A method for swarm control of quad-rotor unmanned aerial vehicle cluster based on quantized communication under communication bandwidth constraint

Abstract

The application discloses a kind of communication bandwidth constraint under the method for swarming control of quad-rotor unmanned aerial vehicle cluster based on quantized communication, comprising: the dynamic model of quad-rotor unmanned aerial vehicle is decomposed into position outer loop and attitude inner loop subsystem;Design a kind of based on collaborative expected swarming trajectory generation and tracking quad-rotor unmanned aerial vehicle swarming control hierarchical architecture;According to the communication ability of unmanned aerial vehicle, based on algebraic graph theory, the communication connection relationship of quad-rotor unmanned aerial vehicle is established;For each unmanned aerial vehicle, design a collaborative expected swarming trajectory generation algorithm based on quantized information, obtain the expected trajectory of each unmanned aerial vehicle;Each unmanned aerial vehicle realizes the tracking of expected trajectory by bottom controller, and the cluster finally realizes speed consistency, collision avoidance and relative aggregation.The application realizes the swarming control of large-scale quad-rotor unmanned aerial vehicle cluster, and is suitable for large-scale quad-rotor unmanned aerial vehicle cluster take-off and landing control, wide-area search, area coverage and the like.

Description

Technical Field

[0001] This application belongs to the field of collaborative control technology for unmanned aerial vehicle (UAV) swarms, specifically relating to a method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints. Background Technology

[0002] Quadrotor drones possess significant advantages such as compact structure, high maneuverability, and vertical takeoff and landing capabilities. Compared to fixed-wing drones, they exhibit greater environmental adaptability and mission flexibility when performing tasks such as reconnaissance and search, communication relay, target tracking, and damage assessment in low-altitude airspace. In recent years, they have received widespread attention in both military and civilian fields. When using quadrotor drones in swarms for collaborative missions, all drones must form a stable, clustered formation during flight, with each drone maintaining a consistent speed and a safe distance to avoid collisions. These behavioral characteristics are essentially consistent with the swarming movements of flocks of birds or schools of fish in nature.

[0003] Regarding the swarm control problem in multi-agent systems, existing research has formed a relatively systematic theoretical framework based on the three basic rules proposed by Reynolds: aggregation, velocity matching, and collision avoidance. Tanner et al. transformed these rules into a control law containing artificial potential field terms and velocity consistency terms, and proved the convergence of swarm behavior on a second-order integrator model. Subsequent research has further extended swarm control to environments with obstacles, heterogeneous multi-agent systems, and pre-defined formation configurations, achieving substantial theoretical results. Inspired by these advances, the swarm control problem of quadcopter UAV swarms has also been preliminarily explored. Some studies have verified the feasibility of swarm behavior on quadcopter platforms by embedding the swarm control law into the outer loop position loop of the UAVs.

[0004] However, most of the aforementioned studies are based on ideal communication conditions and the assumption of a point mass model. In practical quadcopter drone swarm systems, drones exchange state information via a wireless ad hoc network. Communication bandwidth and transmission rate are constrained by both physical hardware and the electromagnetic environment. The problem of communication resource scarcity becomes particularly prominent when the swarm size increases or the state update frequency rises. Information quantization technology, by mapping continuous state quantities to finite-level quantized values ​​before transmission, can effectively reduce the number of bytes required for a single communication, providing an effective way to alleviate communication bandwidth pressure and improve channel utilization. This technology has been systematically studied in the fields of networked control systems and multi-agent consensus. On the other hand, existing swarm control methods typically ignore the physical size of the drones when designing control laws, treating each drone as a geometric point mass and only requiring the distance between drones to be greater than zero. For quadcopter drones carrying wingspans or fuselage sizes that are not negligible, this approach poses a risk of collisions between drones in dense formations or high-maneuver flight scenarios. Directly applying the aforementioned point mass model swarm control algorithm is difficult to meet actual flight safety requirements.

[0005] Therefore, existing quadcopter drone swarm control methods do not simultaneously consider the quantization and transmission of state information under limited communication bandwidth conditions and the collision avoidance problem between drones under physical size constraints, making it difficult to meet the dual requirements of communication resource conservation and flight safety. Summary of the Invention

[0006] To address the shortcomings of existing quadrotor UAV swarm swarm control methods in terms of limited communication bandwidth and non-negligible physical size, this paper proposes a quadrotor UAV swarm swarm swarm control method based on quantized communication under communication bandwidth constraints, enabling UAV swarms to achieve collision avoidance swarm flight under quantized state information interaction conditions.

[0007] To achieve the above technical objectives, this application specifically employs the following technical solution: In one aspect of this application, a method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints is provided, comprising the following steps: Step 1: Initialize the quadcopter drone swarm system, establish a six-degree-of-freedom kinematics and dynamics model for each quadcopter drone, decompose the six-degree-of-freedom kinematics and dynamics model into an outer-loop position subsystem and an inner-loop attitude subsystem, and integrate them into a compact outer-loop position model and an inner-loop attitude model. Step 2: Based on the dynamics and kinematics characteristics of quadrotor UAVs, construct a hierarchical swarm control scheme for quadrotor UAVs based on cooperative expected swarm trajectory generation and tracking; Step 3: Set the communication range according to the communication capability of each quadcopter drone, obtain the communication weight between each quadcopter drone and its neighboring drones, and establish the communication connection relationship between each quadcopter drone based on algebraic graph theory. Step 4: Based on the hierarchical swarm control scheme, introduce a virtual agent for each quadcopter drone, construct a multi-agent system composed of all the virtual agents, and design a cooperative expected swarm trajectory generation algorithm based on quantized information; each quadcopter drone uses quantized relative position information and quantized relative velocity information to calculate the control input of each virtual agent, and integrates the control input of each virtual agent to obtain the expected position and expected velocity of each quadcopter drone; Step 5: Based on the desired position and desired velocity, use the desired position and desired velocity as inputs to the underlying position tracking controller. The desired trajectory is tracked by the position tracking controller and attitude tracking controller of the quadcopter UAV to realize swarm behavior under communication quantization constraints.

[0008] In one implementation, in step one, the outer ring position subsystem and the inner ring attitude subsystem are:

[0009]

[0010] In the formula, For drone index subscripts, For the first The weight of a quadcopter drone Represents gravitational acceleration. Indicates quadcopter drone The position variable of the centroid relative to the ground coordinate system. , , These represent the roll angle, pitch angle, and yaw angle, respectively. respectively gyroplane The moment of inertia about each axis of the machine body, so as to... Taking the rolling motion of the shaft as an example, in the formula... The inter-axis angular rate coupling term causes the quadcopter UAV The control input for pose state changes is: ,in, This represents the total thrust generated by all rotors. , The sub-tables represent the layout around the aircraft body. , The differential thrust generated by the shaft This indicates that the aircraft is surrounded by air. The force generated by the shaft.

[0011] In one implementation, in step one, the compact outer ring model is:

[0012] In the formula, , For the first The orthogonal rotation matrix from the body coordinate system to the ground coordinate system of the quadcopter UAV. Indicates quadcopter drone The velocity variable of the center of mass relative to the ground coordinate system; The compact form of the attitude inner loop model is as follows:

[0013] In the formula, This is the angular velocity transfer matrix. For the first The rotational inertia matrix of a quadcopter drone, For the first Torque control input for a quadcopter drone. denoted as the angular velocity vector of the quadcopter UAV.

[0014] In one implementation, in step two, the hierarchical congestion control scheme includes a collaborative estimation layer and a bottom-level control layer; In the cooperative estimation layer, a virtual second-order agent is introduced, with the same number of agents as the quadcopter drone swarm. Each quadcopter drone is associated with one of these virtual agents. Each quadcopter drone calculates its control input to its corresponding virtual agent, and obtains the quantized relative position and velocity information of each quadcopter drone interacting with its neighboring drones. The control input of each virtual agent is then integrated to obtain the desired position and velocity of each quadcopter drone. Under the control law of cooperative desired swarm trajectory generation, each virtual agent tracks the virtual leader agent or moving target, and no collisions occur between the virtual agents. In the underlying control layer, each quadcopter drone uses the trajectory of the virtual agent as the desired trajectory, and uses the generated desired position and desired velocity as inputs to the position controller. The desired trajectory is tracked by the position tracking controller and attitude tracking controller at the bottom layer of the quadcopter drone.

[0015] In one implementation, in step three, each quadcopter drone is simplified to a radius of... Sphere, quadcopter drone The radius of perception is Its neighborhood set is defined as:

[0016] In the formula, This represents the set of all quadcopter drones; at any given time, each quadcopter drone shares information only with its neighboring drones within its neighborhood set.

[0017] In one implementation, in step four, the... The dynamic model of the virtual agent corresponding to the quadcopter drone is as follows:

[0018] In the formula, Indicates the first The position of the virtual intelligent agent, i.e., the first The desired location for the quadcopter drone; Indicates the first The speed of the first virtual agent, i.e., the speed of the second virtual agent. The expected speed of a quadcopter drone; This is the control input to be designed.

[0019] In one implementation, in step four, the multi-agent system includes a virtual navigation agent that acts as the tracking target, and the trajectory of the virtual navigation agent is represented as:

[0020] In the formula, Let be the position vector of the virtual navigation agent. Let V be the velocity vector of the virtual navigation agent. This is the control input for the virtual navigation agent.

[0021] In one implementation, in step four, each quadcopter UAV uses a quantizer to quantize the interaction information, wherein the quantizer is:

[0022] In the formula, , Represents an absolutely quantized sequence. and A quantization constant greater than 0. This is the sector boundary of the quantizer.

[0023] In one implementation, in step four, each quadcopter UAV uses quantized relative position information and quantized relative velocity information. Control input of a virtual intelligent agent Designed as follows:

[0024] In the formula, Indicates the first A quadcopter drone A collection of drones from neighboring locations at any given moment; Indicates the first The communication relationship between the quadcopter drone and the virtual navigation agent. Indicates the first There is a communication connection between the quadcopter drone and the virtual navigation agent. Indicates the first There is no communication connection between the quadcopter drone and the virtual navigation agent. ; This is the vector integration form of the quantizer; Let gravitational / repulsive potential energy function be used. For communication weighting function, ; The gravitational / repulsive potential energy function Represented as:

[0025] in, basic functions for:

[0026] Parameters satisfy , , S-shaped saturation function, impulse function The definition is as follows: ; In the formula, This indicates the physical radius of the quadcopter drone. To achieve the desired distance between drones when forming a stable swarm, The sensing radius of the quadcopter drone. This is the upper threshold of the full-operation range of the designed impact function. For the normalized distance variable between drones, express -norm; Communication weight function Represented as:

[0027] In the formula, This indicates the relative positions between the two drones. Indicates the first The neighbor set of a quadcopter drone , Indicates the first Location of the drone.

[0028] In one implementation, the virtual navigation agent is a target moving at a constant speed, satisfying... .

[0029] The beneficial effects of this application are as follows: 1. This application proposes a distributed swarm control hierarchy framework for quadrotor UAV swarms, which is a fully distributed control method. Essentially, it is an integrated control strategy for swarm swarming and trajectory tracking, which enables quadrotor UAV swarms to achieve swarming behavior. 2. The swarm control algorithm proposed in this application takes into account the actual physical size of the quadcopter UAV, designs a safe collision avoidance distance, meets the actual collision avoidance requirements between UAVs, enables UAV swarms to achieve collision-free aggregation, and solves the inter-UAV collision problem that may occur when the traditional mass model swarm control method is applied to actual swarm systems. 3. This application takes into account the communication bandwidth limitations of quadcopter UAVs, improves the traditional Reynolds swarm model, and designs a cooperative expected swarm trajectory generation algorithm based on quantized information. The artificial potential field term, velocity matching term, and navigation feedback term in the control law all adopt quantized relative state information, which saves communication bandwidth and makes up for the shortcomings of existing technologies in this regard. Attached Figure Description

[0030] Figure 1 A quadcopter UAV and its control variables in a three-dimensional coordinate system; Figure 2 This is a block diagram of a hierarchical architecture for quadrotor cluster congestion control based on expected congestion trajectory generation and tracking proposed in this application; Figure 3 This is the communication connection topology of a quadcopter UAV cluster established according to the initial position distribution of the cluster in this embodiment of the invention; Figure 4 This is a trajectory diagram of a quadcopter drone swarm tracking a uniformly moving target in scenario 1 of the present invention. Figure 5 To provide the various drones in scenario 1 of the present invention, Three-axis velocity variation curves; Figure 6 The diagram shows the velocity vectors and communication topology of each UAV at different simulation times in scenario 1 according to an embodiment of the present invention. Figure 7 The minimum and maximum distance curves between any two drones in scenario 1 according to the present invention; Figure 8 This is a trajectory diagram of a quadcopter drone swarm tracking a variable-speed target in scenario 2 of the present invention. Figure 9 To provide the various drones in scenario 2 of the present invention, Three-axis velocity variation curves; Figure 10 The following is a diagram showing the velocity vectors and communication topology of each UAV at different simulation times in scenario 2 according to an embodiment of the present invention; Figure 11 The minimum and maximum distance curves between any two drones in scenario 2 of the present invention are shown. Detailed Implementation

[0031] The technical solution of this application will be clearly and completely described below with reference to specific embodiments. However, those skilled in the art will understand that the embodiments described below are only some embodiments of this application, not all embodiments, and are only used to illustrate this application, and should not be regarded as limiting the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0032] This application provides a swarm control method for quadrotor UAVs based on quantized communication under communication bandwidth constraints. On the one hand, considering the complex dynamic coupling relationship of quadrotor UAVs, it is necessary to solve the problem of adapting the swarm control algorithm to the motion of the quadrotor UAV swarm. On the other hand, since the communication bandwidth of quadrotor UAVs is limited and their physical size cannot be ignored, it is necessary to further consider the design problem of collision-free swarm control algorithm for swarm system under communication constraints.

[0033] In one specific implementation, a method for controlling swarm congestion of quadrotor UAVs based on quantized communication under communication bandwidth constraints includes the following steps: Step 1: Initialize the UAV swarm system, establish a six-degree-of-freedom kinematic and dynamic model of the quadcopter UAV, decompose it into an outer-loop position subsystem and an inner-loop attitude subsystem, and integrate them into a compact outer-loop position model and an inner-loop attitude model.

[0034] Specifically, a quadcopter drone swarm system includes The state variables of a drone in a three-dimensional coordinate system are as follows: Figure 1 As shown. A quadcopter UAV is a typical nonlinear, strongly coupled, multivariable underactuated system, containing 4 control inputs and 6 degrees of freedom output. Considering the UAV is only subject to gravity and propeller lift, neglecting air resistance, a quadcopter UAV model is established based on Newton-Euler's method. The six-degree-of-freedom kinematic and dynamic model specifically includes the linear motion equations of the center of mass and the angular motion equations of rotation about the center of mass: (1) (2) In the formula, For drone index subscripts, For the quality of drones, Represents gravitational acceleration. Indicates quadcopter drone The position variable of the centroid relative to the ground coordinate system. , , These represent the roll angle (around the fuselage). (axis), pitch angle (around the fuselage) (axis), yaw angle (around the fuselage) axis), respectively gyroplane The moment of inertia about each axis of the machine body, so as to... Taking the rolling motion of the shaft as an example, in the formula... This is the inter-axis angular rate coupling term. It causes [the following] in quadcopter drones. The control input for pose state changes is: ,in, This represents the total thrust generated by all rotors. , The sub-tables represent the layout around the aircraft body. , The differential thrust generated by the shaft This indicates that the aircraft is surrounded by air. The force generated by the shaft.

[0035] The compact form of the outer ring model is as follows: (3) In the formula, ; Indicates quadcopter drone The velocity variable of the center of mass relative to the ground coordinate system; For rotorcraft The orthogonal rotation matrix from the body coordinate system to the ground coordinate system is specifically expressed as: (4) The compact form of the attitude inner loop model is as follows: (5) In the formula, For the first The rotational inertia matrix of a quadcopter drone; For the first Torque control input for a quadcopter drone; Here is the angular velocity vector of the quadcopter UAV; angular velocity transfer matrix. Represented as: (6) For a quadcopter drone swarm system, the designed swarm control law based on quantized communication should enable each drone in the swarm to achieve the following control objectives: Considering that the system contains a virtual navigation agent or moving target Its motion state variables can be described as And there are: (7) The definition of swarm behavior in a quadcopter drone swarm is as follows: (1) When each drone converges to the same velocity, we have: .

[0036] When a cluster system has a virtual lead drone or a moving tracking target, the speed of each drone should converge to the speed of the virtual lead drone or the moving tracking target, i.e., it should satisfy:

[0037] (2) The distance between any two UAVs is bounded, that is, there exists a bounded positive constant. ,have:

[0038] (3) No collision occurs between any two UAVs, i.e., there exists a bounded positive constant. ,have:

[0039] In the formula, To consider the physical radius of quadcopter drones The safe collision avoidance distance between drones was designed later.

[0040] Furthermore, for the quadcopter drone swarm system described in step one, its initialization state satisfies: (1) For all quadcopter drones Its attitude angle Both are bounded, that is, they have , Therefore, the angular velocity transfer matrix It is a non-singular matrix; (2) Under initial conditions, no collisions will occur between the drones, i.e., for ,have ; (3) The expected trajectory of the virtual navigator or moving target linear velocity and acceleration Both are bounded.

[0041] Step 2: Based on the dynamics and kinematics characteristics of quadrotor UAVs, a hierarchical congestion control scheme for quadrotor UAVs based on cooperative expected congestion trajectory generation and tracking is proposed.

[0042] A hierarchical swarm control architecture for quadrotor drone swarms based on desired trajectory tracking, such as... Figure 2 As shown, the system specifically includes a cooperative estimation layer and a bottom-level control layer. In the cooperative estimation layer, a virtual second-order agent, the same number as the number of drones in the swarm, is introduced. There is a one-to-one correspondence between each drone and a virtual agent. Each drone is responsible for calculating the control input to its corresponding virtual agent. The cooperative expected swarm trajectory generation algorithm module adopts a distributed estimation approach. Each drone interacts with its neighboring drones to obtain quantized relative position and velocity information, thus obtaining the control input to the virtual agent. After integration, the expected position and velocity of each drone are obtained. Under the control law of the cooperative expected swarm trajectory generation, each virtual agent can, on the one hand, track the virtual leader or moving target, with its velocity consistently converging to the moving speed of the leader or target; on the other hand, no collisions occur between virtual agents, meaning the virtual agents exhibit swarm behavior at the swarm level. In the bottom-level control layer, each drone uses the trajectory of the virtual agent as its expected trajectory and generates the expected position. With expected speed As input to the position controller, the desired trajectory is tracked through the position and attitude tracking controller at the bottom layer of the quadcopter UAV, ultimately exhibiting swarming behavior at the swarm level.

[0043] Step 3: Based on the communication capabilities of the drone, set its communication range, obtain the communication weights between the drone and its neighboring drones, and establish the communication connection relationship between the quadcopter drones based on algebraic graph theory.

[0044] Considering an environment with no communication delay, no noise, and no interference, the quadcopter drone is simplified to have a radius of... The sphere, all quadcopter drones have the same sensing capabilities, quadcopter drones The radius of perception is Its neighborhood set is defined as: (8) In a quadcopter drone swarm system, all drones have the same perception radius. At any given moment, a quadcopter drone only shares information with its neighboring drones within its vicinity.

[0045] Step 4: Based on the hierarchical swarm control scheme, introduce a virtual agent for each UAV, construct a multi-agent system composed of all virtual agents, design a cooperative expected swarm trajectory generation algorithm based on quantized information, use the quantized relative position and relative speed information of each UAV to calculate the control input of each virtual agent, and obtain the expected trajectory of each UAV after integration.

[0046] With quadcopter drones The corresponding second-order virtual agent node The dynamic model is as follows: (9) In the formula, Representing virtual intelligent agent nodes Position and speed, i.e., drone The expected position and expected velocity, This is the control input to be designed.

[0047] For multi-agent systems, the trajectory of the introduced virtual navigator or moving target is represented by a second-order integrator model, as follows: (10) In the formula, It is the position and velocity vector of the virtual navigation agent. This represents the control input for the virtual navigation agent.

[0048] To reduce communication load and save transmission bandwidth, the UAV adopts a logistic quantization communication mode, specifically the following logistic quantizer: (11) in, , Represents an absolutely quantized sequence. and For the quantization constant greater than 0 to be designed, the quantizer... The quantization density is . This is called the sector boundary of the quantizer.

[0049] The quantization error of the quantizer can be expressed as: (12) Right now Located in about sector-shaped bounded interval Inside.

[0050] Consider the virtual navigator or target moving at a constant speed, i.e. Based on the quantized relative state information and considering the safe distance between drones, the first... Control input of a virtual intelligent agent Designed as follows: (13) In the formula, Indicates drone exist A collection of neighboring drones at any moment. For drones The communication weight between them, if ,but ,otherwise ; Indicates the first The communication relationship between the quadcopter drone and the virtual navigation agent. Indicates the first There is a communication connection between the quadcopter drone and the virtual navigation agent. Indicates the first There is no communication connection between the quadcopter drone and the virtual navigation agent. ; For the vector integration form of the quantizer in formula (11), the basic function The parameters satisfy , To ensure Impact function The definition is as follows: (14) Gravitational / Repulsive Potential Energy Function Represented as: (15) in, This indicates the physical radius of the quadcopter drone. To achieve the desired distance between drones when forming a stable swarm, The sensing radius of the quadcopter drone. The upper bound threshold of the full action range of the designed impulse function; the expression for the communication weight is: (16) In the formula, This indicates the relative positions between the two drones. Indicates the first The neighbor set of a quadcopter drone , Indicates the first Location of the drone.

[0051] Based on the control input obtained in step four, each quadcopter drone calculates the position and velocity of its corresponding virtual agent in a distributed manner, i.e., the desired position of each drone. With expected speed .

[0052] Step 5: Based on the motion trajectory of the virtual intelligent agent generated in Step 4, each UAV uses the desired position and desired speed as input to the position controller. Through the position tracking and attitude tracking controller at the bottom layer of the quadcopter UAV, the desired trajectory is tracked, and the cluster realizes swarm behavior under the communication quantization constraint.

[0053] Specifically, each drone will determine its desired location. With expected speed As input to the underlying position tracking controller, the desired trajectory is tracked through the position tracking and attitude tracking controller of the quadcopter UAV. By repeating steps four and five, the cluster can achieve collective swarming behavior under the control scheme and algorithm, that is, achieve the cluster swarming control target defined in step one.

[0054] Example This embodiment discloses a method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints. The specific steps include: Step 1: Initialize the drone swarm system, which consists of 30 quadcopter drones, initially deployed randomly at... Within the designated area, at a certain moment, the drone swarm receives instructions from the command center to quickly take off from the ground, achieve collision-free aggregation, and rapidly track the trajectory of the virtual navigation agent / target.

[0055] In this embodiment, the quadcopter UAV is positioned in a three-dimensional coordinate system as follows: Figure 1 As shown, quadcopter drone The compact outer loop position model and inner loop attitude model are as follows:

[0056] .

[0057] The basic physical parameters of the quadcopter UAV used in this embodiment are shown in Table 1.

[0058] Table 1 Physical parameters of the quadcopter UAV

[0059] Preferably, for a quadcopter drone swarm system, the designed distributed hierarchical congestion control scheme and the congestion control law based on quantization communication enable each drone in the swarm to achieve the following control objectives: (1) When each drone converges to the same velocity, we have:

[0060] When a cluster system has a virtual lead drone or a moving tracking target, the speed of each drone should converge to the speed of the virtual lead drone or the moving tracking target, i.e., it should satisfy: .

[0061] (2) The distance between any two drones is bounded, that is, there exists a bounded positive constant. ,have: .

[0062] (3) No collision occurs between any two drones, i.e., there exists a bounded positive constant. ,have: .

[0063] Step Two: Based on the strong coupling and nonlinear characteristics of quadcopter UAVs, a hierarchical swarm control architecture for quadcopter UAV swarms based on desired trajectory tracking is proposed. This architecture mainly includes a cooperative estimation layer and a bottom-level control layer. Specifically, in the cooperative estimation layer, a virtual second-order agent is introduced, the same number as the number of UAVs in the swarm. There is a one-to-one correspondence between each UAV and the virtual agent. Each UAV is responsible for calculating the control input to its corresponding virtual agent. The cooperative desired swarm trajectory generation algorithm module adopts a distributed estimation approach. Each UAV interacts with its neighboring UAVs to obtain the quantized relative position and velocity information, thus obtaining the control input to the virtual agent. After integration, the desired position and velocity of each UAV are obtained. Under the control law of the cooperative desired swarm trajectory generation, each virtual agent can track the virtual leader or moving target, with its velocity consistently converging to the moving speed of the leader or target. Furthermore, no collisions occur between virtual agents, meaning that the virtual agents exhibit swarm behavior at the swarm level. Then, in the bottom-level control layer, each UAV uses the trajectory of the virtual agent as its desired trajectory and generates the desired position. With expected speed As input to the position controller, the desired trajectory is tracked through the underlying position and attitude tracking controllers of the quadrotor UAV, ultimately exhibiting swarming behavior at the swarm level. Overall, the distributed control block diagram of the quadrotor UAV swarm swarm swarm swarm swarm control proposed in this embodiment is as follows: Figure 2 As shown.

[0064] according to Figure 2The distributed swarm control architecture shown implements swarm control for a quadcopter drone swarm. Specifically, the moving target is a virtual leader agent with a second-order integrator model, as detailed below: (17) In the formula, It is the position and velocity vector of the virtual navigation agent. This represents the control input for the virtual navigation agent.

[0065] A virtual intelligent agent corresponding to a quadcopter drone is introduced, and its dynamic model is as follows: (18) In the formula, Representing virtual intelligent agent nodes Position and speed, i.e., drone The expected position and expected velocity, This is the control input to be designed in step four.

[0066] Step 3: Establish the UAV's communication topology based on its communication capabilities. (Communication radius of a quadcopter UAV) If drone and drones The distance between them is less than Then drone and drones They are neighbors and have ,otherwise Similarly, if drones If information about the virtual navigation agent can be obtained, then... ,otherwise Based on the initial position distribution of the quadcopter drone swarm in step one, the initial communication topology of the quadcopter drone swarm is as follows: Figure 3 As shown in the diagram. The blue hollow circles represent the initial positions of the drones, the red asterisks represent the initial positions of the virtual agents, the blue lines connecting the nodes indicate communication connections between the drones, and the red dashed lines represent the communication relationships between the drones and the virtual navigation agent. Figure 3 It can be seen that the communication topology between the drones is connected at the initial moment, and no collision occurs in the system at the initial moment.

[0067] Step 4: Considering the communication bandwidth constraints of the quadcopter UAV, the UAV adopts a logarithmic communication method, as follows: (19) Based on quantized communication and Reynolds' three principles of swarm control, the design of each virtual agent node is carried out. The integrated control inputs for congestion control and trajectory tracking are as follows: (20) For the above quantizer, the quantization parameter in equation (19) In equation (20), , Gravity / repulsion function In the formula , , Impact function as follows: (twenty one) In the formula, .

[0068] Based on the control input obtained in step four (20), each virtual agent integrates according to equation (18) to obtain the desired position of each UAV. and expected speed .

[0069] Step 5: Based on the motion trajectory of the virtual intelligent agent generated in Step 4, each UAV uses the desired position and desired speed as input to the bottom-level position controller of the quadcopter UAV. Through the bottom-level position tracking and attitude tracking controller of the quadcopter UAV, the desired trajectory is tracked, and the cluster achieves collective swarming behavior.

[0070] To better illustrate the effectiveness of the quadrotor UAV swarm control method based on quantization communication proposed in this application, the method is then verified through simulation, as detailed below: The invention was verified through simulation using MATLAB Simulink. In the simulation, the UAV swarm system was set to include 30 quadcopter UAVs, all of which were homogeneous. Specific parameters are shown in Table 1. The detailed flowchart of the quadcopter UAV swarm congestion control algorithm based on cooperative desired congestion trajectory generation used in this embodiment is shown in Table 2.

[0071] Table 2. Flowchart of Quadrotor UAV Swarm Control Algorithm Based on Cooperative Desired Swarm Trajectory Generation and Tracking

[0072] Scene 1 The initial position of the virtual navigation agent is It is in three directions respectively , , The simulation operates at a constant velocity in a straight line for 200 seconds. The simulation results are as follows: Figures 4-7 As shown. Figure 4The simulation demonstrates the motion trajectory of a quadcopter drone swarm in three-dimensional space. The dashed short lines represent the trajectory of the virtual navigator, the dotted lines represent the trajectories of the virtual agents (i.e., the desired trajectories of each drone), and the solid lines represent the actual trajectories of each drone. It can be seen that each virtual agent achieves consistent tracking of the virtual navigator or target, and each drone can track the motion trajectory of its corresponding virtual agent, verifying the effectiveness of the proposed control scheme based on desired swarm trajectory generation and tracking. Figure 5 For each virtual intelligent agent relative to the virtual navigator in The three-axis velocity tracking curves show that, as the simulation runtime increases, the speed of each virtual agent keeps pace with the speed of the virtual navigator / target, and the virtual agents achieve consistent tracking of the virtual navigator. Figure 6 Showing , , and The position and velocity vectors of each UAV at any given time and their communication connections are shown. The direction and length of the arrows represent the direction and magnitude of the speed of each UAV, respectively. It can be seen that the speed of the UAV swarm is randomly distributed at the beginning. Under the control scheme and algorithm, the speed of each UAV gradually becomes consistent, and the UAVs gather around the virtual navigator to track the virtual navigator. The swarm forms a stable cluster and exhibits swarming behavior at the overall level. Figure 7 The graph shows the maximum / minimum distance changes between drones during the simulation. As can be seen, the maximum distance between drones tends to be bounded over time, indicating that the drones are clustered together without any "falling behind" behavior. As the simulation progresses, the minimum distance between drones tends to the set expected distance and is strictly greater than the set safe distance throughout the simulation. No collisions occur between drones, and the cluster forms a stable, relatively clustered unified whole.

[0073] Scene 2 To verify the tracking effectiveness of the method proposed in this embodiment for time-varying targets, the system also includes 30 quadcopter UAVs, whose initial positions are randomly distributed. Within the area. At the start of the simulation, each quadcopter drone climbs at a constant speed, in... Upon receiving a target search command, the quadcopter drone swarm is instructed to execute the search task along an S-shaped curve path. Considering the variable speed movement of the virtual navigation agent, the swarm trajectory generation control law in this embodiment is expected to take the following form: (twenty two) The basic parameters and control law parameters of the quadcopter UAV are consistent with those in Example 1. The simulation results are as follows: Figure 8-11As shown in the figure. It can be seen that under the swarm control architecture based on desired trajectory generation and tracking proposed in this embodiment, the quadcopter drone swarm achieves consistent tracking of the virtual navigator / target, and each quadcopter drone achieves consistent tracking of the desired trajectory generated by the virtual agent. Figure 8 The drones achieved consistent speed. Figure 9 , Figure 10 Furthermore, the maximum and minimum distances between drones tend to stabilize, and the minimum distance between drones meets the requirements for safe collision avoidance. Figure 11 The cluster achieves stable flight with a stable inter-aircraft distance, and the cluster exhibits swarming behavior.

[0074] In summary, the control scheme and algorithm proposed in this application, based on cooperative expected swarm trajectory generation and tracking, can achieve swarm control of quadrotor UAV swarms under information quantization constraints. The proposed quadrotor UAV swarm ... This invention can provide a reference for the swarming and formation control of other unmanned systems (unmanned vehicles, unmanned ships, etc.).

[0075] Although the embodiments of this application have been described above in conjunction with the accompanying drawings, this application is not limited to the specific embodiments and application fields described above. The specific embodiments described above are merely illustrative and instructive, not restrictive. Those skilled in the art can make many other forms based on the guidance of this specification and without departing from the scope of protection of the claims of this application, and these are all within the scope of protection of this application.

Claims

1. A method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints, characterized in that, Includes the following steps: Step 1: Initialize the quadcopter drone swarm system, establish a six-degree-of-freedom kinematics and dynamics model for each quadcopter drone, decompose the six-degree-of-freedom kinematics and dynamics model into an outer-loop position subsystem and an inner-loop attitude subsystem, and integrate them into a compact outer-loop position model and an inner-loop attitude model. Step 2: Based on the dynamics and kinematics characteristics of the quadrotor UAV, construct a hierarchical swarm control scheme for the quadrotor UAV based on cooperative expected swarm trajectory generation and tracking; Step 3: Set the communication range according to the communication capability of each quadcopter drone, obtain the communication weight between each quadcopter drone and its neighboring drones, and establish the communication connection relationship between each quadcopter drone based on algebraic graph theory. Step 4: Based on the hierarchical swarm control scheme, introduce a virtual agent for each quadcopter drone, construct a multi-agent system composed of all the virtual agents, and design a cooperative expected swarm trajectory generation algorithm based on quantized information; each quadcopter drone uses quantized relative position information and quantized relative velocity information to calculate the control input of each virtual agent, and integrates the control input of each virtual agent to obtain the expected position and expected velocity of each quadcopter drone; Step 5: Based on the desired position and desired velocity, use the desired position and desired velocity as inputs to the underlying position tracking controller. The desired trajectory is tracked by the position tracking controller and attitude tracking controller of the quadcopter UAV to realize swarm behavior under communication quantization constraints.

2. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step one, the outer ring position subsystem and the inner ring attitude subsystem are: In the formula, For drone index subscripts, For the first The weight of a quadcopter drone Represents gravitational acceleration. Indicates quadcopter drone The position variable of the centroid relative to the ground coordinate system. , , These represent roll angle, pitch angle, and yaw angle, respectively. respectively gyroplane The moment of inertia about each axis of the machine body, so as to... Taking the rolling motion of the shaft as an example, in the formula... The inter-axis angular rate coupling term causes the quadcopter UAV The control input for pose state changes is: ,in, This represents the total thrust generated by all rotors. , The sub-tables represent the layout around the aircraft body. , The differential thrust generated by the shaft This indicates that the aircraft is surrounded by air. The force generated by the shaft.

3. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 2, characterized in that, In step one, the compact outer ring model is as follows: In the formula, , For the first The orthogonal rotation matrix from the body coordinate system to the ground coordinate system of the quadcopter UAV. Indicates quadcopter drone The velocity variable of the center of mass relative to the ground coordinate system; The compact form of the attitude inner loop model is as follows: In the formula, This is the angular velocity transfer matrix. For the first The rotational inertia matrix of a quadcopter drone, For the first Torque control input for a quadcopter drone. denoted as the angular velocity vector of the quadcopter UAV.

4. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step two, the hierarchical congestion control scheme includes a collaborative estimation layer and a bottom-level control layer; In the collaborative estimation layer, a virtual second-order agent is introduced with the same number as the number of quadcopter drones in the swarm, and there is a one-to-one correspondence between each quadcopter drone and one of the virtual agents. Each quadcopter drone calculates the control input to its corresponding virtual agent, and obtains the quantized relative position and velocity information of each quadcopter drone interacting with neighboring drones to get the control input of each virtual agent. The desired position and desired velocity of each quadcopter drone are obtained by integrating the control input of each virtual agent. Under the action of the cooperative desired swarm trajectory generation control law, each virtual agent tracks the virtual leader agent or moving target, and no collision occurs between the virtual agents. In the underlying control layer, each quadcopter drone uses the trajectory of the virtual agent as the desired trajectory, and uses the generated desired position and desired velocity as inputs to the position controller. The desired trajectory is tracked by the position tracking controller and attitude tracking controller at the bottom layer of the quadcopter drone.

5. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step three, each quadcopter drone is simplified to a radius of... Sphere, quadcopter drone The radius of perception is Its neighborhood set is defined as: In the formula, This represents the set of all quadcopter drones; at any given time, each quadcopter drone shares information only with its neighboring drones within its neighborhood set.

6. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step four, with the first The dynamic model of the virtual agent corresponding to the quadcopter drone is as follows: In the formula, Indicates the first The position of the virtual intelligent agent, i.e., the first The desired location for the quadcopter drone; Indicates the first The speed of the first virtual agent, i.e., the speed of the second... The expected speed of a quadcopter drone; This is the control input to be designed.

7. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step four, the multi-agent system includes a virtual navigation agent that acts as the tracking target, and the trajectory of the virtual navigation agent is represented as follows: In the formula, Let be the position vector of the virtual navigation agent. Let V be the velocity vector of the virtual navigation agent. This is the control input for the virtual navigation agent.

8. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step four, each quadcopter UAV uses a quantizer to quantize the interactive information. The quantizer is: In the formula, , Represents an absolutely quantized sequence. and A quantization constant greater than 0. This is the sector boundary of the quantizer.

9. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 1, characterized in that, In step four, each quadcopter drone uses quantized relative position information and quantized relative velocity information. Control input of a virtual intelligent agent Designed as follows: In the formula, Indicates the first A quadcopter drone A collection of drones from neighboring locations at any given moment; Indicates the first The communication relationship between the quadcopter drone and the virtual navigation agent. Indicates the first The quadcopter drone has a communication connection with the virtual navigation agent. Indicates the first There is no communication connection between the quadcopter drone and the virtual navigation agent. ; This is the vector integration form of the quantizer; Let gravitational / repulsive potential energy function be used. For communication weighting function, ; The gravitational / repulsive potential energy function Represented as: in, basic functions for: Parameters satisfy , , S-shaped saturation function, impulse function The definition is as follows: ； In the formula, This indicates the physical radius of the quadcopter drone. To achieve the desired distance between drones when forming a stable swarm, The sensing radius of the quadcopter drone. This is the upper threshold of the full-operation range of the designed impact function. For the normalized distance variable between drones, express -norm; Communication weight function Represented as: In the formula, This indicates the relative positions between the two drones. Indicates the first The neighbor set of a quadcopter drone , Indicates the first Location of the drone.

10. The method for swarm control of quadrotor UAVs based on quantized communication under communication bandwidth constraints according to claim 7, characterized in that, The virtual navigation agent is a target moving at a constant speed, satisfying the following conditions: .

Images