Aircraft adaptive control method considering attack angle delay constraint

Abstract

The invention discloses an aircraft adaptive control method considering attack angle delay constraint. The method comprises the following steps: firstly, establishing an aircraft longitudinal dynamic model; then, introducing a novel state conversion function to transform the model, and converting an attack angle delay constraint control problem into a stable control problem of a new state; and finally, designing an adaptive backstepping control strategy based on a neural network for the transformed model to ensure the stability of the system, thereby realizing the delay constraint of the attack angle. The method is high in universality and high in calculation precision, and has important engineering practice reference significance.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft, and in particular relates to an aircraft adaptive control method considering attack angle delay constraints. Background Art

[0002] The flight angle of attack refers to the angle between the aircraft's velocity direction and the aircraft's longitudinal axis. Too large a flight angle of attack can cause the aircraft to stall or even lose control, seriously threatening flight safety. Therefore, it is of great significance to consider the constraints on the angle of attack in the design of the flight control system. The angle of attack constraint control method based on the obstacle Lyapunov function is an effective means to ensure that the angle of attack always remains within a safe range. However, this method requires that the desired angle of attack command and the angle of attack tracking error be restricted at the same time. The algorithm is complex and the results are relatively conservative. In addition, this method requires that the flight angle of attack meets the constraint requirements at the initial moment, and cannot be applied to situations where the initial angle of attack range exceeds the limit. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the present invention provides an aircraft adaptive control method considering the delay constraint of the angle of attack. First, the longitudinal dynamics model of the aircraft is established; then, the model is transformed by introducing a new state conversion function to transform the angle of attack delay constraint control problem into a stable control problem of the new state; finally, an adaptive backstepping control strategy based on a neural network is designed for the transformed model to ensure the stability of the system, thereby realizing the delay constraint of the angle of attack. The method has strong versatility and high calculation accuracy, and has important reference significance for engineering practice.

[0004] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0005] Step 1: Establish the longitudinal dynamics model of the aircraft;

[0006] Step 1-1: Establish the longitudinal dynamics model with the flight path angle as output as follows:

[0007]

[0008] Where V, γ, α and q represent the flight speed, track angle, angle of attack and pitch angular velocity respectively; m, I yy ,g and δ e They represent the mass of the aircraft, the moment of inertia of the y-axis, the gravitational acceleration and the elevator deflection angle respectively; L(α), T and M yy (δ e ,α) represent lift, engine thrust and pitch moment respectively, d γ ,d α and d q is a bounded unknown interference term;

[0009] Step 1-2: Define the system state as: x1 = γ, x2 = α, x3 = q, define the system control input as: u = δ e , then model (1) can be re-expressed as follows:

[0010]

[0011] in, is the state vector, and is an unknown nonlinear function;

[0012] Step 1-3: Set the purpose of the invention as:

[0013] Design an adaptive controller to ensure the stability of system (2) and to ensure the initial over-limit flight angle of attack At the specified time T α Converges to the specified constraint range Ω α Inside:

[0014]

[0015] Among them, α u and α l are the upper and lower boundaries of the angle of attack constraint, T α is the specified time for the angle of attack to converge within the constraint range;

[0016] Step 2: Model transformation;

[0017] Step 2-1: Introduce the following state transition function:

[0018]

[0019] The state is transformed as follows using the conversion function (4):

[0020] s1=x1, s2=x2+S(x2), s3=x3 (5)

[0021] in,

[0022]

[0023] Step 2-2: Model formula (2) is transformed into:

[0024]

[0025] in, is the system state after model transformation, and is the unknown nonlinear function related to the transformed state, and is the composite disturbance after model transformation;

[0026] Step 2-3: For the aircraft system, the composite disturbance function satisfies:

[0027]

[0028] in, and is an unknown continuous nonlinear function related to the state;

[0029] Step 3: Design an adaptive backstepping control strategy based on a neural network for the system after model transformation;

[0030] Step 3-1: Define the first step tracking error z1 = s1-γ d , where γ d Represents the expected track angle command, and the virtual control command is designed as follows:

[0031]

[0032] in, is the Gaussian basis function of the neural network, and the parameters k1>0 and λ1>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0033]

[0034] Among them, the parameters σ1>0 and μ1>0 are given by the designer;

[0035] Step 3-2: Define the first step tracking error The virtual control command color is designed as follows:

[0036]

[0037] in, is the Gaussian basis function of the neural network, and the parameters k2>0 and λ2>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0038]

[0039] Among them, the parameters σ2>0 and μ2>0 are given by the designer;

[0040] Step 3-3: Define the first step tracking error The elevator deflection angle control command is designed as follows:

[0041]

[0042] in, is the Gaussian basis function of the neural network, and the parameters k3>0 and λ3>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0043]

[0044] Among them, parameters σ3>0 and μ3>0 are given by the designer.

[0045] Step 4: Based on the obtained control input, return to the dynamic model (1) of the aircraft, so that stable tracking control of the track angle can be achieved, and the initial over-limit angle of attack can converge to the constraint range within the specified time.

[0046] Furthermore, the pitch moment M yy (δ e ,α) is calculated as follows:

[0047]

[0048] Among them, S w , Q and ρ h Respectively represent the wing area, dynamic pressure and air density; c A , C L and C m They represent the mean aerodynamic chord length, lift coefficient and pitch moment coefficient respectively; C D0 , C Dα , C L0 , C Lα , C m0 , C mα , and C mq denotes the corresponding aerodynamic derivative.

[0049] The beneficial effects of the present invention are as follows:

[0050] The present invention considers the aircraft control problem with model uncertainty and angle of attack delay constraint requirements, introduces a new state conversion function to perform nonlinear transformation on the angle of attack, transforms the angle of attack delay constraint control problem into a stable control problem of the new state, and designs an adaptive control algorithm based on a neural network for the system after the model transformation, while ensuring that the system stably tracks the desired track angle, the initial over-limit angle of attack converges to the specified constraint range within a specified time. This method can be widely used in the design of flight control systems with constraints on the available angle of attack range, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0052] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0053] The purpose of the present invention is to provide an aircraft adaptive control method considering the delay constraint of the angle of attack, so as to solve the problem of the angle of attack constraint control of the aircraft under the conditions of uncertainty and over-limit of the initial value of the angle of attack. The method introduces a new state conversion function to transform the angle of attack delay constraint control problem into a stable control problem of the new state, and designs an adaptive backstepping control algorithm based on a neural network for the transformed system, so as to realize the delay constraint of the angle of attack by ensuring the stability of the new system.

[0054] The present invention discloses an aircraft adaptive control method considering angle of attack delay constraint. The innovation of this method lies in that a delay constraint control strategy for the angle of attack of an aircraft is designed by introducing a new state transformation function, which relaxes the requirement of the traditional angle of attack constraint control method that the initial moment of the aircraft angle of attack must meet the set constraint, ensuring that the flight angle of attack with an initial value exceeding the limit can converge to the safety constraint range within a specified time.

[0055] The present invention first establishes the longitudinal dynamics model of the aircraft; then transforms the model by introducing a new state transition function, transforming the angle of attack delay constraint control problem into a new state stability control problem; finally, designs an adaptive backstepping control strategy based on a neural network for the transformed model to ensure the stability of the system, thereby realizing the delay constraint of the angle of attack. The algorithm logic block diagram is shown in the figure. Figure 1 This method has a simple structure, is easy to implement, has strong engineering practice ability, and can be widely used in the design of flight control systems for various aircraft.

[0056] The design steps are described as follows:

[0057] Step 1: Establish the longitudinal dynamics model of the aircraft;

[0058] The longitudinal dynamics model with the flight path angle as output is established as follows

[0059]

[0060] Where V, γ, α and q represent the flight speed, track angle, angle of attack and pitch velocity respectively; m, I yy ,g and δ e They represent the mass of the aircraft, the moment of inertia of the y-axis, the gravitational acceleration and the elevator deflection angle respectively; L(α), T and M yy (δ e,α) represent lift (function of angle of attack α), engine thrust and pitch moment (function of angle of attack α and elevator deflection angle δ) respectively. e function), d γ ,d α and d q is a bounded unknown interference term.

[0061] Define the system state as: x1 = γ, x2 = α, x3 = q, define the system control input as: u = δ e , then the model formula (1) can be re-expressed as follows:

[0062]

[0063] in, is the state vector, and is an unknown nonlinear function.

[0064] The purpose of the invention can be further described as: designing an adaptive controller to ensure the stability of system (2) and to ensure that the initial over-limit flight angle of attack At the specified time T α Converges to the specified constraint range Ω α Inside:

[0065]

[0066] Among them, α u and α l are the upper and lower boundaries of the angle of attack constraint, T α It is the specified time for the angle of attack to converge to the constraint range.

[0067] Step 2: Model transformation;

[0068] Introduce the following state transition function:

[0069]

[0070] The state is transformed as follows using the conversion function (4):

[0071] s1=x1, s2=x2+S(x2), s3=x3 (5)

[0072] in,

[0073]

[0074] Furthermore, the model formula (2) can be expressed as follows after transformation:

[0075]

[0076] in, The system state after the model transformation, and is the unknown nonlinear function related to the transformed state, and is the composite disturbance after model transformation.

[0077] For the aircraft system, these composite disturbance functions satisfy:

[0078]

[0079] in, and is an unknown continuous nonlinear function related to the state.

[0080] Step 3, designing an adaptive backstepping control strategy based on a neural network for the system after model transformation;

[0081] (a) Define the first step tracking error z1 = s1 - γ d , where γ d Represents the expected track angle command, and the virtual control command is designed as follows

[0082]

[0083] in, is the Gaussian basis function of the neural network, and the parameters k1>0 and λ1>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0084]

[0085] Among them, parameters σ1>0 and μ1>0 are given by the designer.

[0086] (b) Define the first step tracking error The virtual control command color design is as follows;

[0087]

[0088] in, is the Gaussian basis function of the neural network, and the parameters k2>0 and λ2>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0089]

[0090] Among them, parameters σ2>0 and μ2>0 are given by the designer.

[0091] (c) Define the first step tracking error The elevator deflection angle control command is designed as follows:

[0092]

[0093] in, is the Gaussian basis function of the neural network, and the parameters k3>0 and λ3>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0094]

[0095] Among them, parameters σ3>0 and μ3>0 are given by the designer.

[0096] Step 4: Based on the obtained control input, the dynamic model (1) of the aircraft is returned to achieve stable tracking control of the track angle, and the initial over-limit angle of attack can converge to the constraint range within the specified time.

[0097] Example:

[0098] The present invention is further described in conjunction with a calculation example of an aircraft system.

[0099] Step 1: Model building;

[0100] The longitudinal dynamics model of the aircraft considering unknown time-varying disturbances is:

[0101]

[0102] Where V, γ, α and q represent the speed, track angle, angle of attack and pitch rate respectively; T represents the engine thrust, L represents the lift, M represents the yy represents the pitch moment; m represents the mass of the aircraft, I yy represents the moment of inertia of the aircraft about the y-axis, g represents the acceleration due to gravity; d γ ,d α ,d q Indicates that the system is subject to unknown external disturbances;

[0103] The calculation method of the relevant aerodynamic force and moment in formula (15) is as follows:

[0104]

[0105] Among them, S w , Q and ρ h Respectively represent the wing area, dynamic pressure and air density; c A , C L and C m They represent the mean aerodynamic chord length, lift coefficient and pitch moment coefficient respectively; CD0 , C Dα , C L0 , C Lα , C m0 , C mα , and C mq denotes the corresponding aerodynamic derivative.

[0106] Define the system state as: x1 = γ, x2 = α, x3 = q, define the system control input as: u = δ e , then model (15) can be re-expressed as follows:

[0107]

[0108] in, is the state vector, and is an unknown nonlinear function.

[0109] The initial value of the angle of attack is x2(0) = 6°, and the aircraft is required to stably track the desired trajectory γ = 5cos(0.1t) (in degrees) and the flight angle of attack is T α = Converge within 5s to the following constraints:

[0110]

[0111] Among them, α u =5° and α l =-5° are the upper and lower boundaries of the attack angle constraint, T α =5s is the specified time for the angle of attack to converge within the constraint range.

[0112] Step 2: Model transformation;

[0113] Introduce the following state transition function:

[0114]

[0115] The state is transformed as follows using the conversion function (19):

[0116] s1=x1, s2=x2+S(x2), s3=x3 (20)

[0117] in,

[0118]

[0119] Furthermore, model (2) can be expressed as follows after transformation:

[0120]

[0121] in, The system state after the model transformation, and is the unknown nonlinear function related to the transformed state, and is the composite disturbance after model transformation. For the aircraft system, these composite disturbance functions satisfy:

[0122]

[0123] in, and is an unknown continuous nonlinear function related to the state.

[0124] Step 3, designing an adaptive backstepping control strategy based on a neural network for the system after model transformation;

[0125] (a) Define the first step tracking error z1 = s1 - γ d , where γ d Represents the expected track angle command, and the virtual control command is designed as follows:

[0126]

[0127] in, is the Gaussian basis function of the neural network, and the parameters k1=5 and λ1=20 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0128]

[0129] Among them, parameters σ1=0.01 and μ1=1 are given by the designer.

[0130] (b) Define the first step tracking error The virtual control command color is designed as follows:

[0131]

[0132] in, is the Gaussian basis function of the neural network, and the parameters k2=10 and λ2=20 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0133]

[0134] Among them, parameters σ2=0.01 and μ2=1 are given by the designer.

[0135] (c) Define the first step tracking error The elevator angle control command is designed as follows

[0136]

[0137] in, is the Gaussian basis function of the neural network, and the parameters k3=10 and λ3=20 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows:

[0138]

[0139] Among them, parameters σ3=0.01 and μ3=1 are given by the designer.

[0140] Step 4: Based on the obtained control input, the dynamic model (17) of the aircraft is returned to achieve stable tracking control of the track angle, and the initial over-limit angle of attack can converge to the constraint range within the specified time.

Claims

1. An adaptive control method for an aircraft considering angle of attack delay constraints, characterized in that: The steps include: Step 1: Establish the longitudinal dynamics model of the aircraft; Step 1-1: Establish the longitudinal dynamics model with the flight path angle as output as follows: Where V, γ, α and q represent the flight speed, track angle, angle of attack and pitch angular velocity respectively; m, I yy ,g and δ e They represent the mass of the aircraft, the moment of inertia of the y-axis, the gravitational acceleration and the elevator deflection angle respectively; L(α), T and M yy (δ e ,α) represent lift, engine thrust and pitch moment respectively, d γ ,d α and d q is a bounded unknown interference term; Step 1-2: Define the system state as: x1 = γ, x2 = α, x3 = q, define the system control input as: u = δ e , then model (1) can be re-expressed as follows: in, is the state vector, and is an unknown nonlinear function; Step 1-3: Set the purpose of the invention as: Design an adaptive controller to ensure the stability of system (2) and to ensure the initial over-limit flight angle of attack At the specified time T α Converges to the specified constraint range Ω α Inside: Among them, α u and α l are the upper and lower boundaries of the angle of attack constraint, T α is the specified time for the angle of attack to converge within the constraint range; Step 2: Model transformation; Step 2-1: Introduce the following state transition function: The state is transformed as follows using the conversion function (4): s1=x1, s2=x2+S(x2), s3=x3 (5) in, Step 2-2: Model formula (2) is transformed into: in, is the system state after model transformation, and is the unknown nonlinear function related to the transformed state, and is the composite disturbance after model transformation; Step 2-3: For the aircraft system, the composite disturbance function satisfies: in, and is an unknown continuous nonlinear function related to the state; Step 3: Design an adaptive backstepping control strategy based on a neural network for the system after model transformation; Step 3-1: Define the first step tracking error z1 = s1-γ d , where γ d Represents the expected track angle command, and the virtual control command is designed as follows: in, is the Gaussian basis function of the neural network, and the parameters k1>0 and λ1>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows: Among them, the parameters σ1>0 and μ1>0 are given by the designer; Step 3-2: Define the first step tracking error The virtual control command color is designed as follows: in, is the Gaussian basis function of the neural network, and the parameters k2>0 and λ2>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows: Among them, the parameters σ2>0 and μ2>0 are given by the designer; Step 3-3: Define the first step tracking error The elevator deflection angle control command is designed as follows: in, is the Gaussian basis function of the neural network, and the parameters k3>0 and λ3>0 are given by the designer. is the adaptive parameter estimate related to the optimal weight vector of the neural network, and its adaptive law is designed as follows: Among them, the parameters σ3>0 and μ3>0 are given by the designer; Step 4: Based on the obtained control input, return to the dynamic model (1) of the aircraft, so that stable tracking control of the track angle can be achieved, and the initial over-limit angle of attack can converge to the constraint range within the specified time.

2. The method for adaptive control of an aircraft considering angle of attack delay constraints according to claim 1, characterized in that: The pitching moment M yy (δ e ,α) is calculated as follows: Among them, S w , Q and ρ h Respectively represent the wing area, dynamic pressure and air density; c A , C L and C m They represent the mean aerodynamic chord length, lift coefficient and pitch moment coefficient respectively; C D0 , C Dα , C Dα2 , C L0 , C Lα , C m0 , C mα , and C mq represents the corresponding aerodynamic derivative.

Images