value of lost data. Document [13] when considering packet loss or UDP attacks, a markov chain is applied, taking into account its control model, with the aim of finding a strategy that maximizes long-term return. Most of the above researches are to attribute the data packet loss to network attack, however, in the formation of an actual unmanned aerial vehicle, the packet loss may be caused by non-aggressive factors such as network congestion, link delay and the like, and more importantly, the current researches generally fail to distinguish and model special scenes in which the attack and various packet losses occur asynchronously. The distributed formation control of unmanned aerial vehicles requires frequent status information communication between neighboring unmanned aerial vehicles, and usually an event trigger mechanism is introduced to relieve the communication pressure of the system. Document [14] is directed at four rotor unmanned aerial vehicle, utilizes the discrete time to interfere the observer and designs a discrete time neural control based on event triggering, can reach effective control to unmanned aerial vehicle's external disturbance and input saturation. The literature [15] considers that the four-rotor unmanned aerial vehicle system has the characteristic of model dynamic uncertainty, and when interference is studied, a self-adaptive event triggering control method is provided, and the update frequency of a designed controller is reduced. Document [16] proposes an event trigger mechanism based on a deep learning control strategy for real-time trajecto