Heterogeneous multi-agent formation control method based on composite disturbance observer

Abstract

The invention discloses a heterogeneous multi-agent formation control method based on a composite disturbance observer, and the method comprises the steps: building a nonlinear model of heterogeneous multi-agents based on the composite disturbance observer, and enabling the agents to comprise an unmanned plane and an unmanned vehicle; the composite interference observer is used for estimating unknown composite interference of each agent; a heterogeneous multi-agent intelligent plane formation controller with an internal safe collision avoidance function is constructed based on a nonlinear model, a conversion error function is introduced to restrain an initial collaborative error in a specified interval, and a safe distance between two adjacent agents is ensured; meanwhile, a preset time height formation controller is constructed, so that the consistency of expected flight heights of the unmanned aerial vehicles is realized; and controlling the heterogeneous multi-agent formation according to the constructed controller. The adverse effect caused by unknown interference is effectively eliminated through the composite interference observer, the collision avoidance problem between the intelligent agents is solved by using a specified performance control method, and meanwhile, the system is ensured to be stable within a preset time and the control precision is improved.

Description

Technical Field

[0001] The present invention relates to intelligent agent formation control allocation, and in particular to a heterogeneous multi-agent formation control method based on a composite interference observer. Background Art

[0002] A heterogeneous multi-agent system composed of multiple types of agents (such as multiple quad-rotor drones and two-wheel drive unmanned vehicles) has better collaborative working capabilities than a single type of multi-agent system. Unmanned vehicles have the ability to carry cargo and are mainly used to detect the ground environment and transmit information to drones in a timely manner. When facing complex environments, such as obstacles on the ground or blocked communication signals, unmanned vehicles alone cannot detect comprehensive information, while multiple drones can use their own flight characteristics to obtain a wide range of environmental information and communicate with unmanned vehicles in a timely manner. Due to the inconsistent dimensions of the mathematical models of quad-rotor drones and two-wheel drive unmanned vehicles, the formation control design problem is more complicated than that of homogeneous systems.

[0003] Heterogeneous multi-UAV unmanned vehicle formations may be subject to various unknown interferences and uncertainties during driving. These interferences and uncertainties will affect the stability of the formation system. It is difficult to measure directly or the cost of measurement is too high. Early research mainly focused on improving the anti-interference ability of the system itself. In addition, more advanced anti-interference technologies such as control methods based on interference observers are widely used in the research of formation systems. However, the anti-interference accuracy of the composite external interference during the operation of the heterogeneous system formation is not enough.

[0004] At present, the problem of collision avoidance in platoon driving has become a popular research direction. Since the distance between adjacent intelligent agents needs to be limited to a certain range to ensure that there is enough space for emergencies, it is considered to specify performance control to constrain the coordination error to ensure the safety of the platoon operation and avoid collisions. The scheduled time stability performance discusses the transient performance of the controlled system, that is, to make the controlled system reach an equilibrium state within a predetermined time. The system needs to satisfy Lyapunov stability and reach a stable state in a short time. As we all know, the scheduled time controller contains exponential power terms. Compared with the infinite time control of the closed-loop system, the scheduled time control has higher accuracy and stronger anti-disturbance ability. Therefore, it is necessary to study the scheduled time formation control problem on the basis of ensuring the safety of the distance between intelligent agents. Summary of the invention

[0005] Purpose of the invention: In view of the above shortcomings, the present invention provides a heterogeneous multi-agent formation control method based on a composite disturbance observer with high control accuracy and good stability.

[0006] Technical solution: To solve the above problems, the present invention adopts a heterogeneous multi-agent formation control method based on a composite interference observer, comprising the following steps:

[0007] (1) A nonlinear model of heterogeneous multi-agents based on a composite disturbance observer is established, wherein the agents include unmanned aerial vehicles (UAVs) and unmanned vehicles (UAVs). The composite disturbance observer is used to estimate the unknown composite disturbance of each agent, wherein the composite disturbance includes the disturbance of external airflow on the agent, the coupling disturbance of surrounding neighbors, and the fuzzy approximation error.

[0008] (2) Based on the nonlinear model, a heterogeneous multi-agent intelligent plane formation controller with internal safety collision avoidance is constructed. In this process, a distributed collaborative error control objective is designed, collaborative error constraints and a predetermined time specified performance function are defined, and a conversion error function is introduced to constrain the initial collaborative error within a specified interval to ensure a safe distance between two adjacent agents. At the same time, a predetermined time height formation controller is constructed to ensure that each UAV achieves the consistency of the desired flight altitude.

[0009] (3) Control the heterogeneous multi-agent formation according to the constructed heterogeneous multi-agent intelligent plane formation controller and the predetermined time altitude formation controller.

[0010] Furthermore, the nonlinear model of heterogeneous multi-agents in step (1) is:

[0011]

[0012] Among them, p mi ,q mi denote the position and velocity of the ith m-type agent, d mi Represents the external composite interference, u mi represents the control input, f mi is the unknown nonlinear term of the heterogeneous multi-agent system;

[0013] Use fuzzy logic system to approximate unknown nonlinear terms:

[0014]

[0015] Among them, θ i is the ideal weight vector, T is the transpose of the vector, The fuzzy basis function satisfies and ε i (ζ i ) is the fuzzy approximation error and satisfies is a normal number.

[0016] Furthermore, the composite interference is specifically expressed as:

[0017]

[0018] Among them, d αi (t) represents the interference of external airflow on the intelligent agent; d βi (t) represents the coupling interference of the neighbors around the agent; a ij Represents the Laplace matrix of topological connections in the formation. When the i-th agent and the j-th agent are neighbors, a ij =1, otherwise a ij =0, and i≠j; λ1>0, k1>0 are adjustable parameters; ||d ij ||=||p mi -p mj || represents the distance between the i-th agent and the j-th agent.

[0019] Furthermore, based on the composite disturbance, the auxiliary variable b is defined as i (t), and obtain the interference estimate as follows:

[0020]

[0021] Among them, K i >0 and ψ>0 are both positive constants, and Respectively represent d mi and θ i k1=(π / 2 1-γ / 2 γT s ), k2=(π / 2 1+γ / 2 γT s ) are two control gains, e mi represents the interference estimation error, 0<γ<1, T s >0,sig β (·)=|· β sgn(·).

[0022] Furthermore, the distributed collaborative error formula in step (2) is:

[0023]

[0024] Among them, d i and d j denote the displacement vectors of the ideal positions of the i-th and j-th agents and the leader formation, respectively, p0 denotes the position of the leader; b i represents the communication status between the ith agent and the leader;

[0025] The collaborative error constraints are:

[0026] -δ min σi (t)<e i (t)<δ max σ i (t)

[0027] Among them, δ min >0,δ max >0 are two adjustable parameters, σ i is a predetermined time-specified performance function;

[0028] The scheduled time specified performance function is:

[0029]

[0030] Among them, σ0>σ ∞ >0 is a constant, T s >0 is the maximum allowed convergence time that can be predefined, η i >0 indicates a preset constant that can adjust the convergence speed;

[0031] Transform the cooperative error:

[0032] e i (t) = Δ(o i (t))σ i (t)

[0033]

[0034] Among them, i is a new conversion error and Δ(o i ) is a smooth and strictly increasing function.

[0035] Furthermore, the heterogeneous multi-agent intelligent plane formation controller includes a plane virtual controller α i1 (t) and the plane actual controller u mi (t):

[0036]

[0037] in, q0 represents the speed of the leader, z i1 is the transformed coordination error, is to use the intermediate control function α i1 and the state variables obtained by the first-order filter.

[0038] Furthermore, the second-order model of the UAV z-axis subsystem is established as:

[0039]

[0040] Furthermore, in step (2), the backstepping method is used to complete the design of the corresponding controller of the altitude system of the UAV, and the coordinate transformation design is as follows:

[0041]

[0042] Among them, p d1 is the desired height, q d1 is the desired speed, α zi is the virtual control quantity;

[0043] The scheduled time altitude formation controller includes an altitude virtual controller α zi (t) and the height actual controller u fzi (t):

[0044] α zi (t) = -k1sig 1-γ (s z1 (t))-k2sig 1+γ (s z1 (t))

[0045]

[0046] Where g is the acceleration due to gravity.

[0047] The present invention also adopts a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the above method when executing the computer program.

[0048] The present invention also adopts a computer-readable storage medium on which a computer program is stored, characterized in that the computer program implements the steps of the above method when executed by a processor.

[0049] Beneficial effects: Compared with the prior art, the significant advantage of the present invention is that it effectively eliminates the adverse effects caused by unknown interference through a composite interference observer, solves the collision avoidance problem between intelligent agents using a specified performance control method, and ensures that the system reaches stability within a predetermined time and improves control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the heterogeneous multi-agent formation control process of the present invention. DETAILED DESCRIPTION

[0051] like Figure 1 As shown, a heterogeneous multi-agent formation control method based on a composite interference observer in this embodiment specifically includes the following steps:

[0052] Step 1: Establish a nonlinear model of heterogeneous multi-agents based on a composite disturbance observer. The agents include drones and unmanned vehicles. According to the different dynamic characteristics and different dimensions of movable space of the heterogeneous multi-agent system composed of quadrotor drones and two-wheel drive unmanned vehicles, a unified nonlinear model of the heterogeneous multi-drone and unmanned vehicle system is established.

[0053] Taking into account the influence of air resistance on the quadrotor drone, the position system of the i-th quadrotor drone can be expressed as:

[0054]

[0055] Among them, p fi =[p fxi ,p fyi ,p fzi ] T ,U fi =[u fxi ,u fyi ,u fzi ] T ,d fi =[d fxi ,d fyi ,d fzi ] T They represent the position, control quantity and unknown external disturbance of the quadrotor drone, respectively, and f fi represents the unknown nonlinear term of the position system, f fi The aerodynamic drag coefficient k included in 1i , k 2i , k 3i is unknown, and the nonlinear term and control quantity can be expanded as follows:

[0056]

[0057] Among them, φ i ,θ i , i are the roll angle, pitch angle and yaw angle of the i-th UAV, respectively, M ai represents the mass of the i-th UAV. g is the acceleration due to gravity. i1 is the total thrust in the direction perpendicular to the fuselage.

[0058] Next, we build a two-wheel drive unmanned vehicle model. Assuming that it has a rigid structure, it can be regarded as a rigid body during the motion process. We only consider the pure rolling motion of the two-wheeled unmanned vehicle on the two-dimensional plane, and there will be no slipping between the two-wheeled unmanned vehicle and the ground. Considering the non-holonomic constraint problem, the position system of the i-th two-wheeled unmanned vehicle can be expressed as:

[0059]

[0060] Among them, p si =[p sxi ,p syi ] T is the location information of the two-wheeled unmanned vehicle, d si =[d sxi ,d syi ] T Indicates unknown external interference, U si represents the control input, f si represents the unknown nonlinear term, which is expanded as follows:

[0061]

[0062] Among them, V i is the linear velocity, θ i is the direction angle, ω i is the angular velocity. i Represents the distance between the front of the i-th two-wheeled unmanned vehicle and the midpoint of the two wheels.

[0063] Combining the above-mentioned positioning systems of the quadcopter and the two-wheeled unmanned vehicle, it can be summarized into a unified form. The planar second-order model of heterogeneous multi-agents is expressed as follows:

[0064]

[0065] Among them, p mi ,q mi denote the position and velocity of heterogeneous multi-agents, d mi represents external composite interference, u mi represents the control input, f mi is the unknown nonlinear term of the system. Definition And m∈{f,s} includes all drones and unmanned vehicles. The model of the virtual leader can be expressed as follows:

[0066]

[0067] where p0, q0 and u0 represent the position, velocity and control input of the leader, respectively.

[0068] The second-order model of the UAV z-axis subsystem is established as:

[0069]

[0070] Among them, p fzi ,q fzi ,u fzi They are the position signal, speed signal, and control signal of the drone in the z-axis direction.

[0071] The unknown nonlinear terms contained in equation (7) are approximated using the fuzzy logic system:

[0072]

[0073] Among them, θ i is the ideal weight vector, The fuzzy basis function satisfies and ε i (ζ i ) is the fuzzy approximation error and satisfies is a normal number.

[0074] The composite disturbance observer is used to estimate the unknown composite disturbance of each agent, which includes the disturbance of external airflow on the agent, the coupling disturbance of surrounding neighbors and the fuzzy approximation error.

[0075] Considering the impact of the composite interference consisting of the interference of external airflow on the i-th agent, the secondary interference on the surrounding neighbors and the fuzzy approximation error on the heterogeneous multi-agent system, the composite interference is specifically expressed as follows:

[0076]

[0077] Among them, a ij Represents the Laplace matrix of the topological connection in the formation. When the i-th agent and the j-th agent are neighbors, the relationship is a ij =1, otherwise a ij =0, and i≠j. λ1>0, k1>0 are adjustable parameters, ||d ij ||=||p mi -p mj || represents the distance between the ith agent and the jth agent. According to the above formula, the influence of the interference of the neighboring agents on the ith agent is d βi , it can be seen that the distance between the agents is related to the coupling interference d βi Closely related. When ||d ij When || is relatively large, it means that the two agents are far apart and the interference from the neighbors has little impact on the i-th agent. On the contrary, when ||d ij When || is relatively small, it means that the distance between the two agents is relatively close, and the influence of the neighbor’s interference on the i-th agent will become larger. c ≤||d ij ||≤r d ,[r c ,r d ] is the safe distance between adjacent intelligent agents to avoid collision and communicate normally.

[0078] Combined with the composite interference d in formula (11)mi The nonlinear compound disturbance observer for compound disturbances estimates the coupled disturbances to each agent as accurately as possible. Define the auxiliary variable b i (t), design interference estimate as follows:

[0079]

[0080] Among them, K i >0 and ψ>0 are both positive constants, and Respectively represent d mi and θ i k1=(π / 2 1-γ / 2 γT s ), k2=(π / 2 1+γ / 2 γT s ) are two control gains, e mi is the interference estimation error, 0<γ<1, T s >0,sig β (·)=|·| β sgn(·).

[0081] Next, define Combining the above formula and The interference estimation error e mi Taking the derivative we get:

[0082]

[0083] Step 2: Based on the nonlinear model, a heterogeneous multi-agent intelligent plane formation controller with internal safety collision avoidance is constructed. The distributed collaborative error control objective is designed, the collaborative error constraint conditions and the scheduled time specified performance function are defined, and the conversion error function is introduced to constrain the initial collaborative error within the specified interval to ensure the safe distance between two adjacent agents. At the same time, a scheduled time height formation controller is constructed to enable each UAV to achieve consistency in the expected flight altitude.

[0084] For the heterogeneous multi-UAV unmanned vehicle system in the two-dimensional xoy plane, a distributed collaborative error control target is designed, and the collaborative error constraint conditions and the scheduled time specified performance function are defined. Then, the conversion error function is introduced to constrain the initial collaborative error within the specified interval to ensure the safe distance between two adjacent agents. On this basis, a heterogeneous multi-agent scheduled time formation controller with internal safety collision avoidance is designed, and the stability of the designed closed-loop error system is proved by the Lyapunov stability theory, and the convergence time of the collaborative error is given.

[0085] Define the coordination error constraint and the scheduled time to specify the performance function. The distributed coordination error for the i-th follower can be designed as:

[0086]

[0087] Among them, d i and d j denote the displacement vectors of the ideal positions of the i-th and j-th agents and the leader formation, respectively, and b i Represents the communication between the ith agent and the leader. If the follower can receive the leader’s message b i =1, otherwise b i = 0. In order to ensure that the tracking error of the heterogeneous multi-agent system remains within the specified boundaries and has ideal transient and steady-state performance, the following constraints are constructed:

[0088] -δ min σ i (t)<e i (t)<δ max σ i (t) (15)

[0089] Among them, δ min >0,δ max >0 are two adjustable parameters and σ i is the specified performance function. The smoothing function σ i The following conditions are met: (1) σ i >0; (2) (3) σ Ts is a pre-designed positive constant. When t ≥ T s When i (t) = σ Ts The scheduled time specified performance function is defined as follows:

[0090]

[0091] Among them, σ0>σ ∞ >0 is a constant, T s >0 is the maximum allowed convergence time that can be predefined, η i >0 represents a preset constant that can adjust the convergence speed. Transforming the tracking error yields:

[0092] e i (t) = Δ(o i (t))σ i (t) (17)

[0093] Among them, i is a new conversion error and Δ(oi ) is a smooth and strictly increasing function, defined as follows:

[0094]

[0095] For i Taking the time derivative we get:

[0096]

[0097] It can be deduced that i The time derivative of is:

[0098]

[0099] Before designing the scheduled time formation controller, the following lemmas and assumptions are used.

[0100] Lemma 1: There exists a positive definite function V and a constant 0<γ<1, T s >0, d>0, if the nonlinear system satisfy:

[0101]

[0102] Then the nonlinear system is the actual scheduled time to be stable, and V can enter and stay in the region:

[0103]

[0104] And the scheduled time T' s Meet T' s ≤2T s .

[0105] Lemma 2: For positive constants p>0, q>0, η>0 real variables x and y, we have:

[0106]

[0107] Lemma 3: For y ≥ x and ρ>1, we have:

[0108]

[0109] Assumption: Due to external disturbance d αi The energy is limited. Their first-order derivatives are also bounded and satisfy and d1 are both positive constants. ||d ij ||≥r c Can ensure βi is bounded, and the fuzzy approximation error is also bounded. Therefore, the composite interference and the first-order derivative are bounded, satisfying

[0110] The control goal in this embodiment is that the two-dimensional time-varying formation of a heterogeneous multi-UAV system in the xoy plane can be characterized by a set of differentiable variables. s >0, the following conditions are met:

[0111]

[0112] The formation can then avoid internal collisions and achieve the desired formation within the predetermined time.

[0113] For the z-axis subsystem of the drone, the desired height and speed are If satisfied:

[0114]

[0115] It is considered that a high degree of consistency has been achieved.

[0116] Considering the existence of external unknown interference and nonlinear terms in heterogeneous multi-UAV unmanned vehicle systems, coordinate transformation is defined, the predetermined time stability theory is combined with the backstepping recursion technology, a nonlinear filter with predetermined time parameters and adaptive parameters is constructed, and a fuzzy adaptive predetermined time formation controller is designed. The Lyapunov stability theory is used to prove that all signals in the system are bounded, and the formation coordination error converges to the specified performance limit within the predetermined time. Next, the control design of the two-dimensional xoy plane and the altitude control of the UAV are carried out.

[0117] Xoy plane model control

[0118] In order to facilitate the implementation of the back-stepping recursion method, the coordinate transformation is first defined as follows:

[0119]

[0120] In the formula, z i1 is the transformed coordination error, is to use the intermediate control function α i1 and the state variable obtained by the first-order filter, ρ i2 is the output error of the first-order filter.

[0121] Design the virtual controller, actual controller and adaptive law as follows:

[0122]

[0123] in, p θ =π(2+γ) / 2γT s β1 γ / 2 (1+γ),qθ =(π / γT s ) 2 / 2-γ , β1>0.

[0124] For a heterogeneous multi-UAV unmanned vehicle system (7)-(9) that considers the problem of safe obstacle avoidance within a formation, under the action of the designed virtual controller (29), actual controller (30), adaptive law (31) and composite disturbance observer (12), the tracking error can converge to The formation closed-loop system can achieve stability within the predetermined time, and collision avoidance can be achieved between intelligent agents.

[0125] Verify as follows:

[0126] Step 1: Transform the coordinates of z i1 Taking the derivative we get:

[0127]

[0128] Choose the following Lyapunov function:

[0129]

[0130] V i1 The time derivative is:

[0131]

[0132] Substituting the virtual controller into the above formula, we can get:

[0133]

[0134] By the intermediate control function α i1 and a first-order filter, designing a new state variable as follows:

[0135]

[0136] According to the output error defined And the above formula, we can get:

[0137]

[0138] Among them, M1 is a continuous function with the following structure:

[0139]

[0140] in, and

[0141] Step 2: For z in (28) i2Taking the derivative we get:

[0142]

[0143] The Lyapunov function is selected as follows:

[0144]

[0145] Among them, β1>0 and β2>0 are design parameters. i2 Taking the time derivative we get:

[0146]

[0147] Substituting (30)-(31) into (41), we can obtain:

[0148]

[0149] Using Young's inequality we can get:

[0150]

[0151] Substituting (43)-(48) into (42), we can obtain the overall Lyapunov function:

[0152]

[0153] According to Lemma 2 and Lemma 3, we can get:

[0154]

[0155] Substituting (50)-(51) into (49), the overall Lyapunov function can be rearranged to obtain:

[0156]

[0157] in

[0158]

[0159] According to the above analysis, It can be further shortened to:

[0160]

[0161] According to Lemma 1, the tracking error is s '≤2T s It can converge to:

[0162]

[0163] Drone altitude control

[0164] According to the second-order model of the UAV z-axis subsystem, the backstepping method is used to complete the design of the corresponding controller of the UAV's altitude system. The coordinate transformation design is as follows:

[0165]

[0166] The desired height is p d1 , the expected speed is p d2 , the virtual control amount is α zi .

[0167] Design the virtual controller and the actual controller as follows:

[0168] α zi (t) = -k1sig 1-γ (s z1 (t))-k2sig 1+γ (s z1 (t)) (57)

[0169]

[0170] Based on the control of the virtual controller and the actual controller, each UAV can achieve the same expected flight height.

[0171] Verify as follows:

[0172] Step 1: Transform s into coordinates z1 The derivative is:

[0173]

[0174] Select the Lyapunov function as follows:

[0175]

[0176] Taking the derivative of the above formula, we can get:

[0177]

[0178] Substituting the virtual control quantity (57) into (61), we can obtain:

[0179]

[0180] Step 2: Transform the coordinates z2 The derivative is:

[0181]

[0182] Select the Lyapunov function as follows:

[0183]

[0184] Taking the derivative of the above formula, we can get:

[0185]

[0186] Substituting the actual control quantity (58) into (65), we can obtain:

[0187]

[0188] According to Lemma 1, we can analyze

Claims

1. A heterogeneous multi-agent formation control method based on a composite disturbance observer, characterized in that: The following steps are involved: (1) Establishing a nonlinear model of heterogeneous multi-agents based on a composite disturbance observer, wherein the agents include unmanned aerial vehicles and unmanned vehicles; The composite disturbance observer is used to estimate the unknown composite disturbance of each intelligent agent, wherein the composite disturbance includes the disturbance of external airflow to the intelligent agent, the coupling disturbance of surrounding neighbors and the fuzzy approximation error; (2) Based on the nonlinear model, a heterogeneous multi-agent intelligent plane formation controller with internal safety collision avoidance is constructed. In this process, a distributed collaborative error control objective is designed, collaborative error constraints and a predetermined time specified performance function are defined, and a conversion error function is introduced to constrain the initial collaborative error within a specified interval to ensure a safe distance between two adjacent agents. At the same time, a predetermined time height formation controller is constructed to ensure that each UAV achieves the consistency of the desired flight altitude. (3) Control the heterogeneous multi-agent formation according to the constructed heterogeneous multi-agent intelligent plane formation controller and the predetermined time altitude formation controller.

2. The heterogeneous multi-agent formation control method according to claim 1, characterized in that: The nonlinear model of heterogeneous multi-agents in step (1) is: Among them, p mi ,q mi denote the position and velocity of the ith m-type agent, d mi Represents the external composite interference, u mi represents the control input, f mi is the unknown nonlinear term of the heterogeneous multi-agent system; Use fuzzy logic system to approximate unknown nonlinear terms: Among them, θ i is the ideal weight vector, T is the transpose of the vector, The fuzzy basis function satisfies and ε i (ζ i ) is the fuzzy approximation error and satisfies is a normal number.

3. The heterogeneous multi-agent formation control method according to claim 2, characterized in that: The composite interference is specifically expressed as: Among them, d αi (t) represents the interference of external airflow on the agent; d βi (t) represents the coupling interference of the neighbors around the agent; a ij Represents the Laplace matrix of topological connections in the formation. When the i-th agent and the j-th agent are neighbors, a ij =1, otherwise a ij =0, and i≠j; λ1>0, k1>0 are adjustable parameters; ||d ij ||=||p mi -p mj || represents the distance between the i-th agent and the j-th agent.

4. The heterogeneous multi-agent formation control method according to claim 3 is characterized in that: Based on the composite disturbance, define the auxiliary variable b i (t), and obtain the interference estimate as follows: Among them, K i >0 and ψ>0 are both positive constants, and Respectively represent d mi and θ i k1=(π / 2 1-γ / 2 γT s ), k2=(π / 2 1+γ / 2 γT s ) are two control gains, e mi represents the interference estimation error, 0<γ<1, T s >0,sig β (·)=|·| β sgn(·).

5. The heterogeneous multi-agent formation control method according to claim 4 is characterized in that: The distributed collaborative error formula in step (2) is: Among them, d i and d j denote the displacement vectors of the ideal positions of the i-th and j-th agents and the leader formation, respectively, p0 denotes the position of the leader; b i represents the communication status between the ith agent and the leader; The collaborative error constraints are: -d min s i (t)<e i (t)<δ max s i (t) Among them, δ min >0,δ max >0 are two adjustable parameters, σ i is a predetermined time-specified performance function; The scheduled time specified performance function is: Among them, σ0>σ ∞ >0 is a constant, T s >0 is the maximum allowed convergence time that can be predefined, η i >0 indicates a preset constant that can adjust the convergence speed; Transform the cooperative error: e i (t)=Δ(o i (t))σ i (t) Among them, i is a new conversion error and Δ(o i ) is a smooth and strictly increasing function.

6. The heterogeneous multi-agent formation control method according to claim 5, characterized in that: The heterogeneous multi-agent intelligent plane formation controller includes a plane virtual controller α i1 (t) and the plane actual controller u mi (t): in, q0 represents the speed of the leader, z i1 is the transformed coordination error, is to use the intermediate control function α i1 and the state variables obtained by the first-order filter.

7. The heterogeneous multi-agent formation control method according to claim 6, characterized in that: The second-order model of the UAV z-axis subsystem is established as:

8. The heterogeneous multi-agent formation control method according to claim 5, characterized in that: In step (2), the backstepping method is used to complete the design of the corresponding controller of the altitude system of the UAV, and the coordinate transformation design is as follows: Among them, p d1 is the desired height, q d1 is the desired speed, α zi is the virtual control quantity; The scheduled time altitude formation controller includes an altitude virtual controller α zi (t) and the height actual controller u fzi (t): α zi (t)=-k1sig 1-γ (s z1 (t))-k2sig 1+γ (s z1 (t)) Where g is the acceleration due to gravity.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images