Constraint-considered hypersonic aircraft composite model predictive control method

Abstract

The invention discloses a hypersonic flight vehicle composite model predictive control method considering constraints, which comprises the following steps of: firstly, establishing a dynamic model of a longitudinal motion system of a hypersonic flight vehicle, and fully considering the constraints of control quantity and increment thereof and the influence of multi-source interference in a modeling process; secondly, based on an extended state observer technology, observers are designed for the pitch angle rate channel and the accelerator opening degree dynamic state respectively, so that lumped disturbance estimation is achieved; designing a model prediction controller under the condition of considering the control quantity amplitude and change rate constraint; and finally, performing feed-forward compensation in combination with interference estimation information to construct a composite model prediction controller. According to the method, the anti-interference capability of the system is remarkably improved through interference estimation and feed-forward compensation, and the amplitude and the change rate of the control quantity are ensured to meet constraint requirements. The method has better anti-interference capability and constraint processing capability, and effectively inhibits the influence of interference on the control performance of the hypersonic aircraft.

Description

Technical Field

[0001] The invention belongs to the technical field of flight control, and in particular relates to a composite model predictive control method for a hypersonic aircraft considering constraints. Background Art

[0002] Hypersonic vehicles have attracted widespread attention due to their advantages such as high flight speed and strong strike capability. However, due to their wide flight envelope and large speed and airspace coverage, control system design not only faces problems such as strong nonlinearity and strong coupling, but is also affected by multiple sources of interference such as external interference, internal parameter perturbations, and unmodeled dynamics. In addition, due to the bandwidth and energy limitations of the actuators, the control aspects of hypersonic vehicles also need to meet the restrictions on rate of change and amplitude.

[0003] In response to the influence of interference and system nonlinearity, domestic and foreign scholars have proposed a variety of advanced control schemes such as adaptive control, robust control, and sliding mode control. Although these schemes have improved the system nonlinearity and interference processing capabilities to a certain extent, the interference suppression strategy of these methods is to passively eliminate the influence of interference through error signals. When the interference is large, it is impossible to achieve rapid suppression of interference; in addition, the existing methods are limited by the limitations of the algorithm itself and rarely consider constraints such as the rate of change and amplitude of the control quantity. Therefore, it is urgent to propose a hypersonic aircraft control method that can quickly suppress interference and handle constraints. Summary of the invention

[0004] Purpose of the invention: The present invention provides a composite model predictive control method for a hypersonic aircraft taking constraints into consideration, which realizes effective estimation of multi-source interference and ensures that the hypersonic aircraft has stronger anti-interference performance and constraint processing capability while satisfying control quantity constraints.

[0005] Technical solution: A hypersonic vehicle composite model predictive control method considering constraints described in the present invention comprises the following steps:

[0006] (1) Establish a disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle taking into account the second-order dynamics of the engine;

[0007] (2) Based on the disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle, extended state observers are designed for the pitch rate channel and throttle opening dynamics to obtain disturbance estimation information;

[0008] (3) Design a model predictive controller considering the control variable amplitude and change rate constraints to obtain the optimal control variable considering the constraints;

[0009] (4) Construct a composite model predictive controller based on the disturbance estimation information obtained in step (2) and the optimal control quantity considering constraints obtained in step (3).

[0010] Furthermore, the disturbed dynamics model of the longitudinal motion system of the hypersonic vehicle in step (1) is:

[0011]

[0012] Where h is the flight altitude; V is the speed; α and γ are the angle of attack and track angle respectively, q is the pitch rate; r is the distance from the center of the earth, r = h + R E , R E is the radius of the Earth; and Represent the first-order derivatives of h, V, γ, α and q respectively; I yy represents the moment of inertia of the hypersonic vehicle around the y-axis; m is the mass of the vehicle itself; μ is the gravitational constant; D q is the disturbance of the pitch rate channel; L, D, T, and M are lift, drag, engine thrust, and pitch moment, respectively;

[0013] The relationship between the engine throttle opening dynamics and the throttle opening command is:

[0014]

[0015] Among them, δ e represents the elevator deflection angle, c e represents the proportional coefficient; β is the engine throttle opening, β c is the throttle opening command; ω n is the angular frequency, represents damping, which are all known constants, D β It is the disturbance in the throttle opening dynamics;

[0016] The actual control variable in the longitudinal motion model of the hypersonic vehicle is the elevator deflection angle δ e and the engine throttle opening command β c , the control target is to fly the aircraft at the desired flight altitude h r and speed V r Flight; elevator rudder deflection angle δ e and engine throttle opening command β c The constraints on amplitude and rate of change need to be met:

[0017]

[0018] The state space model is used to represent the perturbed dynamics model (1)-(2) of the longitudinal motion system of a hypersonic vehicle, and the system state x and control variable u are defined as:

[0019] u=[β c δ e ]

[0020] Ignoring the effect of disturbance in the system (1)-(2), at the equilibrium point u=u0=[β c 0 δ e 0 ] T The linearization nearby gives:

[0021]

[0022] Among them: A m is the state transfer matrix, B m is the input matrix, C m is the output matrix, D m is the direct transfer matrix; considering the influence of the pitch rate channel and the throttle opening dynamic disturbance, the disturbed dynamic in equation (3) is abbreviated as:

[0023]

[0024] Among them, a 52 ,a 54 ,a 55 ,b 52 ,a 76 ,a 77 ,b 71 A m and B m The corresponding coefficient term in D q =b 52 ·d q , D β =b 71 ·d β They are the interferences of the pitch angle rate and throttle opening channels respectively.

[0025] Furthermore, the lift, drag, engine thrust, and pitch moment are respectively:

[0026] L=QSC L ,D=QSC D ,T=QSC T ,

[0027]

[0028] Where Q is the dynamic pressure, ρ is the atmospheric density, S is the reference area, is the average aerodynamic chord length of the wing; C L , C D , C M , C T They are the aerodynamic coefficients of lift, drag, pitch moment and engine thrust coefficient:

[0029] C L =0.6203α,C D =0.6450α 2 +0.0043378α+0.003772,

[0030] C M (α)=-0.035α 2 +0.036617α+5.3261×10 -6 ,C M (δ e )=c e (δ e -α)

[0031]

[0032] Among them, δ e represents the elevator deflection angle, c e represents the proportional coefficient; β is the engine throttle opening, β c is the relationship between the throttle opening command.

[0033] Furthermore, the implementation process of step (2) is as follows:

[0034] The extended state observer is designed for the disturbed system dynamics and D is estimated. q , D β :

[0035]

[0036] in, and is the extended state observer dynamics; and They are and The derivative of and is the observer gain and is a positive constant; To interfere with D q The estimated value of To interfere with D β The estimated value of .

[0037] Furthermore, the implementation process of step (3) is as follows:

[0038] Define a new augmented state x for the system ignoring the influence of interference A :

[0039] e=[hh r VV r ]T

[0040] Considering the reference instruction h r and V r Usually a constant value, the augmented state x A News:

[0041]

[0042] in, is the derivative of the control quantity u, A, B, C are constant matrices; combined with the longitudinal motion control target of the hypersonic aircraft system, the cost function is introduced:

[0043]

[0044] Where 0≤τ≤T p , R = diag{r1 r2}, W = diag{w1 w2}; in the prediction interval [0 T P ] introduces the Laguerre function and uses two sets of standard orthogonal basis functions to approximate Then we get:

[0045]

[0046] In the formula, N1 and N2 represent the number of terms in the function. and are a set of orthogonal basis functions; and is the coefficient vector to be determined,

[0047] According to the properties of Laguerre function:

[0048]

[0049] Where p = [p1 p2] is the time scale factor of the Laguerre function; considering the orthogonality of the Laguerre function, that is, is the identity matrix, and choose a sufficiently large prediction step size T P , so that τ ≥ T P hour, That is, the selected prediction range is greater than the effective time of the control signal, then the second term of equation (8) can be calculated as:

[0050]

[0051] Based on the current state, predict the tracking error of the system at future times:

[0052]

[0053] Formula (11) is rewritten as:

[0054]

[0055] Combining equations (8)-(12) gives the cost function expression:

[0056]

[0057] Where η=[η1 η2] T , η is the variable to be optimized. Without considering the constraints, the η that minimizes the cost function * for:

[0058] η * =-Ω -1 Ψx A (t) (14)

[0059] At any future time t+τ, the control variable derivative can be obtained by solving equation (9): The optimal value of is:

[0060]

[0061] By applying the rolling optimization principle, that is, the control action only uses the derivative of the future signal at τ = 0; without considering the constraints, for any time t, the optimal control amount change rate is:

[0062]

[0063] The feedback gain matrix is:

[0064]

[0065] Without considering the constraints, the optimal control quantity is obtained:

[0066]

[0067] Considering the constraints of the control amplitude and change rate, the model predictive controller is designed: For the constraint of the control amplitude, if the sampling interval is Δt, the control signal at time t is:

[0068]

[0069] Where L(0) = [L 1 (0) L 2 (0)], η=[η1 η2] T ,L(0) T η is the control derivative at the beginning of the optimization window. At any future time τ, the control signal is obtained by integrating the control amount at the previous time plus the control signal derivative:

[0070]

[0071] If the sampling interval is Δt, the future control signal at time t+τ is:

[0072]

[0073] The inequality constraint is of the form:

[0074] u min -u(t-Δt)≤C u η≤u max -u(t-Δt) (21)

[0075] in, and are the upper and lower bounds of the control amplitude constraint, respectively, τ=0, C u =[L 1 (0) T Δt L 2 (0) T Δt] T ; For the constraint of the control variable change rate, a constraint is imposed on the derivative of the control signal at time t. The constraint is expressed as:

[0076]

[0077] The control derivative constraint for the start of the optimization window is implemented as follows:

[0078]

[0079] in, and are the upper and lower bounds of the control variable change rate constraint respectively;

[0080] It is known that the cost function (13) is a quadratic objective function. Combining the control amplitude and change rate constraints of (13), (21) and (23) is a typical quadratic programming problem:

[0081] min J=η T Ωη+2η T ψx A (t) (24)

[0082] stMη≤γ (25)where:

[0083]

[0084] Solve the optimal coefficient solution of the control quantity under the constraints Then we can get the control quantity change rate For the optimal control quantity under the constraints, the integral of the control quantity change rate is obtained:

[0085]

[0086] Furthermore, the composite model predictive controller in step (4) is:

[0087]

[0088] in,

[0089] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention combines the constraint processing advantages of model predictive control and the interference fast processing advantages of active anti-interference, realizes the effective estimation of multi-source interference in the longitudinal motion system of hypersonic aircraft, and compensates for the adverse effects of multi-source interference in the form of feedforward interference estimation information, ensuring that the hypersonic aircraft has stronger anti-interference performance and constraint processing capabilities while meeting the control quantity constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 is a flow chart of the present invention;

[0091] Figure 2 The composite model predictive controller CCMPC considering constraints of the method proposed in the present invention, the model predictive controller MPC used for comparison, and the model predictive controller CMPC considering constraints and their tracking error response curves;

[0092] Figure 3 It is the speed and its tracking error response curve under three methods;

[0093] Figure 4 The following are the response curves of the elevator and its rate of change under the three methods;

[0094] Figure 5 These are the response curves of the throttle opening command and its rate of change under three methods. DETAILED DESCRIPTION

[0095] The present invention is further described in detail below with reference to the accompanying drawings.

[0096] like Figure 1 As shown, the present invention provides a hypersonic aircraft composite model predictive control method considering constraints, and the specific steps are as follows:

[0097] S1. Establish a disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle considering the second-order dynamics of the engine.

[0098] The disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle considering the second-order dynamics of the engine is established:

[0099]

[0100] Where h is the flight altitude; V is the speed; α and γ are the angle of attack and track angle respectively, q is the pitch rate; r is the distance from the center of the earth, r = h + R E , R E is the radius of the Earth; and Represent the first-order derivatives of h, V, γ, α and q respectively; I yy represents the moment of inertia of the hypersonic vehicle around the y-axis; m is the mass of the vehicle itself; μ is the gravitational constant; D q is the disturbance of the pitch rate channel; L, D, T, M are lift, drag, engine thrust, and pitch moment, respectively, and their expressions are as follows:

[0101] L=QSC L , D = QSC D , T=QSC T ,

[0102] M=QSc[C M (α)+C M (δ e )+C M (q)]

[0103] Where ρ is the atmospheric density, S is the reference area, Q is the dynamic pressure, is the average aerodynamic chord length of the wing; C L , C D , C M , C T They are the aerodynamic coefficients of lift, drag, pitch moment and engine thrust coefficient:

[0104] C L =0.6203α,C D =0.6450α 2 +0.0043378α+0.003772,

[0105] C M (α)=-0.035α 2 +0.036617α+5.3261×10 -6 ,C M (δ e )=c e (δ e -α),

[0106]

[0107] Among them, δ e represents the elevator deflection angle, ce represents the proportional coefficient; β is the engine throttle opening, and its dynamics is related to the throttle opening command β c Relationship:

[0108]

[0109] Among them, ω n is the angular frequency, represents damping, which are all known constants, D β It is the disturbance in the throttle opening dynamics.

[0110] The hypersonic vehicle parameters and atmospheric parameters under certain cruise conditions are shown in Table 1.

[0111] Table 1 Parameter list

[0112]

[0113] The model of the hypersonic vehicle is linearized under cruise conditions (considering engine dynamics). Here, the equilibrium cruise conditions are considered to be V0 = 15060ft, γ0 = 0, h0 = 110000ft, q0 = 0, α0 = 0.0312rad, β0 = 0.1762,

[0114] From the above analysis, it can be seen that the actual control amount in the longitudinal motion model of the hypersonic aircraft is the rudder deflection angle δ of the elevator e and engine throttle opening command β c , the control goal is "the aircraft flies at the desired altitude h r = 609.6m and speed V r =30.48m / s flight". Considering the actual situation in the project, the elevator rudder deflection angle δ e and engine throttle opening command β c The constraints on amplitude and rate of change need to be met:

[0115]

[0116] Using the state space model to represent the hypersonic vehicle dynamics model (1)-(2), the system state x and control variable u can be defined as:

[0117] u=[β c δ e ]

[0118] Ignoring the effect of disturbance in the system (1)-(2), at the equilibrium point u=u0=[β c 0 δ e 0] T The linearization nearby gives:

[0119]

[0120] in:

[0121]

[0122] Considering the influence of the pitch rate channel and the throttle opening dynamic disturbance, the disturbed dynamics in equation (3) can be simplified as:

[0123]

[0124] in, a 77 =-2ζω n , D q =b 52 ·d q , D β =b 71 ·d β They are the interferences of the pitch angle rate and throttle opening channels respectively.

[0125] S2. Design extended state observers for the pitch rate channel and throttle opening dynamics respectively.

[0126] The extended state observers are designed for the pitch rate and throttle opening channels respectively. The specific designs are as follows: the extended state observer is designed for the disturbed system dynamics (4) to estimate D q , D β

[0127]

[0128] in, and is the extended state observer dynamics; and They are and The derivative of and is the observer gain and is a positive constant; To interfere with D q The estimated value of To interfere with D β The estimated value of .

[0129] S3. Design a model predictive controller considering the control variable amplitude and change rate constraints.

[0130] Design a model predictive controller without considering constraints. The specific design is as follows:

[0131] Define a new augmented state x for system (3) A :

[0132] e=[h-609.6V-30.48] T

[0133] Considering the reference instruction h r and V r Usually a constant value, then according to system (3) we can get the augmented state x A News:

[0134]

[0135] in, is the derivative of the control quantity u, A, B, C are constant matrices, and their values ​​are:

[0136] C=[o 2×7 I 2×2 ].

[0137] Combined with the longitudinal motion control target of the hypersonic vehicle system, the cost function is introduced:

[0138]

[0139] Where 0≤τ≤T p 、T p =18s, In the prediction interval [0T P ] introduces the Laguerre function and uses two sets of standard orthogonal basis functions to approximate Then we can get:

[0140]

[0141] In the formula, N1=9, N2=15 represents the number of terms in the function. and are a set of orthogonal basis functions; and is the coefficient vector to be determined,

[0142] According to the properties of Laguerre function, we can get:

[0143]

[0144] Where p = [p1 p2] is the time scale factor of the Laguerre function, and p1 = 18, p2 = 7. Considering the orthogonality of the Laguerre function, that is is the identity matrix, and choose a sufficiently large prediction step size T P , so that τ ≥ T P hour, That is, the selected prediction range is greater than the effective time of the control signal, then the second term of equation (8) can be calculated as:

[0145]

[0146] Based on the current state, predict the tracking error of the system at future times:

[0147]

[0148] Formula (11) can be rewritten as:

[0149]

[0150] Combining equations (8)-(12) gives the cost function expression:

[0151]

[0152] Where η=[η1 η2] T , Obviously, this is a typical quadratic objective function, and η is the variable to be optimized. Without considering the constraints, we can get η that minimizes the cost function. * for:

[0153] η * =-Ω -1 Ψx A (t) (14)

[0154] At any future time t+τ, the control variable derivative can be obtained by solving equation (9): The optimal value of is:

[0155]

[0156] By applying the rolling optimization principle, that is, the control action only uses the derivative of the future signal at τ = 0, the optimal control change rate for any time t can be obtained without considering the constraints:

[0157]

[0158] The feedback gain matrix is:

[0159]

[0160] The feedback gain matrix is ​​obtained as K mpc :

[0161]

[0162] Without considering the constraints, the optimal control quantity can be solved:

[0163]

[0164] If the model predictive controller is designed considering the control variable amplitude and change rate constraints, the design process is:

[0165] For the constraint of the control amplitude, if the sampling interval is Δt = 0.01s, the control signal at time t can be calculated as:

[0166]

[0167] Where L(0) = [L 1 (0) L 2 (0)], η=[η1 η2] T ,L(0) T η is the control derivative at the beginning of the optimization window. At any future time τ, the control signal can be obtained by adding the control amount at the previous time to the integral of the control signal derivative.

[0168]

[0169] If the sampling interval is Δt, the future control signal at time t+τ can be expressed as:

[0170]

[0171] Therefore, the inequality constraint is of the form:

[0172] u min -u(t-Δt)≤C u η≤u max -u(t-Δt) (21)

[0173] in, and are the upper and lower bounds of the control amplitude constraint, respectively, τ=0, C u =[L 1 (0) T Δt L 2 (0) T Δt] T In order to constrain the rate of change of the control variable, a constraint is imposed on the derivative of the control signal at time t. The constraint is expressed as:

[0174]

[0175] The control derivative constraint for the start of the optimization window is implemented as follows:

[0176]

[0177] in, and are the upper and lower bounds of the control variable change rate constraint respectively.

[0178] It is known that the cost function (13) is a quadratic objective function. Combining the control amplitude and change rate constraints of (13), (21) and (23) is a typical quadratic programming problem:

[0179] min J=η T Ωη+2η T ψx A (t) (24)

[0180] stMη≤γ (25)where:

[0181]

[0182] The optimal coefficient solution of the control quantity under the constraints can be solved Then we can get the control quantity change rate

[0183] The optimal control quantity under the constraints can be obtained by integrating the rate of change of the control quantity:

[0184]

[0185] S4. Construct a composite model predictive controller based on the disturbance estimation information.

[0186] Based on the disturbance estimation information obtained by observers (5)-(6) in S2 and the optimal control quantity considering constraints obtained in S3, a composite model predictive controller is constructed:

[0187]

[0188] in,

[0189] In this embodiment, the interference is d β =0.1sin(0.3t);d q =0.06sin(0.3t). Using the Simulink toolbox in the simulation software MATLABR2020b, a disturbed longitudinal motion model of a hypersonic vehicle considering constraints was built for simulation and experiments. Figures 2 to 5The height and speed tracking effects and control quantity response curves under three different methods are shown in FIG. 1 , wherein the method of the present invention is a composite constrained model predictive controller (CCMPC); a model predictive controller (MPC) used for comparison; and a constrained model predictive controller (CMPC) used for comparison. Figure 2 The height and tracking error response curves of the three methods are given. It can be seen from the figure that the height tracking error accuracy of the CCMPC method proposed in the present invention is significantly higher than that of the comparison method, and the two comparison methods MPC and CMPC obtain the same tracking accuracy. Figure 3 The speed and tracking error response curves of the three methods are given. It can be seen from the figure that the method of the present invention obtains the best speed tracking accuracy; the two comparison methods obtain similar speed tracking accuracy. Figure 4 The response curve of throttle opening and its change rate is given. It can be seen from the figure that the proposed method CCMPC and the comparative method CMPC ensure that the throttle opening and its change rate are within the constraint range, while the throttle opening and its change rate of the system under the action of the traditional MPC method exceed the constraint range. Figure 5 The response curve of elevator deflection and its rate of change is given. It can be seen from the figure that the proposed method CCMPC and the comparative method CMPC ensure that the elevator deflection and its rate of change are within the constraint range, while the elevator deflection and its rate of change of the system under the action of the traditional MPC method are beyond the constraint range. Compared with the traditional predictive control method, the present invention has better anti-interference ability and constraint processing ability, and ensures high-precision tracking of the height and speed instructions of the hypersonic aircraft under the premise of meeting the constraints of the elevator and throttle opening amplitude and rate of change.

[0190] The above-mentioned embodiments only express the implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A hypersonic vehicle composite model predictive control method considering constraints, characterized in that: The following steps are involved: (1) Establish a disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle taking into account the second-order dynamics of the engine; (2) Based on the disturbed dynamics model of the longitudinal motion system of a hypersonic vehicle, extended state observers are designed for the pitch rate channel and throttle opening dynamics to obtain disturbance estimation information; (3) Design a model predictive controller considering the control variable amplitude and change rate constraints to obtain the optimal control variable considering the constraints; (4) Construct a composite model predictive controller based on the disturbance estimation information obtained in step (2) and the optimal control quantity considering constraints obtained in step (3).

2. The method for complex model predictive control of a hypersonic vehicle considering constraints according to claim 1, characterized in that: The disturbed dynamics model of the longitudinal motion system of the hypersonic vehicle in step (1) is: Where h is the flight altitude; V is the speed; α and γ are the angle of attack and track angle respectively, q is the pitch rate; r is the distance from the center of the earth, r = h + R E , R E is the radius of the Earth; and Represent the first-order derivatives of h, V, γ, α and q respectively; I yy represents the moment of inertia of the hypersonic vehicle around the y-axis; m is the mass of the vehicle itself; μ is the gravitational constant; D q is the disturbance of the pitch rate channel; L, D, T, and M are lift, drag, engine thrust, and pitch moment, respectively; The relationship between the engine throttle opening dynamics and the throttle opening command is: Among them, δ e represents the elevator deflection angle, c e represents the proportional coefficient; β is the engine throttle opening, β c is the throttle opening command; ω n is the angular frequency, represents damping, which are all known constants, D β It is the disturbance in the throttle opening dynamics; The actual control variable in the longitudinal motion model of the hypersonic vehicle is the elevator deflection angle δ e and the engine throttle opening command β c , the control target is to fly the aircraft at the desired flight altitude h r and speed V r Flight; elevator rudder deflection angle δ e and the engine throttle opening command β c The constraints on amplitude and rate of change need to be met: The state space model is used to represent the perturbed dynamics model (1)-(2) of the longitudinal motion system of a hypersonic vehicle, and the system state x and control variable u are defined as: Ignoring the effect of disturbance in the system (1)-(2), at the equilibrium point The linearization nearby gives: Among them: A m is the state transfer matrix, B m is the input matrix, C m is the output matrix, D m is the direct transfer matrix; considering the influence of the pitch rate channel and the throttle opening dynamic disturbance, the disturbed dynamic in equation (3) is abbreviated as: Among them, a 52 ,a 54 ,a 55 ,b 52 ,a 76 ,a 77 ,b 71 A m and B m The corresponding coefficient term in D q =b 52 ·d q , D β =b 71 ·d β They are the interferences of the pitch angle rate and throttle opening channels respectively.

3. The method for predictive control of a hypersonic vehicle with complex model considering constraints according to claim 2, characterized in that: The lift, drag, engine thrust, and pitch moment are respectively: Where Q is the dynamic pressure, ρ is the atmospheric density, S is the reference area, is the average aerodynamic chord length of the wing; C L , C D , C M , C T They are the aerodynamic coefficients of lift, drag, pitch moment and engine thrust coefficient: C L =0.6203α,C D =0.6450a 2 +0.0043378a+0.003772, C M (a)=-0.035a 2 +0.036617α+5.3261×10 -6 ,C M (d e )=c e (d e -a) Among them, δ e represents the elevator deflection angle, c e represents the proportional coefficient; β is the engine throttle opening, β c is the relationship between the throttle opening command.

4. The method for predictive control of a hypersonic vehicle with complex model considering constraints according to claim 1, characterized in that: The implementation process of step (2) is as follows: The extended state observer is designed for the disturbed system dynamics and D is estimated. q , D β : in, and is the extended state observer dynamics; and They are and The derivative of and is the observer gain and is a positive constant; To interfere with D q The estimated value of To interfere with D β The estimated value of .

5. The method for complex model predictive control of a hypersonic vehicle considering constraints according to claim 1, characterized in that: The implementation process of step (3) is as follows: Define a new augmented state x for the system ignoring the influence of interference A : Considering the reference instruction h r and V r Usually a constant value, the augmented state x A News: in, is the derivative of the control quantity u, A, B, C are constant matrices; combined with the longitudinal motion control target of the hypersonic aircraft system, the cost function is introduced: Where 0≤τ≤T p , R = diag{r1 r2}, W = diag{w1 w2}; in the prediction interval [0 T P ] introduces the Laguerre function and uses two sets of standard orthogonal basis functions to approximate Then we get: In the formula, N1 and N2 represent the number of terms in the function. and are a set of orthogonal basis functions; and is the coefficient vector to be determined, According to the properties of Laguerre function: Where p = [p1p2] is the time scale factor of the Laguerre function; considering the orthogonality of the Laguerre function, that is is the identity matrix, and choose a sufficiently large prediction step size T P , so that τ ≥ T P hour, That is, the selected prediction range is greater than the effective time of the control signal, then the second term of equation (8) can be calculated as: Based on the current state, predict the tracking error of the system at future times: Formula (11) is rewritten as: Combining equations (8)-(12) gives the cost function expression: Where η=[η1 η2] T , η is the variable to be optimized. Without considering the constraints, the η that minimizes the cost function * for: or * =-Ω -1 Ψx A (t) (14) At any future time t+τ, the control variable derivative can be obtained by solving equation (9): The optimal value of is: By applying the rolling optimization principle, that is, the control action only uses the derivative of the future signal at τ = 0; without considering the constraints, for any time t, the optimal control amount change rate is: The feedback gain matrix is: Without considering the constraints, the optimal control quantity is obtained: Considering the constraints of the control amplitude and change rate, the model predictive controller is designed: For the constraint of the control amplitude, if the sampling interval is Δt, the control signal at time t is: Where L(0) = [L 1 (0) L 2 (0)], η=[η1 η2] T ,L(0) T η is the control derivative at the beginning of the optimization window. At any future time τ, the control signal is obtained by integrating the control amount at the previous time plus the control signal derivative: If the sampling interval is Δt, the future control signal at time t+τ is: The inequality constraint is of the form: u min -u(t-Δt)≤C u η≤u max -u(t-Δt) (21) in, and are the upper and lower bounds of the control amplitude constraint, respectively, τ=0, C u =[L 1 (0) T Δt L 2 (0) T Δt] T ; For the constraint of the control variable change rate, a constraint is imposed on the derivative of the control signal at time t. The constraint is expressed as: The control derivative constraint for the start of the optimization window is implemented as follows: in, and are the upper and lower bounds of the control variable change rate constraint respectively; It is known that the cost function (13) is a quadratic objective function. Combining the control amplitude and change rate constraints of (13), (21) and (23) is a typical quadratic programming problem: min J=η T Oh+2h T x A (t) (24) stMη≤γ (25) in: Solve the optimal coefficient solution of the control quantity under the constraints Then we can get the control quantity change rate For the optimal control quantity under the constraints, the integral of the control quantity change rate is obtained:

6. The method for complex model predictive control of a hypersonic vehicle considering constraints according to claim 1, characterized in that: The composite model predictive controller in step (4) is: in,

Images