Six-rotor unmanned aerial vehicle control method and system with disturbance and performance constraints

Abstract

The invention discloses a six-rotor unmanned aerial vehicle control method and system with disturbance and performance constraints. The method comprises the steps that 1, a six-rotor unmanned aerial vehicle power model is built; 2, errors are determined based on the dynamic model, functions are combined, and a finite time scheme is determined; step 3, based on the dynamic model, establishing an observer, and knowing a disturbance state; step 4, a performance control scheme and an unknown disturbance state are stipulated in finite time, an error dynamic model is established, and the error dynamic model is established based on reinforcement learning; 5, setting a base scheme, setting a virtual controller and an actual controller, and setting an adaptive law for the base neural network; 6, the controller follower can track the leader, and tracking control is completed; establishing a controller by adopting a backstepping method, establishing a disturbance observer, and compensating negative effects; a reinforcement learning algorithm and a backstepping method technology are combined, and a controller is arranged to realize control; according to the method, the stability is proved by using a Lyapunov judgment stability theory, and the feasible effectiveness is verified by a simulation example.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-rotor UAV control, and particularly to a control method and system for a six-rotor UAV with disturbances and performance constraints. Background Art

[0002] In recent years, multi-rotor unmanned aerial vehicles (UAVs) have been widely used in disaster monitoring, rescue missions, and military reconnaissance due to their advantages such as hovering, high mobility, and agility. A six-rotor UAV system (UAVs) is usually divided into a position subsystem and an attitude subsystem. The present invention mainly focuses on the research of the position subsystem. A six-rotor UAV is a very complex non-linear system, with uncertainties such as external disturbances, actuator failures, and dead zones. Therefore, it is of great significance to study the modeling of a six-rotor UAV system, consider uncertainties and non-linearity, and then realize the motion trajectory planning of the UAV under different tasks.

[0003] Since non-linearity is the most common and essential feature of dynamic systems in the real world, non-linear control has been widely applied in various fields and has received increasing attention in the past decade. As is well known, non-linear strict-feedback systems are the most common model forms of engineering systems. Therefore, the control of such systems is a valuable research topic. In recent years, several control methods have been proposed for non-linear strict-feedback systems. However, these methods first convert the high-order system into a compact form and then design the control by treating the converted system as a normal system. Therefore, they cannot guarantee that the entire backstepping control sequence is optimal. It is worth mentioning that a new control technique has been proposed, called optimized backstepping (OB) for non-linear strict-feedback systems. Its basic idea is to design the control of each step back as the optimal solution of the corresponding subsystem, so as to optimize the control of the entire system. However, several common defects of optimal control have not been eliminated, namely: 1) The reinforcement learning (RL) update rule is very complex and intractable; 2) It is assumed that the system dynamics are known. Therefore, this technique is difficult to apply to real-world engineering.

[0004] For the tracking control problem of multi-UAV systems, good control performance and small residuals are required; however, it is difficult to determine the residual set without using the trial-and-error method, and the size of the residual set is often difficult to determine; thus, the prescribed performance control method, as one of the effective methods to avoid this problem, has received extensive attention from scholars; however, it is worth noting that when considering the finite-time control method, the backstepping control method will have singularity problems, which encourages us to design a suitable method to improve the transient performance of multi-UAVs with prescribed performance.

[0005] In practical applications, UAVs are often subject to interference from the environment, resulting in a decline in system performance; if these adverse effects are not properly compensated, the performance of UAVs will be severely reduced, and even cause the instability of UAVs; to enhance the anti-interference performance of the system, the disturbance observer can avoid the problem that the existence of disturbances may make it difficult for multiple UAVs to achieve the tracking control performance, and improve the robustness of the hexacopter UAV system; at the same time, considering that factors such as harsh environments, limited resources, and equipment losses of UAVs will affect the control performance of the system and have a certain impact on the completion of control tasks, the control of UAV systems is also worthy of further research.

[0006] In summary, it is important to consider the external environmental interference problem in practice, and it has certain practical significance to study the optimal tracking control of hexacopter UAVs on this basis; in addition, studying the control problem of UAV systems based on reinforcement learning and disturbance observers helps to improve the system control performance, and at the same time provides a theoretical basis for the optimal tracking control of UAV systems in complex environments. Summary of the Invention

[0007] The purpose of the present invention is to provide a control method and system for a hexacopter UAV with disturbance and performance constraints to solve the problems existing in the above-mentioned prior art.

[0008] A control method for a hexacopter UAV with disturbance and performance constraints includes the following steps:

[0009] Step 1: Establish a dynamic model of a hexacopter UAV with lumped disturbances;

[0010] Step 2: Based on the dynamic model, determine the synchronization error, and combine the prescribed performance function with the velocity function to determine the finite-time prescribed performance control scheme;

[0011] Step 3: Based on the dynamic model, establish a disturbance observer to determine the unknown composite disturbance state information of the hexacopter UAV;

[0012] Step 4: Construct an error dynamic model according to the finite-time prescribed performance control scheme and the unknown composite disturbance state information, and establish a control scheme for the six-rotor UAV based on the error dynamic model by using the reinforcement learning method;

[0013] Step 5: Design a virtual controller and an actual controller based on the control scheme, and design an unknown parameter adaptive law based on a neural network;

[0014] Step 6: According to the virtual controller, the actual controller and the unknown parameter adaptive law, enable the follower of the six-rotor UAV system to track the leader, and complete the tracking control of the six-rotor UAV system.

[0015] Preferably, establishing the dynamic model of the six-rotor UAV system with lumped disturbances in step 1 specifically includes:

[0016] Define the angular velocity vector and the centroid linear velocity vector of the dynamic model in the body coordinate system of a fixed object, and define the attitude angle vector and the position vector of the dynamic model in the inertial coordinate system of the fixed earth;

[0017] Establish a dynamic model of the six-rotor UAV system according to the angular velocity vector, the centroid linear velocity vector, the attitude angle vector and the position vector. The dynamic model is:

[0018]

[0019] where Y = [l, m, n] T is the position vector in the roll, pitch and yaw directions in the position subsystem, is the derivative of Y, T is the symbol for the transpose operation; Γ = [ν l , ν m , ν n T is the centroid linear velocity vector in three directions, is the derivative of the centroid linear velocity vector Γ; represents the force generated by the motor, M represents the mass, P1 and P2 represent transformation matrices, G represents the gravitational acceleration; the matrix Ξ = diag{e1, e2, e3}, where e1, e2 and e3 are air resistance coefficients; the matrix E3 = [0, 0, 1] T ; ) represents the attitude angle vector in the roll, pitch and yaw directions in the attitude subsystem, is the derivative of Λ, is the angular velocity vector in three directions; H = diag{j l , j m , j n} is the moment of inertia in three directions, The gyroscopic torques in three directions; B is the torque generated by six rotors, and D represents the disturbance vector;

[0020] The system dynamic equation of the i-th follower in the dynamic model is:

[0021]

[0022] where, represents the state of the i-th follower system, x i,1 is the first state of the i-th follower system, x i,2 is the second state of the i-th follower system; x i,1 = [x i,11 , x i,21 , x i,31 T , x i,2 = [x i,12 , x i,22 , x i,32 T where x i,11 = l i , x i,21 = m i and x i,31 = n i are the position vectors of the i-th follower system in the roll, pitch, and yaw directions respectively; x i,12 = ν li , x i,22 = ν mi and x i,32 = ν ni are the linear velocity vectors of the center of mass of the i-th follower system in the roll, pitch, and yaw directions respectively; represents the output of the i-th follower system, represents an unknown function; represents the lumped disturbance, u i represents the control input; is the derivative of x i,1 ; is the derivative of x i,2 .

[0023] Preferably, in step 1, before establishing the dynamic model of the six-rotor UAV system with lumped disturbance, it further includes: describing the communication topology structure between six-rotor UAVs;

[0024] When describing the communication topology structure between six-rotor UAVs, a directed graph is used to describe the directed communication topology relationship between multiple UAVs, specifically including:

[0025] ​​Use a directed graph \(F=(B, J, A)\) to represent the directed communication topology relationship among multiple UAVs; where \(B = (1,\cdots,N)\) and are non-empty sets of nodes and directed edges respectively, \(N\) is the number of nodes, \(A=[a b,j \in N×N is the relevant adjacency weight matrix, and \(a b,j is the weight between node \(b\) and node \(j\);

[0026] If node \(b\) can receive communication information from node \(j\), \(a b,j >0; otherwise \(a b,j = 0\), and node \(b\) cannot receive communication information from node \(j\);

[0027] (B j ,B b )\in J\) represents an edge from node \(b\) to node \(j\); define the neighbor set of node \(b\) as Then define as the in-degree matrix, \(\theta b is the in-degree of node \(b\), is the Laplacian matrix;

[0028] If node \(b\) can receive the information sent from node \(0\), then \(a b,0 >0, otherwise \(a b,0 = 0\); where \(0\) represents the leader, and \(a b,0 is the weight between node \(b\) and the leader.

[0029] Preferably, based on the dynamic model, determine the synchronization error, specifically including:

[0030] Based on the leader signal given in the dynamic model, construct the synchronization error; the synchronization error is:

[0031]

[0032] where, represents the synchronization error, represents the output of the \(i\)-th follower system of the hexacopter UAV, represents the output of the \(j\)-th follower system of the hexacopter UAV, \(y d represents the output of the leader system of the hexacopter UAV, \(a i,j is the weight between the \(i\)-th follower UAV and the \(j\)-th follower UAV, \(a i,0 is the weight between the \(i\)-th follower UAV and the leader.

[0033] Preferably, in step 2, combine the prescribed performance function with the velocity function to determine the finite-time prescribed performance control scheme, specifically including:

[0034] To ensure that the synchronization error converges to a predetermined neighborhood of the origin and all closed-loop signals are bounded, a prescribed performance is combined with the velocity function to determine a finite-time prescribed performance control scheme; the finite-time prescribed performance control scheme is as follows:

[0035]

[0036] where η(t) is the velocity function, is the prescribed finite time, σ(t) is expressed as a non-decreasing smooth function, t represents time, α is a constant, 0 < α < 1, to ensure is continuous everywhere, define Define as

[0037] Preferably, according to the error dynamic model, a control scheme for a six-rotor UAV is established based on the reinforcement learning method, which also includes:

[0038] Use a radial basis neural network to identify the unknown and uncertain nonlinear dynamics in UAVs; on the compact set Ω x the radial basis neural network W *T χ(x) can approximate the unknown nonlinear function f(x) with arbitrary precision ;

[0039] f(x) = W *T χ(x) + ε(x),

[0040] where x is the neural network input, x ∈ Ω x , ε(x) is the approximation error, is a constant; W * is the ideal weight vector, χ(x) = [χ1, χ2,..., χ p T is the basis function vector, χ i is the Gaussian function, i = 1, 2, 3,..., p, p is the number of radial basis neural network nodes, p > 1, W is the weight vector; p×m is the set of real matrices with p rows and m columns, and m is the number of columns of the matrix.

[0041] Preferably, based on the dynamic model, a disturbance observer is established to estimate the unknown composite disturbance state information of the six-rotor UAV, which specifically includes:

[0042] Use to estimate the unknown composite disturbance state information of the six-rotor UAV;

[0043] where, define​ for yes Estimate of ε fi,2 is the radial basis neural network approximation error, is the lumped disturbance, is a constant, is an auxiliary variable, for The derivative of Is to identify the ideal weights of radial basis neural network The estimate of χ fi,2 is the basis function vector; u i represents the control input, x i,2 is the system state, z i,2 is the virtual tracking error.

[0044] Preferably, based on the control scheme, designing a virtual controller and an actual controller specifically includes:

[0045] In response to the design process of the actual controller, a virtual controller is designed based on the system dynamic equations of the dynamic model; the virtual controller is:

[0046]

[0047] Among them, η i (t) is the velocity function, is η i The derivative of (t), is the synchronization error, λ i (t) is the specified performance function, λ i (t)=(λ i,0 -λ i,∞ )exp(-òt)+λ i,∞ , define λ i,∞ for λ i,0 is λ i,0 =λ i (0)>0, is λ i The derivative of (t), is a smooth function, is the upper bound of the specified performance, is the lower bound of the specified performance, definition

[0048] The goal of specifying performance control is to ensure Converge to the specified bound in; τ i (t) is the conversion error; θ iis the in-degree of node i, N is the number of follower UAVs, j is the neighbor follower of the i-th follower UAV, a i,j is the weight between follower i and neighbor follower j, is the adaptive law of the critic neural network, z i,1 is the state transformation defined according to the specified performance; is the estimation of the ideal weight of the identification neural network χ fi,1 is the basis function vector; is the derivative of the leader output y d ; a i,0 is the weight between follower i and the leader.

[0049] A six-rotor UAV control system with disturbances and performance constraints adopts the six-rotor UAV control method according to any one of the above claims 1-8. The six-rotor UAV control system with disturbances and performance constraints includes:

[0050] A data acquisition module for obtaining the physical characteristics of the six-rotor UAV;

[0051] A data processing module for obtaining the dynamic model of the six-rotor UAV system based on the physical characteristics;

[0052] A control module for constructing an adaptive controller by using the backstepping technique based on the data acquisition module and the data processing module, estimating the unknown composite disturbance state information of the six-rotor UAV by using a disturbance observer, and designing a virtual controller for the six-rotor UAV by combining the backstepping technique and a reinforcement learning algorithm to perform adaptive control design on the six-rotor UAV.

[0053] Preferably, the control module specifically includes:

[0054] An observation and adaptive update law design unit for constructing an adaptive controller by using the backstepping technique, constructing a disturbance observer to observe the unknown composite disturbance state information of the six-rotor UAV, and designing an adaptive update law for identifying-executing-criticizing the neural network weights according to the stability judgment method of the Lyapunov function stability theory;

[0055] [[ID=4(1)]]An adaptive law design unit for designing an adaptive law for unknown nonlinear terms based on the neural network according to the adaptive update law; the neural network is a radial basis neural network;

[0056] An adaptive control unit for performing adaptive control on the six-rotor UAV based on the adaptive law and according to the constructed controller.

[0057] Note: In the translation, for item , the number in the original text seems incorrect. I translated it as [[ID=4(1)]] according to the format you provided. If there is a specific correct number, please let me know and I will adjust it.Compared with the prior art, the present invention provides a control method and system for a six-rotor UAV with disturbances and performance constraints, having the following beneficial effects:

[0058] 1. The technical solution of the present invention proposes a finite-time prescribed performance strategy. Different from the traditional prescribed performance method, the present invention introduces a speed function, which can effectively ensure that the tracking error converges to a prescribed bounded set within a finite time without singularity problems.

[0059] 2. The technical solution of the present invention designs a disturbance observer to compensate for the negative impact of lumped disturbances on the six-rotor UAV system, avoiding the problem that the existence of disturbances may make it difficult for multiple UAVs to achieve tracking control performance, and improving the robustness of the six-rotor UAV system.

[0060] 3. The technical solution of the present invention designs a performance index function that can evaluate the tracking error and energy loss through the learning ability of the reinforcement learning algorithm, ensuring less energy consumption while the system is stable; in addition, by designing a simple positive definite function, an adaptive update law for the evaluation and execution neural networks is constructed, which ensures the smooth execution of the reinforcement learning algorithm.

[0061] Other features and advantages of the present invention will be partially given in the following description, partially will become obvious from the following description, or will be understood through the practice of this aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Figure 1 is the communication topology diagram of the six-rotor UAV system;

[0064] Figure 2 is the tracking control flow chart of the six-rotor UAV system;

[0065] Figure 3 is the trajectory diagram of the six-rotor UAV system to achieve the tracking control task;

[0066] Figure 4 is when i = 1, 2, 3, 4 and when and the trajectory diagram;

[0067] Figure 5Trajectory diagrams of the synchronization error and performance boundaries of the six-rotor UAV system when i = 1, 2, 3, 4;

[0068] Figure 6 Control input curves of the six-rotor UAV system under performance constraints and disturbance observers. Specific implementation manners

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0070] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific implementation manners.

[0071] As Figure 1-6 shown, the present invention provides a control method and system for a six-rotor unmanned aerial vehicle with disturbances and performance constraints. The control structure and overall process of the control system are as Figure 1 and Figure 2 shown;

[0072] Taking the six-rotor unmanned aerial vehicle system as an example, a control method for a six-rotor unmanned aerial vehicle with disturbances and performance constraints will be described in detail. The detailed implementation process includes:

[0073] Step 1: Establish the dynamic model of the six-rotor unmanned aerial vehicle as:

[0074]

[0075] where Y = [l, m, n] T is the position vector in the roll, pitch, and yaw directions in the position subsystem, and Γ = [ν l , ν m , ν n T is the linear velocity vector of the center of mass in three directions; represents the force generated by the motor, M represents the mass, P1 and P2 represent transformation matrices, and G represents the gravitational acceleration. Matrix Ξ = diag{e1, e2, e3}, where e1, e2, and e3 are air resistance coefficients, and matrix E3 = [0, 0, 1] T ; represents the attitude angle vector in the roll, pitch, and yaw directions in the attitude subsystem, is the derivative of Λ, is the angular velocity vector in three directions, and H = diag{jl ,j m ,j n} is the moment of inertia in three directions, It is defined as the gyroscopic torque in three directions, B is the torque generated by the six rotors, and D represents the disturbance vector;

[0076] The system dynamic equation of its i-th follower can be written as:

[0077]

[0078] in, Represents the state of the i-th follower system, and defines x i,1 =[x i,11 ,x i,21 ,x i,31 ] T , x i,2 =[x i,12 ,x i,22 ,x i,32 ] T , where x i,11 =l i , x i,21 =m i and x i,31 =n i are the position vectors of the i-th follower system in the roll, pitch and yaw directions respectively; x i,12 =ν li , x i,22 =ν mi and x i,32 =ν ni are the center-of-mass linear velocity vectors of the i-th follower system in the roll, pitch and yaw directions respectively; represents the system output of the ith follower, Represents an unknown function; Expressed as lumped disturbance, u i Indicates control input.

[0079] Step 2: Design synchronization error:

[0080]

[0081] in, represents the synchronization error, represents the output of the i-th position subsystem of the hexacopter UAV, represents the output of the j-th position subsystem of the hexacopter UAV, y d Represents the leader output of the hexacopter UAV position subsystem.

[0082] Step 3: Combine the specified performance with the speed function and propose a finite-time specified performance control scheme as follows:

[0083]

[0084] Among them, the speed function is defined as η(t), and the specified finite time is defined as σ(t) is expressed as a non-decreasing smooth function, t represents time, and the constant α is defined as 0 < α1, where to ensure is continuous everywhere, define

[0085] Design the virtual tracking error conversion function:

[0086] z i,2 = x i,2 - y i,2 (1.5)

[0087] s i,2 = y i,2 - α i,1 (1.6)

[0088] Among them, z i,2 represents the virtual tracking error, α i,1 is the virtual control signal of the position subsystem, y i,2 represents the output of the filter, and s i,2 is the filter error.

[0089] Step 4: Establish a disturbance observer to estimate the unknown composite disturbance state information of the hexacopter UAV:

[0090]

[0091] Among them, define is 's estimate, ε fi,2 is the neural network approximation error, is the lumped disturbance, define the constant Define as the auxiliary variable, is the estimate of the ideal weight of the identification neural network, χ fi,2 is the basis function vector; ui represents the control input. x i,2 is the system state, and z i,2 is the virtual tracking error.

[0092] Step 5: Combine the backstepping technique and the reinforcement learning algorithm to design a tracking controller;

[0093] According to the system dynamic equation of the dynamic model of the given six-rotor UAV system, design the virtual controller and the actual controller respectively:

[0094] First step, establish the following performance index function:

[0095]

[0096] where, Ω is a compact set containing the origin, Ψ(Ω) is the admissible control set, α i,1 is the virtual controller, is the virtual controller, is the cost function, and z i,1 is the state transformation defined according to the specified performance.

[0097] Differentiate both sides of formula (1.8) to obtain the Hamilton–Jacobi–Bellman (HJB) equation as follows:

[0098]

[0099] By solving obtain the virtual controller as shown below:

[0100]

[0101] Decompose into:

[0102]

[0103] where,

[0104] l i,1 is a positive design parameter;

[0105] Substitute formula (1.11) into formula (1.10) to get:

[0106]

[0107] Use the reinforcement learning algorithm based on the actor-critic neural network to approximate the virtual controller to obtain:

[0108]

[0109] where, is 's estimated value, is 's estimated value; the neural network approximates 's unknown dynamics, is the ideal estimated value for judging the weights of the neural network; the neural network approximates the unknown dynamics in is the ideal estimated value for the weights of the execution neural network; χ Ji,1 is the fuzzy basis function.

[0110] The approximate HJB function is as follows:

[0111]

[0112] Define the Bellman residual as follows:

[0113]

[0114] Step 2: Establish the performance index function as follows:

[0115]

[0116] where, is the cost function, u i is the virtual controller, is the virtual controller.

[0117] The corresponding HJB equation is as follows:

[0118]

[0119] Similar to the first step, an approximate controller is obtained by using the reinforcement learning algorithm:

[0120]

[0121] where, the neural network is used to approximate the unknown dynamics in is the estimated value of the ideal weights of the execution neural network. χ Ji,2 is the fuzzy basis function.

[0122] Step 6: To ensure the smooth execution of the reinforcement learning algorithm, design a simple positive definite function and construct the adaptive update laws for the evaluation and execution neural networks as follows;

[0123]

[0124] where, γ ci,1 and γ ci,2 are the design parameters of the evaluation neural network, γ ai,1 and γ ai,2 are the design parameters of the execution neural network;

[0125] To prove the feasibility, effectiveness, and correctness of this example, the present invention conducts the following simulation experiments:

[0126] In this simulation experiment, for a six-rotor UAV system with disturbances and performance constraints, an adaptive controller based on a disturbance observer and reinforcement learning is designed to achieve the tracking control of the six-rotor UAV system.

[0127] During the controller design process, the system model parameters are set as follows:

[0128] The trajectories of the UAV leader in three directions are respectively: The nonlinear function is: f i,1 (x i,1 ) = 1.9sin(x i,1 ) - 1.4x i,1 , The initial values of the follower system state are:

[0129] The system disturbance is: The relevant parameters of the finite-time prescribed performance are: α = 0.9, λ1(t) = (1 - 0.04)exp(-2t) + 0.04, λ2(t) = λ3(t) = λ4(t) = (2 - 0.04)exp(-2t) + 0.04,

[0130] The evaluation neural network and and the execution neural network and both contain 7 neurons, and the center points of the neural network are evenly distributed on [-3, 3]; the initial weights of the neural network are:

[0131] The parameters related to the virtual and actual controllers are designed as: Λ i,1 = 1.8, Λ i,2 = 1.4, τ i,1 = 3.6, τ i,2 = 2.8; the adaptive law parameters are: γ ai,1 = 18, γ ai,2 = 15, γ ci,1 = 14, γ ci,2 = 13.

[0132] Combined with the accompanying drawings, the effectiveness of the simulation of this example is further illustrated:

[0133] Using the Matrix Laboratory (MATLAB) and Simulation and Link (SIMULINK) software, the mathematical model established in the control method of this embodiment is simulated. Figure 1 In Figure 1 , "leader" means leader, and the following four are followers. Figure 1 Figure 1 depicts the communication relationship between the follower UAV and the leader UAV; the simulation results Figures 3-5 are the simulation results. Figure 4 and Figure 5 In Figure 5 , "Time(s)" means time. Figure 4 In Figure 4 , "Time(sec)" means time, and the unit is seconds. Figure 5 In Figure 5 , "Upper bound" means upper limit or maximum value, representing that the black line from high to low on the upper side of the figure is the numerical upper limit or maximum value. Figure 5 In Figure 5 , "Lower bound" means lower limit or minimum value, representing that the black line from low to high on the lower side of the figure is the numerical lower limit or minimum value. Figure 3 Figure 3 shows the tracking trajectory of the hexacopter UAVs system, achieving precise tracking of the reference signal. Figure 4 Figure 4 shows when i = 1, 2, 3, 4 and when and The trajectory curves of can be seen to track the leader according to the designed control scheme, proving the effectiveness of the strategy proposed in the present invention. Figure 5 Figure 5 shows the trajectory diagram of the synchronization error and performance boundary of the hexacopter UAVs system when i = 1, 2, 3, 4. The synchronization error converges within the finite-time performance boundary range to ensure that all closed-loop signals are bounded. Figure 6 Figure 6 shows the trajectory of the control input under performance constraints and disturbance observers, indicating that the control input signal converges uniformly and successfully realizes the control task of the system.

[0134] In summary, all signals in the system are uniformly ultimately bounded, and the simulation results prove the effectiveness of the proposed tracking control scheme.

[0135] This embodiment takes the backstepping recursion and reinforcement learning techniques as the design framework, and studies the adaptive tracking control problem for a six-rotor UAV system with disturbances and performance constraints. In addition, the present invention assumes that the six-rotor UAVs system has unmeasurable states and unknown disturbance information, making the system more general; by designing a disturbance observer, the simultaneous observation of the system state and disturbance information is realized. This embodiment utilizes the approximation ability of the reinforcement learning neural network for nonlinear functions, uses the reinforcement learning algorithm to approximate the actual controller, solves the difficult problem of directly solving the HJB equation, designs a controller that meets the control requirements, and adjusts the weights of the neural network through an adaptive law, so that the designed approximate controller can approximate the actual controller. Finally, through simulation, it is verified that the proposed adaptive tracking control strategy can ensure that all signals are bounded; the popularization and application of the present invention in the tracking control of six-rotor UAVs is one of the important research directions in the future.

[0136] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention; it should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions; these computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in one Figure 1 process or multiple processes and / or blocks Figure 1 block or multiple blocks.

[0137] In the description of the present invention, it should be understood that the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features; thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0138] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention; thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A control method for a six-rotor UAV with disturbance and performance constraints, characterized in that including the following steps, Step 1: Establish the dynamic model of a six-rotor UAV with lumped disturbances; Step 2: Based on the dynamic model, determine the synchronization error, and combine the prescribed performance function with the velocity function to determine the finite-time prescribed performance control scheme; Step 3: Based on the dynamic model, establish a disturbance observer to determine the unknown composite disturbance state information of the six-rotor UAV; Step 4: According to the finite-time prescribed performance control scheme and the unknown composite disturbance state information, construct an error dynamic model, and based on the error dynamic model, establish a control scheme for the six-rotor UAV using the reinforcement learning method; Step 5: Based on the control scheme, design a virtual controller and an actual controller, and design an unknown parameter adaptive law based on a neural network; Step 6: According to the virtual controller, the actual controller, and the unknown parameter adaptive law, enable the follower of the six-rotor UAV system to track the leader, and complete the tracking control of the six-rotor UAV system.

2. The control method of a six-rotor UAV with disturbances and performance constraints according to claim 1, characterized in that, In Step 1, establishing the dynamic model of the six-rotor UAV system with lumped disturbances specifically includes: Define the angular velocity vector and the centroid linear velocity vector of the dynamic model in the body coordinate system of a fixed object, and define the attitude angle vector and the position vector of the dynamic model in the inertial coordinate system of the fixed earth; According to the angular velocity vector, the centroid linear velocity vector, the attitude angle vector, and the position vector, establish the dynamic model of the six-rotor UAV system, and the dynamic model is: where Y = [l, m, n] T is the position vector in the roll, pitch, and yaw directions of the position subsystem, is the derivative of Y, T is the symbol for the transpose operation; Γ = [ν l , ν m , ν n T is the linear velocity vector of the center of mass in the three directions, is the derivative of the linear velocity vector Γ of the center of mass; represents the force generated by the motor, M represents the mass, P1 and P2 represent transformation matrices, G represents the acceleration due to gravity; the matrix Ξ = diag{e1, e2, e3}, where e1, e2, and e3 are the air resistance coefficients; the matrix E3 = [0, 0, 1] T ; represents the attitude angle vector in the roll, pitch, and yaw directions of the attitude subsystem, is the derivative of Λ, Δ = [ζ l , ζ m , ζ n T is the angular velocity vector in the three directions; H = diag{j l , j m , j n} is the moment of inertia in the three directions, is the gyroscopic torque in the three directions; B is the torque generated by the six rotors, and D represents the disturbance vector;​​ The system dynamic equation of the i-th follower in the dynamic model is: Among them, represents the state of the \(i\)-th follower system, \(x\) i,1 is the first state of the \(i\)-th follower system, \(x\) i,2 is the second state of the \(i\)-th follower system; \(x\) i,1 = [x i,11 , x i,21 , x i,31 T , where \(x\) i,11 = \(l\) i , \(x\) i,21 = \(m\) i and \(x\) i,31 = \(n\) i are the position vectors of the \(i\)-th follower system in the roll, pitch, and yaw directions respectively; \(x\) i,2 = [x i,12 , x i,22 , x i,32 T , \(x\) i,12 = \(ν\) li , \(x\) i,22 = \(ν\) mi and \(x\) i,32 = \(ν\) ni are the centroid linear velocity vectors of the \(i\)-th follower system in the roll, pitch, and yaw directions respectively; represents the output of the \(i\)-th follower system, represents an unknown function; represents a lumped disturbance, \(u\) i represents the control input; is the derivative of \(x\) i,1 ; is the derivative of \(x\) i,2 .​​ 3. A control method for a six-rotor UAV with disturbances and performance constraints according to claim 1, characterized in that, Before Step 1, establishing the dynamic model of the six-rotor UAV system with lumped disturbances also includes: describing the communication topology structure between six-rotor UAVs; When describing the communication topology structure between six-rotor UAVs, use a directed graph to describe the directed communication topology relationship between multiple UAVs, specifically including: Use a directed graph \(F=(B, J, A)\) to represent the directed communication topology relationship among multiple UAVs; where \(B = (1,\ldots,N)\) and are non-empty sets of nodes and directed edges respectively, \(N\) is the number of nodes, \(A=[a b,j \in N×N is the relevant adjacency weight matrix, and \(a b,j is the weight between node \(b\) and node \(j\); If node b can receive communication information from node j, a b,j > 0; otherwise a b,j = 0, and node b cannot receive communication information from node j; (B j , B b ) ∈ J represents an edge from node b to node j; define the neighbor set of node b as Then define as the in-degree matrix, as the in-degree of node b, as the Laplacian matrix; If node b can receive the information sent from node 0, then a b,0 > 0, otherwise a b,0 = 0; where 0 represents the leader, and a b,0 is the weight between node b and the leader.

4. A control method for a six-rotor UAV with perturbations and performance constraints according to claim 1, characterized in that Based on the dynamic model, determining the synchronization error specifically includes: Based on the leader signal given in the dynamic model, construct the synchronization error; the synchronization error is: Among them, represents the synchronization error, represents the output of the i-th follower system of the hexacopter UAV, represents the output of the j-th follower system of the hexacopter UAV, y d represents the output of the leader system of the hexacopter UAV, a i,j is the weight between the i-th follower UAV and the j-th follower UAV, a i,0 is the weight between the i-th follower UAV and the leader.

5. A control method for a six-rotor UAV with perturbations and performance constraints according to claim 1, characterized in that, In Step 2, combining the prescribed performance function with the velocity function to determine the finite-time prescribed performance control scheme specifically includes: To ensure that the synchronization error converges to a predetermined neighborhood of the origin and all closed-loop signals are bounded, combine the prescribed performance with the velocity function to determine the finite-time prescribed performance control scheme; the finite-time prescribed performance control scheme is: where η(t) is the velocity function, is the specified finite time, σ(t) is expressed as a non-decreasing smooth function, t represents time, α is a constant, 0 < α < 1, to ensure continuous everywhere, define Define as 6. A control method for a six-rotor unmanned aerial vehicle with perturbations and performance constraints according to claim 1, characterized in that, Before establishing the control scheme for the six-rotor UAV based on the error dynamic model using the reinforcement learning method, it also includes: Identifying unknown and uncertain nonlinear dynamics in UAVs using a radial basis neural network; on the compact set Ω x the radial basis neural network W *T χ(x) can approximate the unknown nonlinear function f(x) with arbitrary precision ; where \(x\) is the neural network input, \(x\in\Omega\) x , \(\varepsilon(x)\) is the approximation error, is a constant; \(W\) * is the ideal weight vector, \(\chi(x)=[\chi_1,\chi_2,...,\chi\) p T is the basis function vector, \(\chi\) i is the Gaussian function, \(i = 1,2,3,...,p\), \(p\) is the number of nodes of the radial basis neural network, \(p>1\); \(W\) is the weight vector; p×m is the set composed of real matrices of \(p\) rows and \(m\) columns, \(m\) is the number of columns of the matrix.​ 7. A control method for a six-rotor UAV with disturbances and performance constraints according to claim 1, characterized in that, Based on the dynamic model, establish a disturbance observer to estimate the unknown composite disturbance state information of the six-rotor UAV, specifically including: Utilize Estimate the unknown composite disturbance state information of the hexacopter UAV; Among them, it is defined that is is the estimation of, ε fi,2 is the approximation error of the radial basis neural network, is the lumped disturbance, is a constant, is the auxiliary variable, is the derivative of, is the estimation of the ideal weight of the radial basis neural network, χ fi,2 is the basis function vector; u i represents the control input, x i,2 is the system state, z i,2 is the virtual tracking error.

8. A control method for a six-rotor UAV with perturbations and performance constraints according to claim 1, characterized in that Based on the control scheme, designing a virtual controller and an actual controller specifically includes: In response to the design process of the actual controller, design a virtual controller based on the system dynamic equation of the dynamic model; the virtual controller is: Among them, η i (t) is the velocity function, is the derivative of η i (t), is the synchronization error, λ i (t) is the prescribed performance function; λ i (t) = (λ i,0 - λ i,∞ ) exp(-òt) + λ i,∞ , define λ i,∞ as λ i,0 is λ i,0 = λ i (0) > 0, is the derivative of λ i (t), is a smooth function; is the upper bound of the prescribed performance, is the lower bound of the prescribed performance, Definition The goal of the prescribed performance control is to ensure that converges to the prescribed bounds ; τ i (t) is the transformation error; θ i is the in-degree of node i, N is the number of follower UAVs, j is the neighbor follower of the i-th follower UAV, a i,j is the weight between follower i and neighbor follower j, is the critic neural network adaptation law, z i,1 is the state transformation defined according to the prescribed performance; is the estimate of the ideal weight of the identification neural network ; χ fi,1 is the basis function vector; is the derivative of the leader output y d ; a i,0 is the weight between follower i and the leader.

9. A six-rotor UAV control system with disturbance and performance constraints, characterized in that, The six-rotor UAV control system with disturbances and performance constraints adopts a six-rotor UAV control method described in any one of the above claims 1-8. The six-rotor UAV control system with disturbances and performance constraints specifically includes: A data acquisition module for obtaining the physical characteristics of the six-rotor UAV; A data processing module for obtaining the dynamic model of the six-rotor UAV system based on the physical characteristics; A control module for constructing an adaptive controller using the backstepping technique based on the data acquisition module and the data processing module, estimating the unknown composite disturbance state information of the six-rotor UAV according to the disturbance observer, and designing a virtual controller for the six-rotor UAV by combining the backstepping technique and the reinforcement learning algorithm to perform adaptive control design on the six-rotor UAV.

10. The six-rotor UAV control system with disturbance and performance constraints according to claim 9, characterized in that, The control module specifically includes: An observation and adaptive update law design unit for constructing an adaptive controller using the backstepping technique, constructing a disturbance observer to observe the unknown composite disturbance state information of the six-rotor UAV, and designing an adaptive update law for identifying-executing-judging the neural network weights according to the stability judgment method of the Lyapunov function stability theory; An adaptive law design unit for designing an adaptive law for the unknown nonlinear term based on the neural network according to the adaptive update law; the neural network is a radial basis neural network; An adaptive control unit for performing adaptive control on the six-rotor UAV based on the adaptive law and the constructed controller.

Images