Multi-quadrotor unmanned aerial vehicle preset performance hunting fault-tolerant control method

Abstract

The invention relates to a multi-quadrotor unmanned aerial vehicle preset performance formation flight control method with event triggering, and the method comprises the steps: estimating an unknown state in a system through a reduced-order dynamic gain Kalman filter, and designing a controller through a distributed adaptive control method; meanwhile, a dynamic surface control technology and an event triggering technology are adopted in the design process of the controller, and state performance constraint on each unmanned aerial vehicle is achieved on the basis that the initial state of each four-rotor unmanned aerial vehicle is not limited and only the position information of the unmanned aerial vehicle needs to be obtained. Under the condition that the initial state of each quadrotor is not limited and only the position information of the quadrotor unmanned aerial vehicle is measured, the transient and steady state performance constraint of the flight state of each unmanned aerial vehicle can be realized, and communication bandwidth resources between an unmanned aerial vehicle controller and an actuator motor as well as between other unmanned aerial vehicles can be saved; and the system has good anti-interference and fault-tolerant capabilities.

Description

Technical Field

[0001] The present invention relates to the control technology of unmanned aerial vehicles, and particularly to a multi-quadrotor unmanned aerial vehicle preset performance pursuit fault-tolerant control method. Background Art

[0002] The pursuit control of multi-quadrotor unmanned aerial vehicles is widely applied in military and civilian fields. In these applications, it is usually necessary to ensure that multiple quadrotor unmanned aerial vehicles achieve pursuit with set performance requirements, while ensuring the safety and effectiveness of the system.

[0003] The preset performance pursuit control of multi-quadrotor unmanned aerial vehicles is a control method aiming to ensure that the transient and steady-state performance of each quadrotor unmanned aerial vehicle meets the pre-set requirements. However, traditional control methods usually require strict restrictions on the initial positions of each quadrotor unmanned aerial vehicle, which may not conform to the actual applications in many cases.

[0004] In addition, since the unmanned aerial vehicle is equipped with devices such as a drive module and a high-precision sensor, while improving the task execution ability of the unmanned aerial vehicle, it also poses huge challenges to the energy supply and limited communication bandwidth resources. Considering the actual situation, the state information and its own parameters of the quadrotor unmanned aerial vehicle are not easy to measure. Therefore, reasonably saving the communication bandwidth resources of the quadrotor unmanned aerial vehicle and accurately estimating the unknown states and parameters can improve the stability of the system and enable it to better achieve the preset control objectives.

[0005] In the existing multi-quadrotor preset performance pursuit control methods, the initial positions of each quadrotor are strictly restricted. However, in engineering applications, the initial positions can often be arbitrary values. If the initial position exceeds the preset value, the quadrotor unmanned aerial vehicle cannot well achieve the desired performance, and most of the existing control methods do not consider the limited communication bandwidth resources of the system and possible actuator failures. Summary of the Invention

[0006] The object of the present invention is to provide a multi-quadrotor unmanned aerial vehicle preset performance pursuit fault-tolerant control method with unrestricted initial positions and reduced communication resource consumption.

[0007] The present invention adopts the following technical solutions:

[0008] A multi-quadrotor unmanned aerial vehicle preset performance pursuit fault-tolerant control method estimates the unknown states in the system through a reduced-order dynamic gain Kalman filter, and then designs an adaptive position controller and an attitude controller for each unmanned aerial vehicle by using dynamic surface control technology and event-triggering technology, and simultaneously designs a consistent yaw angle controller for each unmanned aerial vehicle to achieve formation pursuit control.

[0009] Further, the reduced-order dynamic gain Kalman filter is:

[0010]

[0011] Among them, n = x, y, z represent the three directions of x, y, and z in three-dimensional space, ξ in , o in , Ξ in , ν in are functions of filter dynamic variables for estimating system state variables, and q in is the filter constant to be designed. l in is the filter dynamic gain. is the system control input.

[0012] Furthermore, the dynamic gain of the Kalman filter is:

[0013]

[0014] Among them, l i (0) = 1, σ i14 , σ i15 are positive constants.

[0015] Furthermore, the position controller is:

[0016]

[0017] Among them, c in2 , δ in1 , are positive constants, is the second-order error of the i-th quadrotor UAV, l i , v in are terms in the filter, is the derivative of the virtual controller α in1 .

[0018] Furthermore, the virtual controller is:

[0019]

[0020] Among them,

[0021]

[0022] π in1 = π in βρ in , ∈ i1 σ i11 , σ i12 , σ i14 , σ i15

[0023] is a positive constant, is β i β iq λ i χ i estimated values;

[0024] The adaptation rate is designed as:

[0025]

[0026] where is a positive constant.

[0027] Furthermore, the attitude controller is:

[0028]

[0029] The adaptation rate is:

[0030]

[0031] where c im2 , and are positive constants.

[0032] Furthermore, the yaw angle controller is:

[0033]

[0034] where

[0035]

[0036] The beneficial effects of the present invention are as follows: The present invention adopts a barrier function to limit the transient and steady-state performance of a quadrotor UAV. This new type of barrier function can relax the initial position of the quadrotor UAV to any value. At the same time, an event-triggered controller is designed using event-triggering technology, saving the waste of communication bandwidth resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a three-dimensional diagram of four UAVs capturing one UAV in a simulation experiment.

[0038] Figure 2 is the error limitation of four UAVs in the x, y, and z directions in a simulation experiment.

[0039] Figure 3 is the triggering time of the event-triggered controller designed for each UAV in a simulation experiment.

[0040] Figure 4It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 1 second.

[0041] Figure 5 It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 2 seconds.

[0042] Figure 6 It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 3 seconds.

[0043] Figure 7 It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 12 seconds.

[0044] Figure 8 It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 22 seconds.

[0045] Figure 9 It is a spatial position diagram of 1 target crazylife drone and 4 security crazylife drones under the multi-unmanned collaborative control motion capture experimental platform at 35 seconds. Specific implementation manners

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. It should be noted that the protection scope of the present invention is not limited to these embodiments. Any change or equivalent substitution that does not deviate from the concept of the present invention is included in the protection scope of the present invention.

[0047] A multi-quadrotor UAV preset performance formation pursuit fault-tolerant control method with event triggering estimates the unknown states of the quadrotor UAVs through a reduced-order dynamic gain Kalman filter, designs a controller through a distributed adaptive control method, and simultaneously adopts a dynamic surface control technology and an event triggering technology in the design process of the controller to constrain the transient and steady-state performance of the multi-quadrotor UAV system when the initial position states of the quadrotor UAVs are arbitrary and some states are unknown.

[0048] 1. Dynamic model of uncertain quadrotor UAVs

[0049]

[0050] Among them, [x i , y i , z i and [φ i , θ i , ψ i are the position and attitude angles of the quadrotor UAV, are uncertain parameters. I ix , I iy and I iz represent the inertia matrices of the quadrotor UAV in the x, y, and z directions respectively. J ip is the moment of inertia of the propeller, W iφ and W iθ are the roll angle and yaw angular velocity respectively. Ω ir represents the propeller speed margin. g represents the acceleration due to gravity. d ix , d iy , d iz , d φ , d θ , d ψ represent the disturbances of the system in different directions, u φ , u θ , u ψ represent the control inputs of the system in three attitudes u ix = u i1 (cosφ i sinθ i cosψ i + sinφ i sinψ i ), u iy = u i1 (cosφ i sinθ i sinψ i - sinφ i cosψ i ), u iz = u i1 (cosφ i cosθ i ). The mathematical model of the actuator fault is:

[0051] 2. Design of the controller

[0052] (1) To better describe the dynamic model of the quadrotor UAV and facilitate the design of the controller, the system state variables are redefined as follows:

[0053]

[0054] (2) Due to the strong coupling of the quadrotor UAV, the system is decoupled into a position subsystem and an attitude subsystem.

[0055] Position subsystem:

[0056]

[0057] Attitude subsystem:

[0058]

[0059] Among them, are gravity-related parameters f iφ1 = x iθ2 x iψ2 f iφ2 = J ip x iθ2 Ω ir f iθ1 = x iφ2 x iψ f iθ2 = J ip x iφ2 Ω ir f iφ2 = J ip x iθ2 Ω ir f iψ1 = x iφ2 x iθ f iψ2 = 0.

[0060] (3) Define the encirclement protocol

[0061]

[0062] This protocol is an encirclement control protocol based on the consensus control protocol, where a iq represents the communication situation between UAVs. If the i-th UAV can receive the information of the q-th UAV, then a iq = 1; otherwise, a iq = 0. h iqn represents the error between the i-th and q-th UAVs in the direction n.

[0063] (4) The reduced-order filter is designed as:

[0064]

[0065] Among them, n = xyz represents the x, y, and z directions in three-dimensional space. ξ in o in Ξ in νin q is a function for the filter to estimate the system state variables in l is the filter constant to be designed in is the dynamic gain of the filter is the system control input

[0066] Dynamic gain:

[0067]

[0068] where l i (0) = 1 π i1 is a preset performance-related function in the controller is σ i14 σ i15 is a positive constant

[0069] (5) Design a barrier function for the multi-agent system according to the time-varying function β(t) to achieve the preset performance control of the multi-agent with unrestricted initial state

[0070] β(t) must be a strictly increasing function and satisfy β(0) = 1 and 0 < b f 01. The preset function constructed according to the time-varying function is: where l is a positive constant

[0071] The barrier function designed for the position and attitude of each UAV according to the time-varying function is:

[0072]

[0073] where ζ in (t) = β(t)η in (t) Λ in is the encirclement protocol of the i-th UAV in the direction n

[0074] (6) Design of the quadrotor UAV position controller

[0075] Perform the following coordinate transformation according to the position system equation and the barrier function

[0076]

[0077] Use the backstepping method to recursively design the controller, where z in1 is the first-step error of the i-th UAV in the direction n during the recursive process, and z in2 is the second-step error of the i-th UAV in the direction n during the recursive process. H in2For the error between the virtual controller and the estimated virtual controller.

[0078] Taking the derivative of z in1 can design the virtual controller in the first step of the recurrence process as:

[0079]

[0080] where, π in1 = π in βρ in , ∈ i1 σ i11 , σ i12 , σ i14 , σ i15 are positive constants. For the estimated values of β i , β iq , λ i , χ i .

[0081] The adaptive rate is designed as:

[0082]

[0083] where, is a positive constant.

[0084] Design a position controller in the following form:

[0085]

[0086] where, c in2 , δ in1 , are positive constants, is the second-order error of the i-th quadrotor UAV, l i , v in are terms in the reduced-order filter, is the derivative of the virtual controller α in1 in the first step.

[0087] (7) Design of the attitude controller for the quadrotor UAV

[0088] Connect the position subsystem and the attitude subsystem through the attitude extraction algorithm. The existing expression relationships are as follows:

[0089]

[0090] Decouple the formula and combine it with the attitude equation mentioned above to obtain:

[0091]

[0092] where \(m = (\varphi,\theta)\).

[0093] Using the same derivation process as the position controller, the attitude controller can be obtained:

[0094]

[0095] Adaptive rate:

[0096]

[0097] where c im2 , and are positive constants.

[0098] (8) Design of the yaw angle controller for a quadrotor UAV

[0099] To achieve formation hunting control, we will design a consistent yaw angle controller \(u\) iψ for each UAV, as follows:

[0100]

[0101] where

[0102] The yaw angle controller is as follows:

[0103]

[0104] 3. Simulation experiments

[0105] The present invention uses Matlab software to simulate four quadrotors forming a formation to hunt one UAV. Among them, the dynamic model of the quadrotor UAV and related algorithms are programmed using.M files, and the model is built using Simlink modules.

[0106] During the simulation, the running duration and step size of the control system are reasonably set, and compared and tested with the corresponding simulation. To further prove the effectiveness of the present invention, an experiment on 4 security UAVs forming a formation to hunt 1 target UAV is carried out using a multi-UAV cooperative control dynamic hunting experiment platform.

[0107] The trajectory of the hunted target is set as follows: [y xd , y yd , y zd , ψ d T = [0.5sin(t), 0.5t, 0.1t, 0] T , and the disturbance of each UAV in the direction \(j=(x,y,z,\varphi,\theta,\psi)\) is

[0108] The initial state is set as follows: x 1x1 (0) = 1.5, x 1y1 (0) = 1.5, x 1z1 (0) = 1.5, x 2x1 (0) = 0.5, x 2y1 (0) = 0.5, x 2z1 (0) = 0.5, x 3x1 (0) = 0.5, x 3y1 (0) = 0.5, x 3z1 (0) = 0.2, x 4x1 (0) = 0.3, x 4y1 (0) = 0.2, x 4z1 (0) = 1.5. The acceleration due to gravity g = 9.8 m / s 2 , q in = q iψ = 10, l in (0) = 1, l iψ (0) = 1.

[0109] By observing Figure 1 it is found that according to the trajectory of the target UAV and the positions of the four chasing UAVs at multiple moments, the target can be effectively captured. By observing Figure 2 it can be found that the errors of each UAV in the x, y, and z directions can converge well between the upper and lower bounds of the set performance function. The output signal of the controller is as Figure 3 shown. The controller signal changes from a continuous signal to an intermittent signal, which can effectively reduce the number of controller updates, reduce the energy consumption of the controller, and the occupancy of the communication bandwidth. By observing Figures 4 to 9 , which are the spatial positions of 1 target UAV and 4 security UAVs under the multi-UAV cooperative control and capture experiment platform at 1 second, 2 seconds, 3 seconds, 12 seconds, 22 seconds, and 35 seconds respectively, it can be found that the four UAVs in the experiment can effectively capture the target UAV. In summary, it shows that this control method is effective.

[0110] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A fault-tolerant control method for capturing multiple quad-rotor drones with preset performance, characterized in that: A reduced-order dynamic gain Kalman filter is used to estimate the unknown states in the system, and dynamic surface control technology and event triggering technology are used to design adaptive position controllers and attitude controllers for each UAV. At the same time, a consistent yaw angle controller is designed for each UAV to achieve formation capture control.

2. A method for controlling the preset performance of a multi-quadrotor drone according to claim 1, characterized in that: The reduced-order dynamic gain Kalman filter is: Where n = x, y, z represents the three directions x, y, z in three-dimensional space; ξ in , o in ,Ξ in , ν in is the function of the filter dynamic variable used to estimate the system state variable, q in is the filter constant to be designed; l in is the filter dynamic gain; It is the system control input.

3. A method for controlling the preset performance of a multi-quadrotor drone according to claim 2, characterized in that: The dynamic gain of the Kalman filter is: in, σ i14 , σ i15 Is a normal number.

4. A method for controlling the preset performance of a multi-quad-rotor drone according to claim 3, characterized in that: The position controller is: Among them, c in2 , δ in1 , is a normal number, is the second-order error of the i-th quadrotor drone, l i ,v in is the term in the filter, For virtual controller α in1 The derivative of .

5. The method for controlling the preset performance of a multi-quad-rotor drone according to claim 4 is characterized in that: The virtual controller is: in, p in1 =π in br in , ∈ i1 s i11 ,s i12 ,s i14 、 σ i15 is a normal number, β i , β iq , i , χ i An estimated value of The adaptive rate is designed as: in, Is a normal number.

6. A method for controlling the preset performance of a multi-quadrotor drone according to claim 5, characterized in that: The attitude controller is: The adaptive rate is: in, c im2 , and Is a normal number.

7. A method for controlling the preset performance of a multi-quadrotor drone according to claim 6, characterized in that: The yaw angle controller is: in,

Images