CN104880946B - A kind of carrier-borne aircraft auto landing on deck control method based on robust preview control - Google Patents

Abstract

A kind of carrier-borne aircraft auto landing on deck control method based on robust preview control disclosed by the invention, so as to carry out prediction control to carrier-borne aircraft using the Future Information and present information of glide path track and longitudinal deck motion, so as to realize the compensation of tracking and longitudinal deck motion to glide path height, can effectively suppress and eliminate laterally offset simultaneously, carrier-borne aircraft is kept lateral stability.Wherein feedforward control is carried out using Future Information, feedback control is carried out using current information, rudder face and throttle that can in advance to carrier-borne aircraft implement average operation to reach tracing compensation purpose, reduce instantaneous energy, and accelerate response speed, it is ensured that carrier-borne function on aircraft carrier safety warship.

Description

A kind of carrier-borne aircraft auto landing on deck control method based on robust preview control

Technical field

The present invention relates to a kind of carrier-borne aircraft auto landing on deck control method, more particularly to a kind of warship based on robust preview control Carrier aircraft auto landing on deck control method.

Background technology

The general principle of carrier-borne aircraft auto landing on deck control system measures the reality of carrier-borne aircraft for the tracking radar on aircraft carrier Border position, while measuring the motion conditions of aircraft-carrier flight deck by deck motion sensor, warship is calculated by predictor method Ideal position residing for carrier aircraft, by the ideal position of carrier-borne aircraft and physical location input instruction computer, is compared and is missed Difference signal, the control instruction of carrier-borne aircraft is calculated according to error signal, then sent out by wireless software download via leading law Carrier-borne aircraft is given, the error signal that the automatic pilot on carrier-borne aircraft is received according to reception device manipulates carrier-borne aircraft and eliminates error, Precalculated position safety warship.Zhou Xin et al. is disclosed《Carrier landing longitudinal direction deck motion prediction and compensation technique》(2013 October in year, Nanjing Aero-Space University's journal, the 5th phase of volume 45) in, describe and deck motion is designed based on particle filter algorithm The method that prediction device carries out deck motion state estimations, obtains deck motion state variable x_kIn the optimal estimation at following τ moment The expression formula of value is

Carrier landing typically warship using glide-slope tracking.So-called glide-slope tracking warship (downslide of carrier-borne aircraft isogonism), It is to enter warship and the final stage of warship, after carrier-borne aircraft intercepts and captures suitable glide path, be always maintained at identical glide paths angle, pitching Angle, speed and deflection ratio, until carrier-borne aircraft is collided with aircraft carrier flight-deck, realize that impacting type warship.Due to the shadow of deck motion Ring, carrier-borne aircraft glide-slope tracking warship overall process can be divided into two stages, and one is the glide-slope tracking stage, and two is deck motion Compensated stage.The general 12.5s before warship of carrier-borne aircraft will during deck motion adds auto landing on deck control system, allow carrier-borne aircraft with Track glides and tracks deck motion during warship simultaneously.Actual downslide tracking phase, traditional PID controller is difficult to make carrier-borne Machine rapidly tracks glide path track；Actual deck compensated stage, traditional PID controller is difficult to make carrier-borne aircraft last Warship stage perfect tracking deck motion so that warship success rate it is low.Therefore, carrier-borne aircraft auto landing on deck design of control method for Carrier-borne aircraft on aircraft carrier safety warship it is particularly important, be directly connected to the success rate and security of carrier-borne aircraft auto landing on deck.

The content of the invention

Goal of the invention：For above-mentioned prior art, there is provided a kind of carrier-borne aircraft auto landing on deck control based on robust preview control Method processed, realizes that carrier-borne aircraft warship after entering the downslide stage along glide path track following, can effectively suppress and eliminate lateral inclined Move, carrier-borne aircraft is kept lateral stability, and realize deck motion prediction with compensation before warship.

Technical scheme：A kind of carrier-borne aircraft auto landing on deck control method based on robust preview control, including deck motion is pre- Estimate module, Longitudinal Control Law module and horizontal lateral control law module；Deck motion feeding deck motion prediction module is obtained first To deck motion prediction information, and deck motion prediction information is introduced into carrier-borne aircraft auto landing on deck control system before warship；Its In, the computational methods of the deck motion prediction module are as follows：

For the deck motion of known transmission function, prediction device is designed using particle filter algorithm, for estimating the τ moment Deck motion information, the expression formula for obtaining deck motion in the optimal estimation value at following τ moment is (k+m, is k) state-transition matrix to wherein Φ, and k is time, m=τ/T_s, T_sIt is systematic sampling time, state transfer matrixA is the Coefficient Space matrix corresponding to deck motion transmission function,It is deck motion k Moment state estimation；According to describedRespectively obtain τ moment deck lengthwise movements and estimate information △ h_dIt is horizontal with deck lateral Motion estimation information △ y_d；

The computational methods of the Longitudinal Control Law module are as follows：

The longitudinal dispersion model ∑ of aircraft_lonFor：

Wherein, longitudinal state variable matrix △ x_blon=[△ V △ α △ q △ θ]^T, △ V are velocity feedback quantity and trim The difference being worth, △ α are angle of attack feedback quantity and the difference with level values, and △ q are pitch rate feedback quantity and the difference with level values, and △ θ are to bow Elevation angle feedback quantity and the difference with level values；△u_blonIt is longitudinally controlled input moment matrix, △ u_blon=[△ δ_e △δ_T]^T, △ δ_eTo rise Drop angle of rudder reflection increment, △ δ_TIt is throttle increment；△ h are practical flight height value；△ rin are the foreseeable preferable height that glides； △x_pIt is the information Store value of foreseeable preferable downslide height；△h_erIt is practical flight altitude feedback amount and the preferable height that glides Difference；It is T in the sampling time_s, preferable downslide height it is contemplated that time t_pWhen, it is contemplated that step number N=t_p/T_s；A_blon、B_blon、 A_eblonIt is longitudinal state space matrices of aircraft；A_p、B_p、C_pThe state space coefficient of delayer is walked for N；△ rin in formula (1) (k)=△ h_c(k+N) be the k moment preferable downslide height N step prediction value, △ h_cIdeal glides highly, △ x_pK () is expressed as：

Longitudinal Control Law module designs controller, full information robust preview control using full information robust preview control method Device is made up of feedback control component and feed-forward control component two parts；In the glide-slope tracking stage, by practical flight altitude feedback The difference △ h of amount and the preferable height that glides_erAnd longitudinal state variable error delta x_blonFeeding feedback control component, by foreseeable reason Think the information Store value △ x of downslide height △ rin and foreseeable preferable downslide height_pFeeding feed-forward control component, under realization Slideway height tracing, its control law is：

In formula, F_blonIt is longitudinal state variable error parameter matrix, F_elonIt is height error parameter matrix, F_plonManaged for N is walked Think downslide height preview information parameter matrix, F_rlonFor N walks preferable downslide height preview information parameter matrix；

In the deck motion compensation stage, by practical flight altitude feedback amount and the difference △ h of the preferable height that glides_er, longitudinal shape State variable error delta x_blonAnd information △ h are estimated in deck lengthwise movement_dFeeding feedback control component, by foreseeable preferable downslide Information △ h are estimated in height △ rin and deck lengthwise movement_dFeeding feed-forward control component, realizes glide path height tracing and deck Porpoising is compensated, and its control law is：

Wherein, F_plonJ () is the F_plonIn (j+1) item；Matrix F_blon,F_elon,F_plon,F_rlonComputing formula such as Under：

In formula,C_glon=[0I], W₁For longitudinal full information robust preview control is restrained Output weight, W₂For longitudinal full information robust preview control restrains input weight, F_glonIt is longitudinal state variable feedback coefficient matrix, j It is intermediate variable, Z is integer set,S、A_cgIt is intermediate variable coefficient matrix, X_ggIt is the discrete algebraically Riccati of formula (6) The steady state solution of equation；

When system meets three below constraints, and cause Control performance standard J_lonWhat is minimized is longitudinally controlled defeated Enter moment matrix △ u_blonIt is target control signal；Wherein, the Control performance standard J_lonFor：

The constraints is：

(1)(A_glon B_glon) it is stable；

(2)W₂'W₂>0；

(3) Full rank；

Wherein, χ is intermediate variable；

The computational methods of the lateral control law module of described horizontal stroke are as follows：

The lateral discretization model ∑ of horizontal stroke of aircraft_latFor：

Wherein, horizontal lateral state variable matrix △ x_blat=[△ β △ p △ r △ φ △ ψ]^T, △ β are yaw angle feedback Amount and the difference with level values, △ p are roll angle Rate Feedback amount and the difference with level values, and △ r are yawrate feedback quantity and trim The difference of value, △ φ are roll angle feedback quantity and the difference with level values, and △ ψ are yaw angle feedback quantity and the difference with level values；△u_blatFor Horizontal lateral control input moment matrix, △ u_blat=[△ δ_a △δ_r]^T, △ δ_aIt is aileron drift angle increment, △ δ_rFor rudder increases Amount；△ y be actual lateral deviation away from；△y_erFor actual lateral deviation away from feedback quantity and with level values difference；A_blat、B_blat、A_eblatIt is aircraft Horizontal lateral state space matrices；

Horizontal lateral control law module designs controller using full information robust preview control method, in glide-slope tracking rank Section, by horizontal lateral state variable error delta x_blatWith actual lateral deviation away from feedback quantity and with level values difference △ y_erFeeding feedback control point Measure to control the horizontal lateral movement of carrier-borne aircraft, make carrier-borne aircraft keep zero lateral deviation away from and lateral stability, its control law is：

In formula, F_blatIt is horizontal lateral state variable feedback coefficient matrix, F_elatIt is lateral deviation away from error feedback coefficient matrix；

In the deck motion compensation stage, by actual lateral deviation away from feedback quantity and the difference △ y with level values_er, horizontal lateral quantity of state misses Difference △ x_blatAnd information △ y are estimated in the horizontal lateral movement in deck_dFeeding feedback control component, information is estimated by the horizontal lateral movement in deck △y_dFeeding feed-forward control component, realize zero lateral deviation away from and horizontal lateral deck motion tracing compensation, its control law is：

In formula, F_platFor N walks lateral deviation away from preview information parameter matrix, F_rlatFor N walks lateral deviation away from preview information parameter square Battle array；Matrix F_blat,F_elat,F_plat,F_rlatComputing formula it is as follows：

Wherein,C_glat=[0 I], W₃It is horizontal lateral full information robust prediction control System rule output weight, W₄It is horizontal lateral full information robust preview control rule input weight；F_glatIt is horizontal lateral feedback factor matrix；S_g、A_cggIt is intermediate variable coefficient matrix；X_gIt is the steady state solution of the discrete algebraic riccati equation of formula (12)；

When system meets three below constraints, and cause Control performance standard J_latThe crosswise joint of minimum is defeated Enter moment matrix △ u_blatIt is target control signal；Wherein, the Control performance standard J_latFor：

The constraints is：

(1)(A_glat B_glat) it is stable；

(2)W₄'W₄>0；

(3) Full rank；

Wherein, χ is intermediate variable.

Beneficial effect：Regarding to the issue above, a kind of carrier-borne aircraft auto landing on deck control based on robust preview control of the invention Method processed, so as to be carried out to carrier-borne aircraft using the Future Information and present information of glide path track and longitudinal deck motion pre- See control, so that the compensation of tracking and longitudinal deck motion to glide path height is realized, while can effectively suppress and eliminate Laterally offset, makes carrier-borne aircraft keep lateral stability.Feedforward control wherein is carried out using Future Information, is carried out instead using current information Feedback control, rudder face and throttle that can in advance to carrier-borne aircraft implement average operation to reach tracing compensation purpose, reduce instantaneous Energy, and accelerate response speed, it is ensured that carrier-borne function on aircraft carrier safety warship.

Brief description of the drawings

Fig. 1 is the carrier-borne aircraft auto landing on deck control method composition structure chart based on robust preview control of the present invention；

Fig. 2 is longitudinal deck motion of the present invention and its curve estimated based on particle filter；

Fig. 3 is the carrier-borne aircraft auto landing on deck glide path height tracing curve based on robust preview control of the present invention；

Fig. 4 is that the carrier-borne aircraft auto landing on deck based on robust preview control of the present invention eliminates lateral deviation away from simulation curve；

Fig. 5 is the carrier-borne aircraft auto landing on deck longitudinal direction deck motion tracing compensation based on robust preview control of the present invention Curve.

Specific embodiment

The present invention is done below in conjunction with the accompanying drawings further is explained.

Understand technical scheme for the ease of the public, entered physical quantity of the present invention and parameter with table 1 below Row explanation：

Table 1

As shown in figure 1, a kind of carrier-borne aircraft auto landing on deck control method based on robust preview control includes that deck motion is pre- Estimate module, Longitudinal Control Law module and horizontal lateral control law module.Deck motion feeding deck motion prediction module is obtained first To deck motion prediction information, and deck motion prediction information is introduced carrier-borne aircraft auto landing on deck control system by 12.5 seconds before warship System.Wherein, the computational methods of deck motion prediction module are as follows：

For the deck motion of known transmission function expression-form, prediction device is designed using particle filter algorithm.First The state space equation of deck motion is obtained by transmission function：

In formula, x is deck motion state variable；ω is deck motion system dynamic noise；Z is observation signal；υ is observation Noise.

The discrete model of deck motion will be obtained after above formula discretization：

In formula, Φ_k,k-1It is deck motion state vector x from t_k-1Moment is transferred to t_kThe transfer matrix at moment,T_sIt is the sampling time；Γ_k,k-1It is t_k-1The system dynamic noise vector w at moment_k-1To t_kThe deck motion shape at moment State vector x_kThe noise coefficient matrix of influence,H_kIt is observed differential matrix；w_k-1It is system dynamic noise, Its variance matrix is Q_k-1；v_kIt is observation noise, its variance matrix is R_k。

Deck motion state variable x_kIt is in the expression formula of the optimal estimation value at following τ moment Wherein m=τ/T_s, state transfer matrixA is that the coefficient corresponding to deck motion transmission function is empty Between matrix,It is deck motion k moment state estimations.According toRespectively obtain τ moment deck lengthwise movements and estimate information △h_dInformation △ y are estimated with the horizontal lateral movement in deck_d。

The computational methods of the Longitudinal Control Law module are as follows：

First by longitudinal state equation discretization of aircraft, the longitudinal dispersion model ∑ of aircraft is obtained_lonFor：

Wherein, longitudinal state variable matrix △ x_blon=[△ V △ α △ q △ θ]^T, △ V are velocity feedback quantity and trim The difference being worth, △ α are angle of attack feedback quantity and the difference with level values, and △ q are pitch rate feedback quantity and the difference with level values, and △ θ are to bow Elevation angle feedback quantity and the difference with level values；△u_blonIt is longitudinally controlled input moment matrix, △ u_blon=[△ δ_e △δ_T]^T, △ δ_eTo rise Drop angle of rudder reflection increment, △ δ_TIt is throttle increment；△ h are practical flight height value；△ rin are the foreseeable preferable height that glides； △x_pIt is the information Store value of foreseeable preferable downslide height；△h_erIt is practical flight altitude feedback amount and the preferable height that glides Difference；It is T in the sampling time_s, preferable downslide height it is contemplated that time t_pWhen, it is contemplated that step number N=t_p/T_s；A_blon、B_blon、 A_eblonIt is longitudinal state space matrices of aircraft；A_p、B_p、C_pThe state space coefficient of delayer is walked for N,

△ rin (k)=△ h in formula (1)_c(k+N) be the k moment preferable downslide height N step prediction value, △ h_cDuring for k The preferable height, △ x of gliding at quarter_pK () is expressed as：

Longitudinal Control Law module designs controller, full information robust preview control using full information robust preview control method Device is made up of feedback control component and feed-forward control component two parts；In the glide-slope tracking stage, by practical flight altitude feedback The difference △ h of amount and the preferable height that glides_erAnd longitudinal state variable error delta x_blonFeeding feedback control component, by foreseeable reason Think the information Store value △ x of downslide height △ rin and foreseeable preferable downslide height_pFeeding feed-forward control component, by two Divide to combine and carry out full information prediction control, realize glide path height tracing, its control law is：

In formula, F_blonIt is longitudinal state variable error parameter matrix, F_elonIt is height error parameter matrix, F_plonManaged for N is walked Think downslide height preview information parameter matrix, F_rlonFor N walks preferable downslide height preview information parameter matrix；

In the deck motion compensation stage, by practical flight altitude feedback amount and the difference △ h of the preferable height that glides_er, longitudinal shape State variable error delta x_blonAnd information △ h are estimated in deck lengthwise movement_dFeeding feedback control component, by foreseeable preferable downslide Information △ h are estimated in height △ rin and deck lengthwise movement_dFeeding feed-forward control component, two parts is combined and is believed entirely Breath prediction control, realizes glide path height tracing and the compensation of deck porpoising, and its control law is：

Wherein, F_plonJ () is the F_plonIn (j+1) item；

By longitudinal dispersion model ∑_lonCan be derived by：

Matrix F_blon,F_elon,F_plon,F_rlonComputing formula it is as follows：

In formula,C_glon=[0I], W₁For longitudinal full information robust preview control is restrained Output weight, W₂For longitudinal full information robust preview control restrains input weight, F_glonIt is longitudinal state variable feedback coefficient matrix, j It is intermediate variable, Z is integer set,S、A_cgIt is intermediate variable coefficient matrix, X_ggIt is the discrete algebraically Riccati of formula (9) The steady state solution of equation；

When system meets three below constraints, and cause Control performance standard J_lonWhat is minimized is longitudinally controlled defeated Enter moment matrix △ u_blonIt is target control signal；Wherein, the Control performance standard J_lonFor：

The constraints is：

(1)(A_glon B_glon) it is stable；

(2)W₂'W₂>0；

(3) Full rank；

Wherein, χ is intermediate variable.

The computational methods of the lateral control law module of described horizontal stroke are as follows：

The lateral discretization model ∑ of horizontal stroke of aircraft_latFor：

Wherein, horizontal lateral state variable matrix △ x_blat=[△ β △ p △ r △ φ △ ψ]^T, △ β are yaw angle feedback Amount and the difference with level values, △ p are roll angle Rate Feedback amount and the difference with level values, and △ r are yawrate feedback quantity and trim The difference of value, △ φ are roll angle feedback quantity and the difference with level values, and △ ψ are yaw angle feedback quantity and the difference with level values；△u_blatFor Horizontal lateral control input moment matrix, △ u_blat=[△ δ_a △δ_r]^T, △ δ_aIt is aileron drift angle increment, △ δ_rFor rudder increases Amount；△ y be actual lateral deviation away from；△y_erFor actual lateral deviation away from feedback quantity and with level values difference；A_blat、B_blat、A_eblatIt is aircraft Horizontal lateral state space matrices；

Horizontal lateral control law module designs controller using full information robust preview control method, in glide-slope tracking rank Section, because the preferable lateral deviation in carrier-borne aircraft tracking glide path auto landing on deck stage is away from being always zero, preview information now can be neglected Slightly disregard, therefore by horizontal lateral state variable error delta x_blatWith actual lateral deviation away from feedback quantity and with level values difference △ y_erFeeding feedback Control component controls the horizontal lateral movement of carrier-borne aircraft, make carrier-borne aircraft keep zero lateral deviation away from and lateral stability, its control law is：

In formula, F_blatIt is horizontal lateral state variable feedback coefficient matrix, F_elatIt is lateral deviation away from error feedback coefficient matrix

In the deck motion compensation stage, by actual lateral deviation away from feedback quantity and the difference △ y with level values_er, horizontal lateral quantity of state misses Difference △ x_blatAnd information △ y are estimated in the horizontal lateral movement in deck_dFeeding feedback control component, information is estimated by the horizontal lateral movement in deck △y_dFeeding feed-forward control component, two parts are combined carries out full information prediction control, realize zero lateral deviation away from and it is horizontal lateral Deck motion tracing compensation, its control law is：

In formula, F_platFor N walks lateral deviation away from preview information parameter matrix, F_rlatFor N walks lateral deviation away from preview information parameter square Battle array；Matrix F_blat,F_elat,F_plat,F_rlatComputing formula it is as follows：

Wherein,C_glat=[0 I], W₃It is horizontal lateral full information robust preview control Rule output weight, W₄It is horizontal lateral full information robust preview control rule input weight；F_glatIt is horizontal lateral feedback factor matrix； It is coefficient matrix；S_g、A_cggIt is intermediate variable coefficient matrix；X_gIt is the steady state solution of the discrete algebraic riccati equation of formula (15)；

When system meets three below constraints, and cause Control performance standard J_latThe crosswise joint of minimum is defeated Enter moment matrix △ u_blatIt is target control signal；Wherein, the Control performance standard J_latFor：

The constraints is：

(1)(A_glat B_glat) it is stable；

(2)W₄'W₄>0；

(3) Full rank；

Wherein, χ is intermediate variable.

In order to verify validity of the present invention in the control of carrier-borne aircraft auto landing on deck, following emulation experiment is carried out.Emulation work Tool uses MATLAB softwares, and carrier-borne aircraft kinetic model uses the relevant parameter of F/A-18, aircraft carrier object to use " Niemi during analysis Hereby " number aircraft carrier, warship in emulation experiment using glide-slope tracking, and the glide-slope tracking time should be 56.3s, the inclination of glide path Angle is 3.5 °, the initial velocity V of carrier-borne aircraft₀It is 70m/s, constant angle of attack is 8.5 °, and the angle of pitch is 5 °, and elemental height is 240.7m, Away from being -1m, the sampling time is 0.1s to initial lateral deviation, it is therefore foreseen that step number is 12, and the prediction device estimated time is 1.2s.First in carrier-borne aircraft The auto landing on deck stage is chosen and matches somebody with somebody flat spot, as shown in table 2, corresponding longitudinal direction and horizontal lateral linear equation is obtained, then by above-mentioned tool Body implementation method carries out simulation calculation.

Table 2

V(m/s)	α(°)	q(°/s)	θ(°)	β(°)	p(°/s)	r(°/s)
							69.3	8.5	0	5	0	0	0
φ(°)	ψ(°)					h(m)
							0	0	-3.17	43	0	0	240.7

In emulation experiment by taking longitudinal deck motion as an example, estimated.Unit white noise is passed through into longitudinal deck motion to pass Delivery functionThe longitudinal deck motion information in time domain can be obtained, its Middle dynamic noise w_kPower is 1, observation noise v_kPower is 0.0225, is sent to estimate module, wherein observer matrix coefficient It is H_k=[1 00 0], obtain longitudinal deck motion to estimate information as shown in Figure 2.As shown in Figure 2, based on particle filter It is fine that the prediction device of design estimates effect.If the whole process of carrier-borne aircraft auto landing on deck does not introduce deck motion prediction information, obtain The simulation curve for arriving is as shown in Figure 3 and Figure 4.Wherein Fig. 3 is longitudinal height tracing simulation curve, as can be seen from Figure carrier-borne aircraft Under full information robust preview control warship height be almost highly completely superposed with preferable warship, tracking effect is very good；And Under PID control warship height there is obvious tracking error always in preceding 20s, and just start to eliminate tracking error after 20s, most Realize that glide-slope tracking warship eventually.Wherein Fig. 4 is horizontal lateral elimination lateral deviation simulation curve, as can be seen from Figure, carrier-borne Machine response time under full information robust preview control is realized reaching stable state without lateral deviation, i.e. tracking when being 15s；PID control Under lateral deviation be that 30s or so is realized without lateral deviation away from the response time.When deck motion prediction information is added warship by 12.5s before warship Emulated in carrier aircraft auto landing on deck system, by taking longitudinal deck motion as an example, obtained longitudinal deck motion tracing compensation curve such as Shown in Fig. 5.As shown in Figure 5, compared to PID control, the longitudinal deck motion tracking response based on full information robust preview control Faster, tracking effect is more preferable for speed.

Can be drawn by emulation experiment, a kind of carrier-borne aircraft auto landing on deck controlling party based on prediction control of the present invention Method can realize well carrier-borne aircraft auto landing on deck glide paths tracking and deck motion compensation, it can be ensured that carrier-borne aircraft safety On aircraft carrier warship, improve warship success rate.

The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should It is considered as protection scope of the present invention.

Claims (1)

1. a kind of carrier-borne aircraft auto landing on deck control method based on robust preview control, it is characterised in that：It is pre- including deck motion Estimate module, Longitudinal Control Law module and horizontal lateral control law module；Deck motion feeding deck motion prediction module is obtained first To deck motion prediction information, and deck motion prediction information is introduced into carrier-borne aircraft auto landing on deck control system before warship；Its In, the computational methods of the deck motion prediction module are as follows：

For the deck motion of known transmission function, prediction device, the first for estimating the τ moment are designed using particle filter algorithm Plate movable information, the expression formula for obtaining deck motion in the optimal estimation value at following τ moment isIts (k+m, is k) state-transition matrix to middle Φ, and k is time, m=τ/T_s, T_sIt is systematic sampling time, state transfer matrixA is the Coefficient Space matrix corresponding to deck motion transmission function,It is deck motion k Moment state estimation；According to describedRespectively obtain τ moment deck lengthwise movements and estimate information Δ h_dIt is horizontal with deck lateral Motion estimation information Δ y_d；

The computational methods of the Longitudinal Control Law module are as follows：

The longitudinal dispersion model ∑ of aircraft_lonFor：

${\mathrm{\Σ}}_{\mathrm{lon}}:\left\{\begin{array}{c}{\mathrm{\Δx}}_{\mathrm{blon}}(k+1)=A_{\mathrm{blon}}{\mathrm{\Δx}}_{\mathrm{blon}}\left(k\right)+B_{\mathrm{blon}}{\mathrm{\Δu}}_{\mathrm{blon}}\left(k\right)\\ \mathrm{\Δh}(k+1)=A_{\mathrm{eblon}}{\mathrm{\Δx}}_{\mathrm{blon}}\left(k\right)+\mathrm{\Δh}\left(k\right)\\ {\mathrm{\Δx}}_p(k+1)=A_p{\mathrm{\Δx}}_p\left(k\right)+B_p\mathrm{\Δrin}\left(k\right)\\ {\mathrm{\Δh}}_{\mathrm{er}}\left(k\right)=\mathrm{\Δh}\left(k\right)-C_p{\mathrm{\Δx}}_p\left(k\right)\end{array}\right.---\left(1\right)$

Wherein, longitudinal state variable matrix Δ x_blon=[Δ V Δ α Δ q Δs θ]^T, Δ V be velocity feedback quantity with level values it Difference, Δ α is angle of attack feedback quantity and the difference with level values, and Δ q is pitch rate feedback quantity and the difference with level values, and Δ θ is the angle of pitch Feedback quantity and the difference with level values；Δu_blonIt is longitudinally controlled input moment matrix, Δ u_blon=[Δ δ_e Δδ_T]^T, Δ δ_eIt is elevator Drift angle increment, Δ δ_TIt is throttle increment；Δ h is practical flight height value；Δ rin is the foreseeable preferable height that glides；Δx_pFor The information Store value of foreseeable preferable downslide height；Δh_erIt is practical flight altitude feedback amount and the difference of the preferable height that glides； It is T in the sampling time_s, preferable downslide height it is contemplated that time t_pWhen, it is contemplated that step number N=t_p/T_s；A_blon、B_blon、A_eblonFor Longitudinal state space matrices of aircraft；A_p、B_p、C_pThe state space coefficient of delayer is walked for N；Δ rin (k)=Δ in formula (1) h_c(k+N) be the k moment preferable downslide height N step prediction value, Δ h_cIdeal glides highly, Δ x_pK () is expressed as：

${\mathrm{\Δx}}_p\left(k\right)=\left[\begin{array}{c}\mathrm{\Δrin}(k-N)\\ .\\ .\\ .\\ \mathrm{\Δrin}(k-1)\end{array}\right]=\left[\begin{array}{c}\mathrm{\Δh}\left(k\right)\\ .\\ .\\ .\\ \mathrm{\Δh}(k+N-1)\end{array}\right]---\left(2\right)$

Longitudinal Control Law module using full information robust preview control method design controller, full information robust preview control device by Feedback control component and feed-forward control component two parts are constituted；In the glide-slope tracking stage, by practical flight altitude feedback amount with The difference Δ h of ideal downslide height_erAnd longitudinal state variable error delta x_blonFeeding feedback control component, by it is foreseeable ideally The information Store value Δ x of sliding height Δ rin and foreseeable preferable downslide height_pFeeding feed-forward control component, realizes glide path Height tracing, its control law is：

In formula, F_blonIt is longitudinal state variable error parameter matrix, F_elonIt is height error parameter matrix, F_plonFor N steps ideally Sliding height preview information parameter matrix, F_rlonFor N walks preferable downslide height preview information parameter matrix；

In the deck motion compensation stage, by practical flight altitude feedback amount and the difference Δ h of the preferable height that glides_er, longitudinal state becomes Amount error delta x_blonAnd information Δ h is estimated in deck lengthwise movement_dFeeding feedback control component, by the foreseeable preferable height that glides Information Δ h is estimated in Δ rin and deck lengthwise movement_dFeeding feed-forward control component, realizes that glide path height tracing and deck are drifted along Motion compensation, its control law is：

Wherein, F_plonJ () is the F_plonIn (j+1) item；Matrix F_blon,F_elon,F_plon,F_rlonComputing formula it is as follows：

$\left\{\begin{array}{c}\left[\begin{array}{cc}{F}_{\mathrm{blon}}& F_{\mathrm{elon}}\end{array}\right]=-{\stackrel{\&OverBar;}{R}}^{-1}\left(B_{\mathrm{glon}}^T{X}_{\mathrm{gg}}{A}_{\mathrm{glon}}\right)\\ F_{\mathrm{plon}}\left(0\right)=0\\ F_{\mathrm{plon}}\left(j\right)=-{\stackrel{\&OverBar;}{R}}^{-1}{B}_{\mathrm{glon}}^T{\left(A_{\mathrm{cg}}^T\right)}^{j-1}S,0<j<N,j\&Element;Z\\ F_{\mathrm{rlon}}=-{\stackrel{\&OverBar;}{R}}^{-1}{B}_{\mathrm{glon}}^T{\left(A_{\mathrm{cg}}^T\right)}^{N-1}S\\ \stackrel{\&OverBar;}{R}={W_2}^T{W}_2+B_{\mathrm{glon}}^T{X}_{\mathrm{gg}}{B}_{\mathrm{glon}}\\ S=C_{\mathrm{glon}}^T{W_1}^T{W}_1\\ A_{\mathrm{cg}}=A_{\mathrm{glon}}+B_{\mathrm{glon}}{F}_{\mathrm{glon}}\end{array}---\left(5\right)\right.$

In formula, $A_{\mathrm{glon}}=\left[\begin{array}{cc}{A}_{\mathrm{blon}}& 0\\ A_{\mathrm{eblon}}& I\end{array}\right],B_{\mathrm{glon}}=\left[\begin{array}{c}{B}_{\mathrm{blon}}\\ 0\end{array}\right],$ C_glon=[0 I], W₁For longitudinal full information robust preview control rule is defeated Go out weight, W₂For longitudinal full information robust preview control restrains input weight, F_glonIt is longitudinal state variable feedback coefficient matrix, j is Intermediate variable, Z is integer set,S、A_cgIt is intermediate variable coefficient matrix, X_ggIt is the discrete algebraically Riccati of formula (6) The steady state solution of equation；

$X_{\mathrm{gg}}={A_{\mathrm{glon}}}^{\&prime;}{X}_{\mathrm{gg}}{A}_{\mathrm{glon}}-{F_{\mathrm{glon}}}^{\&prime;}\stackrel{\&OverBar;}{R}{F}_{\mathrm{glon}}+{C_{\mathrm{glon}}}^{\&prime;}{W_1}^{\&prime;}{W}_1{C}_{\mathrm{glon}}---\left(6\right)$

When system meets three below constraints, and cause Control performance standard J_lonThe longitudinally controlled input quantity for minimizing Matrix Δ u_blonIt is target control signal；Wherein, the Control performance standard J_lonFor：

$J_{\mathrm{lon}}=\underset{k=-N+1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}[{{\mathrm{\&Delta;h}}_{\mathrm{er}}}^T\left(k\right){W_1}^T{W}_1{\mathrm{\&Delta;h}}_{\mathrm{er}}\left(k\right)+{\mathrm{\&Delta;u}}_{\mathrm{blon}}^T\left(k\right){W_2}^T{W}_2{\mathrm{\&Delta;u}}_{\mathrm{blon}}\left(k\right)]---\left(7\right)$

The constraints is：

(1)(A_glon B_glon) it is stable；

(2)W₂'W₂>0；

$\left(3\right)---\&ForAll;\mathrm{\&chi;}\&Element;(-\mathrm{\&pi;},\mathrm{\&pi;}],\left[\begin{array}{cc}{A}_{\mathrm{glon}}-e^{\mathrm{j\&chi;}}I& B_{\mathrm{glon}}\\ C_{\mathrm{glon}}& 0\end{array}\right]$ Full rank；

Wherein, χ is intermediate variable；

The computational methods of the lateral control law module of described horizontal stroke are as follows：

The lateral discretization model ∑ of horizontal stroke of aircraft_latFor：

${\mathrm{\&Sigma;}}_{\mathrm{lat}}:\left\{\begin{array}{c}{\mathrm{\&Delta;x}}_{\mathrm{blat}}(k+1)=A_{\mathrm{blat}}{\mathrm{\&Delta;x}}_{\mathrm{blat}}\left(k\right)+B_{\mathrm{blat}}{\mathrm{\&Delta;u}}_{\mathrm{blat}}\left(k\right)\\ \mathrm{\&Delta;y}(k+1)=A_{\mathrm{eblat}}{\mathrm{\&Delta;x}}_{\mathrm{blat}}\left(k\right)+\mathrm{\&Delta;y}\left(k\right)\\ {\mathrm{\&Delta;y}}_{\mathrm{er}}\left(k\right)=\mathrm{\&Delta;y}\left(k\right)\end{array}\right.---\left(8\right)$

Wherein, horizontal lateral state variable matrix Δ x_blat=[Δ β Δ p Δ r Δ φ Δs ψ]^T, Δ β be yaw angle feedback quantity with Difference with level values, Δ p be roll angle Rate Feedback amount with level values difference, Δ r be yawrate feedback quantity with match somebody with somebody level values it Difference, Δ φ is roll angle feedback quantity and the difference with level values, and Δ ψ is yaw angle feedback quantity and the difference with level values；Δu_blatIt is horizontal side To control input moment matrix, Δ u_blat=[Δ δ_a Δδ_r]^T, Δ δ_aIt is aileron drift angle increment, Δ δ_rIt is rudder increment； Δ y be actual lateral deviation away from；Δy_erFor actual lateral deviation away from feedback quantity and with level values difference；A_blat、B_blat、A_eblatIt is the horizontal side of aircraft To state space matrices；

Horizontal lateral control law module designs controller using full information robust preview control method, in the glide-slope tracking stage, will Horizontal lateral state variable error delta x_blatWith actual lateral deviation away from feedback quantity and with level values difference Δ y_erFeeding feedback control component comes Control the horizontal lateral movement of carrier-borne aircraft, make carrier-borne aircraft keep zero lateral deviation away from and lateral stability, its control law is：

In formula, F_blatIt is horizontal lateral state variable feedback coefficient matrix, F_elatIt is lateral deviation away from error feedback coefficient matrix；

In the deck motion compensation stage, by actual lateral deviation away from feedback quantity and the difference Δ y with level values_er, horizontal lateral quantity of state error delta x_blatAnd information Δ y is estimated in the horizontal lateral movement in deck_dFeeding feedback control component, information Δ y is estimated by the horizontal lateral movement in deck_d Feeding feed-forward control component, realize zero lateral deviation away from and horizontal lateral deck motion tracing compensation, its control law is：

In formula, F_platFor N walks lateral deviation away from preview information parameter matrix, F_rlatFor N walks lateral deviation away from preview information parameter matrix；Square Battle array F_blat,F_elat,F_plat,F_rlatComputing formula it is as follows：

$\left\{\begin{array}{c}\left[\begin{array}{cc}{F}_{\mathrm{blat}}& F_{\mathrm{elat}}\end{array}\right]=-{\stackrel{\&OverBar;}{R}}_g^{-1}\left(B_{\mathrm{glat}}^T{X}_g{A}_{\mathrm{glat}}\right)\\ F_{\mathrm{plat}}\left(0\right)=0\\ F_{\mathrm{plat}}\left(j\right)=-{\stackrel{\&OverBar;}{R}}_g^{-1}{B}_{\mathrm{glat}}^T{\left(A_{\mathrm{cgg}}^T\right)}^{j-1}{S}_g,0<j<N,j\&Element;Z\\ F_{\mathrm{rlat}}=-{\stackrel{\&OverBar;}{R}}_g^{-1}{B}_{\mathrm{glat}}^T{\left(A_{\mathrm{cgg}}^T\right)}^{N-1}{S}_g\\ {\stackrel{\&OverBar;}{R}}_g=W_4^T{W}_4+B_{\mathrm{glat}}^T{X}_g{B}_{\mathrm{glat}}\\ S_g=C_{\mathrm{glat}}^T{W_3}^T{W}_3\\ A_{\mathrm{cgg}}=A_{\mathrm{glat}}+B_{\mathrm{glat}}{F}_{\mathrm{glat}}\end{array}---\left(11\right)\right.$

Wherein, $A_{\mathrm{glat}}=\left[\begin{array}{cc}{A}_{\mathrm{blat}}& 0\\ A_{\mathrm{eblat}}& I\end{array}\right],B_{\mathrm{glat}}=\left[\begin{array}{c}{B}_{\mathrm{blat}}\\ 0\end{array}\right],$ C_glat=[0 I], W₃It is horizontal lateral full information robust preview control rule Output weight, W₄It is horizontal lateral full information robust preview control rule input weight；F_glatIt is horizontal lateral feedback factor matrix； S_g、A_cggIt is intermediate variable coefficient matrix；X_gIt is the steady state solution of the discrete algebraic riccati equation of formula (12)；

$X_g={A_{\mathrm{glat}}}^{\&prime;}{X}_g{A}_{\mathrm{glat}}-{F_{\mathrm{glat}}}^{\&prime;}{{\stackrel{\&OverBar;}{R}}_{g}F}_{\mathrm{glat}}+{C_{\mathrm{glat}}}^{\&prime;}{W_3}^{\&prime;}{W}_3{C}_{\mathrm{glat}}---\left(12\right)$

When system meets three below constraints, and cause Control performance standard J_latThe crosswise joint input quantity of minimum Matrix Δ u_blatIt is target control signal；Wherein, the Control performance standard J_latFor：

$J_{\mathrm{lat}}=\underset{k=-N+1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}[{{\mathrm{\&Delta;y}}_{\mathrm{er}}}^T\left(k\right){W_3}^T{W}_3{\mathrm{\&Delta;y}}_{\mathrm{er}}\left(k\right)+{\mathrm{\&Delta;u}}_{\mathrm{blat}}^T\left(k\right){W_4}^T{W}_4{\mathrm{\&Delta;u}}_{\mathrm{blat}}\left(k\right)]---\left(13\right)$

The constraints is：

(1) (A_glat B_glat) it is stable；

(2) W₄'W₄>0；

$\left(3\right)---\&ForAll;\mathrm{\&chi;}\&Element;(-\mathrm{\&pi;},\mathrm{\&pi;}],\left[\begin{array}{cc}{A}_{\mathrm{glat}}-e^{\mathrm{j\&chi;}}I& B_{\mathrm{glat}}\\ C_{\mathrm{glat}}& 0\end{array}\right]$ Full rank；

Wherein, χ is intermediate variable.