Method and System for Fault Isolation and Formation Reconfiguration of UAV Swarms Based on Bird Flock Algorithm

Abstract

The present invention relates to a method and system for fault isolation and formation reconstruction of an unmanned aerial vehicle (UAV) swarm based on a bird flock algorithm, belonging to the technical field of UAV formation control. Based on the bird flock algorithm, improvements are made according to the characteristics of the UAV swarm. By improving the bird flock algorithm, when a faulty UAV appears in the UAV swarm, the faulty UAV can be made to leave the formation and return for landing automatically, so as to avoid threatening the flight safety of other UAVs in the formation. At the same time, other UAVs are controlled to automatically fill in the formation positions to achieve formation reconstruction. The present invention has high formation control stability, structural flexibility and fault tolerance, and can effectively meet the requirements for the UAV swarm to perform formation flight tasks.

Description

Technical Field

[0001] The present invention belongs to the technical field of UAV formation control, in particular to a method and system for UAV swarm fault isolation and formation reconfiguration based on the bird swarm algorithm. Background Art

[0002] In recent years, with the continuous increase in the application scenarios of unmanned aerial vehicles (UAVs), the application form has changed from a single UAV to multiple UAVs. The UAV swarm has strong flexibility and stability, a large activity range, and high task execution efficiency, and can be competent for tasks in military or civilian fields such as logistics transportation, power inspection, pollution investigation, battlefield reconnaissance, and strikes. The swarm usually adopts the formation flight method to perform these complex tasks. Therefore, researching the problem of multi-UAV formation has important theoretical and application values.

[0003] However, with the expansion of the scale of the UAV swarm formation, the failure probability of a single UAV will also increase sharply. Therefore, there is a significant positive correlation between the scale of the UAV swarm formation and the failure probability of the UAVs within the formation. In addition, with the increase in the flight distance and flight duration of the UAVs, for a single UAV, its failure probability increases significantly. For a UAV formation, there is a high probability that a UAV within the formation will fail.

[0004] When a serious failure occurs to a UAV within the formation, the flight attitude of the UAV will change, and its flight trajectory will also change. In a UAV formation with a short inter-aircraft distance, the flight trajectory of the failed UAV changes significantly, which may cause UAV collisions and the crash of the formation UAVs, posing a greater threat to the flight safety of the formation.

[0005] The current fixed-wing UAV formation reconfiguration technology uses the method of the ground station inputting a preset flight path into the failed UAV to lead the failed UAV away from the fixed-wing UAV formation, thereby achieving the isolation of the failed UAV, and then realizing formation reconfiguration by adjusting the inter-aircraft distance. However, in some failure states, the failed UAV cannot determine its own position. At this time, the existing technology cannot ensure the safe separation of the UAV from the formation and safe landing, and in severe cases, serious accidents such as collisions between the UAVs of the formation members may occur.

[0006] The Boid algorithm is an artificial life project that simulates the flocking behavior of birds. It was developed by Craig Reynolds in 1986. This model is often used in computer animation or computer-aided design for three-dimensional computer geometry. His paper on this topic was published in the proceedings of the 1987 ACM SIGGRAPH (the top annual conference on computer graphics organized by the Special Interest Group on Computer Graphics of the Association for Computing Machinery). "Boid" is an abbreviation for "bird-oid object", referring to a bird-like object. The rules applicable in the simplest Boids world are as follows, which describe how individuals in a flock move based on the positions and velocities of their neighboring companions: Separation, that is, moving to avoid overcrowding in the group; Alignment, that is, moving in the average direction of the surrounding companions; and Cohesion, that is, moving towards the average position (centroid) of the surrounding companions. Summary of the Invention

[0007] The technical problem to be solved by the present invention is:

[0008] To avoid UAV accidents caused by the formation reconstruction by adjusting the inter-machine distance in the prior art and improve the safety of UAVs. The present invention provides a method and system for UAV swarm fault isolation and formation reconstruction based on the Boid algorithm, which removes the faulty UAVs from the formation and makes them return and land on their own, while completing the formation and continuing to perform tasks.

[0009] To solve the above technical problem, the technical solution adopted by the present invention is:

[0010] A method for UAV swarm fault isolation and formation reconstruction based on the Boid algorithm, characterized by the following steps:

[0011] S1: After the UAV swarm completes assembly, the wingman obtains the flight speed vectors in the left front, directly in front, and right front of the aircraft through in-formation communication within the formation, and performs real-time summation on the obtained flight speed vectors to obtain the expected flight speed vector of the aircraft in three-dimensional space, and inputs the expected flight speed vector into the flight control subsystem to complete formation control;

[0012] S2: Each UAV performs periodic fault diagnosis by organizing the time-frequency characteristics of the signal into a feature map for image diagnosis;

[0013] S3: Each UAV in the formation broadcasts its own information within the formation and simultaneously sends it to the ground station; the said own information includes the aircraft number, the aircraft status, and the fault type;

[0014] S4: When the ground station receives that the status of the aircraft is faulty, it uses the Boid formation control algorithm for formation fault isolation and formation reconstruction;

[0015] The formation fault isolation is as follows: The ground station sends virtual formation status information to the faulty UAV. The faulty UAV disengages through the virtual formation status information. After disengaging from the formation, the ground station sends an instruction to turn off the flock tracking function of the faulty UAV and sends a control instruction to the faulty UAV to control it to return or make an emergency landing.

[0016] The formation shape reconstruction is as follows: The ground station judges according to the missing position of the faulty UAV in the formation, shields the UAV for filling positions in the formation communication network, sends the status data of the virtual UAV to the UAV for filling positions, drives the UAV for filling positions to fill to the expected position, and then restores the communication between each UAV in the formation.

[0017] A further technical solution of the present invention: S1 The real-time summation formula for the obtained flight speed vectors is:

[0018]

[0019] In the formula, the vector represents the own speed vector, represents the speed vector of the left front that the own aircraft tracks, represents the speed vector of the dead ahead that the own aircraft tracks, represents the speed vector of the right front that the own aircraft tracks. Except for the own speed vector, the speed vector expressions of the above three other UAVs are as follows:

[0020]

[0021] In the formula, represents the speed vector of a single UAV tracked by the own aircraft in the body axis system, represents the speed vector of a single UAV tracked by the own aircraft in the inertial coordinate system, represents the pitch angle of a single UAV tracked by the own aircraft, the yaw angle of a single UAV tracked by the own aircraft, represents the roll angle of a single UAV tracked by the own aircraft, i = 1, 2, 3.

[0022] A further technical solution of the present invention: S2 The fault diagnosis is specifically as follows: The collected signals of the sensor subsystem are subjected to A / D sampling and noise reduction processing to obtain the preprocessed signals for fault diagnosis; several signal data points are extracted and sorted into a three-dimensional matrix of time-domain signals; the three-dimensional matrix of time-domain signals is subjected to Fourier transform to obtain a frequency-domain diagram, and the frequency-domain diagram is converted into RGB data and added to the three-dimensional matrix of time-domain signals to form a three-dimensional input signal matrix for the fault diagnosis module; the three-dimensional input signal matrix for the fault diagnosis module is input into the trained fault diagnosis module to output the fault diagnosis result.

[0023] A further technical solution of the present invention: The sensor subsystem includes a triaxial accelerometer, a gyroscope, a barometric altimeter, an airspeed sensor, a GPS sensor, and a temperature sensor.

[0024] An unmanned aerial vehicle (UAV) swarm fault isolation and formation reconstruction system based on a bird flock algorithm, characterized by comprising a fault diagnosis module and a formation fault-tolerant control module;

[0025] The fault diagnosis module performs diagnosis by organizing the time-frequency characteristics of the signal into a feature map for image diagnosis; the frequency-domain signal of the signal is decomposed through fast Fourier transform, a feature map is formed according to the amplitude-frequency characteristics of the signal, and finally the feature map is identified and compared to obtain the final fault type, which serves as the trigger condition for formation fault-tolerant control;

[0026] The formation fault-tolerant control module uses a bird flock formation control algorithm for formation control, formation fault isolation, and formation reconstruction.

[0027] The beneficial effects of the present invention are as follows:

[0028] An unmanned aerial vehicle (UAV) swarm fault isolation and formation reconstruction method and system based on a bird flock algorithm provided by the present invention is improved according to the characteristics of the UAV swarm on the basis of the bird flock algorithm. By improving the bird flock algorithm, when a faulty UAV appears in the UAV swarm, the faulty UAV can automatically leave the formation and return for landing to avoid threatening the flight safety of other UAVs in the formation. At the same time, other UAVs are controlled to automatically fill in the formation positions to achieve formation reconstruction. The present invention has high formation control stability, structural flexibility, and fault tolerance, and can effectively meet the requirements of the UAV swarm for performing formation flight tasks. Description of the Drawings

[0029] The drawings are only for the purpose of illustrating specific embodiments and are not considered to be a limitation of the present invention. Throughout the drawings, the same reference signs denote the same components.

[0030] Figure 1 It is a structural diagram of an unmanned aerial vehicle (UAV) swarm fault isolation and formation reconstruction device.

[0031] Figure 2 It is a schematic flow chart of an unmanned aerial vehicle (UAV) swarm fault isolation and formation reconstruction method based on a bird flock algorithm. Detailed Embodiments

[0032] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0033] As Figure 1 shown, an embodiment of the present invention provides a device for fault isolation and formation reconfiguration of a UAV swarm based on a bird flock algorithm. It includes a software platform system and a hardware platform system. The software platform system is used to implement periodic fault diagnosis of a single UAV, fault tolerance control of a single UAV, and formation fault tolerance control based on the improved bird flock algorithm and virtual bird flock algorithm proposed by the present invention; the hardware platform system is used to receive and transmit the numbers, status information, and fault information of the member UAVs in the formation, the ground station receives and transmits control instruction information, and uses the flight control computer and the ground station computer for flight control and formation calculation. Among them, the hardware platform is composed of three subsystems, namely the flight control subsystem, the sensor subsystem, and the communication subsystem, and the software platform consists of two application software, namely the fault diagnosis module and the formation fault tolerance control module.

[0034] Regarding the UAV carried by the present invention, to achieve this design objective, it is necessary to separately design the software and hardware involved in the present invention. The software platform system stores parameters such as the fault data of the UAV and the precise model of the UAV, and drives the hardware platform to achieve the expected functions, while the hardware platform provides the hardware platform for the software platform system to implement functions.

[0035] I. Design of the software platform system

[0036] The software platform system involved in the present invention refers to a control algorithm designed based on the existing UAV software framework, and this algorithm includes two parts: fault diagnosis and formation fault tolerance control.

[0037] 1. Fault diagnosis module

[0038] The fault diagnosis module involved in the present invention relies on a signal-based fault diagnosis program to perform periodic fault diagnosis on the UAV. Each UAV is equipped with a fault diagnosis module for diagnosing its own faults.

[0039] This module performs fault diagnosis based on the processed sensor data. After the UAV enters the stable state, that is, after the UAV enters the stable climbing, cruising, and descending stages, the UAV first performs A / D sampling on the signals of the sensor subsystem including the signals of the three-axis accelerometer, gyroscope, barometric altimeter, airspeed sensor, GPS sensor, and temperature sensor, decomposes the continuous quantity of the original signal into discrete quantities, and then performs noise reduction processing on the discrete quantities to obtain the preprocessed signal for fault diagnosis, and transmits the preprocessed signal to the fault diagnosis module. After the signal is preprocessed, the data of the six sensors each take the latest 2,500 signal data points collected, and are organized into a three-dimensional time-domain signal matrix of 50*50*6. The matrix is subjected to Fourier transform to obtain a frequency-domain diagram, the frequency-domain diagram is converted into RGB data, and is organized into a 50*50 format, and added to the three-dimensional time-domain signal matrix to form a three-dimensional fault diagnosis module input signal matrix of 50*50*7. After being trained with empirical data, the fault diagnosis module can classify the fault types. The input signal matrix of the fault diagnosis module is input into the fault diagnosis module, and its diagnosis and classification results are used as the trigger conditions for the formation fault-tolerant control program.

[0040] 2. Formation Fault-Tolerant Control Module

[0041] The formation fault-tolerant control module involved in the present invention relies on the flight control hardware to execute. As a fault-tolerant control strategy added on the basis of the existing single-aircraft flight control, it includes the functions of formation fault isolation and formation reconstruction.

[0042] In this embodiment, after each aircraft in the UAV swarm performs fault diagnosis, the diagnosis results are fed back to the ground station. The ground station controls the faulty UAV to break away from the formation, and then the faulty UAV returns or lands on its own according to the on-board single-aircraft fault-tolerant control program. During the flight of the UAV swarm formation, the heading control and altitude control of a single UAV both adopt an improved bird swarm algorithm. The algorithm idea is as follows: The lead aircraft of the formation performs path planning according to the task. The wingman obtains the flight speed vectors in its left front, directly in front, and right front within the formation through in-formation communication. The UAV performs real-time solution operations on the sum of the above UAV speed vectors. Its specific calculation expression is shown in Equation (1):

[0043] (1)

[0044] In the formula, the vector represents the speed vector of the aircraft itself, represents the speed vector in the left front that the aircraft itself tracks, represents the speed vector directly in front that the aircraft itself tracks, represents the speed vector in the right front that the aircraft itself tracks. Except for the speed vector of the aircraft itself, the speed vector expressions of the other three UAVs are shown in Equation (2):

[0045] (2)

[0046] In the formula, represents the velocity vector of a single UAV tracked by the host in the body axis system, represents the velocity vector of a single UAV in the inertial coordinate system, that is, the velocity vector of the single UAV tracked by the host, i = 1, 2, 3. represents the pitch angle of this UAV, represents the yaw angle of this UAV, represents the roll angle of this UAV.

[0047] The desired flight velocity vector of the host in three-dimensional space is obtained, and the desired flight velocity vector is input into the flight control system, thereby completing the formation control.

[0048] When the fault diagnosis module outputs a fault type signal, the formation fault tolerance control module of the UAV is triggered to isolate the influence of the faulty UAV on the formation and manipulate the remaining UAVs to perform formation reconstruction. When the ground station receives the fault signal, while shielding the information of the actual front UAV, it sends the data of the virtual UAV directly in front to the faulty UAV. After detecting the fault, the faulty UAV executes the fault tolerance control program by itself, moves downward to the lower part of the UAV group according to the virtual UAV directly in front sent by the ground station. When it breaks away from the formation flight range, the faulty UAV will return or make an emergency landing according to the fault situation. At this time, the formation transformation program stored in the ground station judges according to the missing position of the faulty UAV in the formation, then shields the UAVs that need to adjust their positions in the formation communication network, and the ground station sends the status data of the virtual UAV to the UAVs, driving the UAVs that need to adjust their positions in the formation to fill up to the expected positions, and then resumes the communication between each UAV in the formation.

[0049] II. Hardware Platform Design

[0050] 1. Design of Flight Control Subsystem

[0051] The flight control subsystem takes the flight control computer as the operation core, providing the hardware basis for the flight control and formation control of the UAV. The flight control subsystem includes a flight control computer and a single-machine fault-tolerant control computer. The flight control model is installed in the flight control computer, and real-time path planning is carried out according to the flight mission. This data is transmitted to the trajectory controller, and after the desired attitude angle is calculated by the trajectory controller, it is sent to the attitude controller. The attitude controller tracks the desired attitude angle and then sends it to the UAV, and provides the feedback data of the sensor subsystem to the position controller and the attitude controller. Among them, the three-axis position and acceleration information are provided to the position controller, and the UAV attitude angle data is provided to the attitude controller. Based on the flight control computer, the single-machine fault-tolerant control computer generates fault-tolerant control instructions according to the fault type by the stored fault-tolerant control program. In addition, this computer has the function of judging whether it has the ability to return.

[0052] 2. Design of Sensor Subsystem

[0053] The sensor subsystem is used to obtain flight state information and UAV fault information, broadcast the flight state information within the formation, and send the fault information report to the ground station, and at the same time provide it to the flight control subsystem as feedback information. The sensor subsystem includes six sensors: a three-axis accelerometer, a gyroscope, a barometric altimeter, an airspeed sensor, a GPS sensor, and a temperature sensor. Among them, the three-axis accelerometer is used to calculate the acceleration of the UAV in three axes to serve as an internal feedback signal for the flight control system; the gyroscope is used to calculate the attitude angle of the UAV; the barometric altimeter is used to measure the barometric altitude of the UAV to provide a reference for path planning and flight control; the airspeed sensor is used to measure the flight speed of the UAV relative to the airflow to provide a reference for the formation control between sub-aircraft groups; the data of the GPS sensor is used to calculate the distance between each UAV and calculate the formation position; the temperature sensor is used to monitor the temperature of the UAV itself and measure the atmospheric temperature at the same time to correct the atmospheric parameters of the flight controller.

[0054] 3. Design of Communication Subsystem

[0055] The communication subsystem is used for inter-UAV communication and communication between each UAV in the formation and the ground station. The communication subsystem includes a transceiver and a transceiver antenna. The transceiver is directly connected to the flight control subsystem. When the flight control subsystem or the ground station computer needs to transmit data, the flight control subsystem sends the serial data to the transceiver, and the transceiver broadcasts it to all UAVs in the UAV group or sends instructions to a single UAV separately through the antenna at a specific frequency, so as to achieve the purpose of transmitting data.

[0056] As Figure 2 shown, the embodiment of the present invention provides a method for UAV group fault isolation and formation reconstruction based on the bird swarm algorithm.

[0057] First, write the formation flight program according to the mission requirements and set the communication frequency for the UAV swarm. Subsequently, the UAVs in the swarm take off one by one. The UAVs that have taken off circle and wait in the airspace designated by the ground station near the airport. When the UAVs in the swarm are assembled, the lead UAV flies forward according to the pre-entered formation flight program (which is written according to the mission situation). The UAVs that take off later track the lead UAV based on the improved three-dimensional bird swarm algorithm proposed by the present invention to complete the formation assembly.

[0058] Secondly, after the UAV swarm completes the assembly, the wingman obtains the flight speed vectors including the UAV heading and climb angle in the left front, directly in front, and right front of the aircraft through in-formation formation communication. The UAV performs real-time solution operations on the above UAV speed vectors to obtain the expected flight speed vector of the aircraft in three-dimensional space, and inputs the expected flight speed vector into the flight control system to complete the conventional formation control. At the same time, when the UAV enters a stable state, that is, the stable climb, cruise, and descent states, each UAV conducts periodic diagnosis. The UAV first performs A / D sampling on the signals output by the sensor subsystem, decomposes the continuous quantity of the original signal into discrete quantities, and then processes the discrete quantities through noise reduction to obtain the preprocessed data for fault diagnosis. The preprocessed signal is transmitted to the fault diagnosis module. The fault diagnosis module first decomposes the signal frequency domain signal through fast Fourier transform, then forms a feature map according to the amplitude-frequency characteristics of the signal, and finally identifies and compares the feature map to obtain the final fault type.

[0059] Thirdly, after detecting a fault, the faulty UAV determines by itself whether it meets the return conditions. If it does, it leaves the formation and flies to the nearest airport according to the guidance of the ground station. If it does not, it chooses a nearby site for a forced landing. The faulty UAV sends a fault report containing information such as its own number, its own status, and the type of fault to the ground station. The ground station randomly sends a shielding instruction to the faulty UAV and sends virtual formation status information to the UAV. The faulty UAV tracks the virtual formation through the bird swarm algorithm to disengage from the real formation. After disengaging from the formation, the ground station sends an instruction to turn off the bird swarm tracking function of the faulty UAV and sends a control instruction to the faulty UAV to control it to land nearby. At the same time, the ground station judges according to the stored formation transformation program the missing position of the faulty UAV in the formation, then shields the UAVs that need to adjust their positions in the formation communication network. The ground station sends the status data of the virtual UAV to the UAVs, drives the UAVs that need to adjust their positions in the formation to fill in the expected positions, and then restores the communication between the UAVs in the formation.

[0060] Finally, the remaining UAVs in the formation continue to execute the mission, and the ground station controls the faulty UAV to land. If it meets the return landing conditions, it calculates the position and route of the nearest airport where it can land. If it does not, it avoids residential areas and densely populated areas for a forced landing.

[0061] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A method for fault isolation and formation reconfiguration of an unmanned aerial vehicle swarm based on a bird flock algorithm, characterized in that, The steps are as follows: S1: After the UAV swarm completes assembly, the wingman obtains the flight speed vectors in the left front, directly front, and right front of the aircraft through in-formation formation communication, and sums up the obtained flight speed vectors in real time to obtain the expected flight speed vector of the aircraft in three-dimensional space. The expected flight speed vector is input into the flight control subsystem to complete formation control; S2: Each UAV performs periodic fault diagnosis by organizing the time-frequency characteristics of the signal into a feature map for image diagnosis; S3: Each UAV in the formation broadcasts its own information within the formation and simultaneously sends it to the ground station; the said own information includes the aircraft number, the aircraft status, and the fault type; S4: When the ground station receives that the status of the aircraft is a fault, the bird swarm formation control algorithm is used for formation fault isolation and formation shape reconstruction; The said formation fault isolation: The ground station sends virtual formation status information to the faulty UAV. The faulty UAV detaches through the virtual formation status information. After detaching from the formation, the ground station sends an instruction to turn off the bird swarm tracking function of the faulty UAV and sends a control instruction to the faulty UAV to control the faulty UAV to return or make an emergency landing; The said formation shape reconstruction: The ground station judges according to the missing position of the faulty UAV in the formation, shields the UAV to fill the position in the formation communication network, sends the status data of the virtual UAV to the UAV to fill the position, drives the UAV to fill the position to the expected position, and then restores the communication between each UAV in the formation.

2. The method for fault isolation and formation reconstruction of an unmanned aerial vehicle swarm based on the bird flock algorithm according to claim 1, wherein The real-time summation formula for the flight speed vectors obtained in S1 is: In the formula, the vector represents the local velocity vector, represents the velocity vector of the left front being tracked by the local aircraft, represents the velocity vector of the dead ahead being tracked by the local aircraft, represents the velocity vector of the right front being tracked by the local aircraft. Except for the local velocity vector, the velocity vector expressions of the other three UAVs are as follows: In the formula, represents the velocity vector of a single UAV tracked by the host in the body-axis system, represents the velocity vector of a single UAV tracked by the host in the inertial coordinate system, represents the pitch angle of a single UAV tracked by the host, the yaw angle of a single UAV tracked by the host, represents the roll angle of a single UAV tracked by the host, i = 1, 2, 3.

3. A method for fault isolation and formation reconstruction of an unmanned aerial vehicle swarm based on the bird flock algorithm according to claim 1, characterized in that, The specific fault diagnosis in S2 is: The acquisition signal of the sensor subsystem is subjected to A / D sampling and noise reduction processing to obtain the preprocessed signal for fault diagnosis; several signal data points are extracted and organized into a three-dimensional matrix of time-domain signals; the three-dimensional matrix of time-domain signals is subjected to Fourier transform to obtain a frequency-domain map, and the frequency-domain map is converted into RGB data and added to the three-dimensional matrix of time-domain signals to form the input signal matrix of the three-dimensional fault diagnosis module; The input signal matrix of the three-dimensional fault diagnosis module is input into the trained fault diagnosis module to output the fault diagnosis result.

4. A method for UAV swarm fault isolation and formation reconstruction based on the bird flock algorithm according to claim 3, characterized in that, The said sensor subsystem includes a three-axis accelerometer, a gyroscope, a barometric altimeter, an airspeed sensor, a GPS sensor, and a temperature sensor.

5. A system for implementing the method for fault isolation and formation reconstruction of an unmanned aerial vehicle swarm based on the bird flock algorithm according to claim 1, characterized in that, It includes a fault diagnosis module and a formation fault tolerance control module; The said fault diagnosis module diagnoses by organizing the time-frequency characteristics of the signal into a feature map for image diagnosis; the frequency-domain signal of the signal is decomposed through fast Fourier transform, a feature map is formed according to the amplitude-frequency characteristics of the signal, and finally the feature map is identified and compared to obtain the final fault type as the trigger condition for formation fault tolerance control; The said formation fault tolerance control module uses the bird swarm formation control algorithm for formation control, formation fault isolation, and formation shape reconstruction.

Images