CN106200383A - A kind of three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net - Google Patents

A kind of three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net Download PDF

Info

Publication number
CN106200383A
CN106200383A CN201610645413.4A CN201610645413A CN106200383A CN 106200383 A CN106200383 A CN 106200383A CN 201610645413 A CN201610645413 A CN 201610645413A CN 106200383 A CN106200383 A CN 106200383A
Authority
CN
China
Prior art keywords
control
adaptive
neural network
axle inertially
inertially stabilized
Prior art date
Application number
CN201610645413.4A
Other languages
Chinese (zh)
Other versions
CN106200383B (en
Inventor
李志毅
Original Assignee
北京宇鹰科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京宇鹰科技有限公司 filed Critical 北京宇鹰科技有限公司
Priority to CN201610645413.4A priority Critical patent/CN106200383B/en
Publication of CN106200383A publication Critical patent/CN106200383A/en
Application granted granted Critical
Publication of CN106200383B publication Critical patent/CN106200383B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance

Abstract

A kind of three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net, relate to the Disturbance Rejection design of control parameter On-line Estimation based on adaptive neural network and expansion state state observer, first, according to three axle inertially stabilized platform kinetic models, for the three uncertain feedback control parameters time-varying characteristics caused of axle inertially stabilized platform model parameter, build adaptive neural network and feedback control matrix parameter is carried out On-line Estimation, make three axle inertially stabilized platform control accuracies approach expectational model control accuracy;Secondly, the interference compensation controlled quentity controlled variable that the upper bound disturbed for adaptive neural network estimation difference and three axle inertially stabilized platforms the builds impact on control accuracy, build extended state observer interference is estimated and suppresses, it is achieved three axle inertially stabilized platform high accuracy under complex environment control.The present invention has that real-time is good, dynamic parameter response is fast, to advantages such as multi-source interference strong adaptabilities, can be used for the high accuracy control etc. under complicated multi-source interference environment of the three axle inertially stabilized platforms.

Description

A kind of three axle inertially stabilized platforms based on model reference adaptive neutral net control Method
Technical field
The present invention relates to a kind of three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net, suitable For aerial mapping stabilized platform high accuracy control field.
Background technology
Three axle gondola platforms are fixed on flight carrier by pedestal, support and stablize remotely sensed image load, and isolation multi-source is done Disturb the impact on the remotely sensed image load optical axis, improve remote sensing load pointing accuracy over the ground, be with a wide range of applications.
Three axle inertially stabilized platform interference types in the course of the work are various, do not only exist uncertain RANDOM WIND disturb and By aircraft body vibration caused angular movement interference, there is also owing to load system barycenter causes not with system origin is misaligned Trimming moment, the internal interference that platform sensor part measurement error causes, therefore, complexity disturbs the high accuracy controlling party under environment more Method has become as one of key technology of three axle inertially stabilized platform researchs.
For improving performance, PID control method, robust control, intelligent control method are used for three axle inertially stabilized platforms High accuracy controls.PID controller simple in construction, but poor anti jamming capability, it is difficult to ensure that three axle inertia under multi-source interference effect Platform stable precision.Robust control can disturb the impact on control accuracy with suppression system model parameter uncertainty and multi-source, But control accuracy has bigger conservative.By substantial amounts of sample training, neutral net infinite can approach nonlinear system, Thus solve the non-linear of three axle inertially stabilized platform models, it is achieved high-precision gesture stability, but traditional neutral net needs Substantial amounts of sample data is wanted to be trained, the shortcoming with poor real.
Summary of the invention
Present invention solves the technical problem that and be: three axle inertially stabilized platform control accuracies are non-linear and multi-source by system model The impact of interference, is estimated that by adaptive neural network feedback control parameters solves the non-linear of model in real time, and builds expansion Observer improves the capacity of resisting disturbance of system, it is achieved three axle inertially stabilized platform high accuracy under complex environment control.
The technical solution of invention is: set up kinetic model, according to expectation first against three axle inertially stabilized platforms System performance index, builds three axle inertially stabilized platform reference driving force models, by adaptive network in line tracking feedback control Parameter processed, makes three axle inertially stabilized platform kinetic models approach and reference model;Secondly, build expansion observer, reduce certainly The sign function gain that the upper bound of adaptation neutral net estimation difference and external disturbance the builds impact on system control accuracy.Its Realize step as follows:
(1) according to three axle inertially stabilized platform kinetic models, design a model reference adaptive neural network control method, For the three uncertain feedback control parameters time-varying characteristics caused of axle inertially stabilized platform model parameter, build adaptive neural network net Network carries out On-line Estimation to feedback control matrix parameter, makes three axle inertially stabilized platform control accuracies approach expectational model and controls essence Degree;
(2) upper bound disturbed for adaptive neural network estimation difference and three axle inertially stabilized platforms build interference Compensate the controlled quentity controlled variable impact on system control accuracy, build extended state observer and interference is estimated and suppresses, it is achieved be multiple Three axle inertially stabilized platform high accuracy under heterocycle border control;
The three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net of the present invention, wherein said Step (1) control based on model reference adaptive ANN Control input uj, adaptive neural network more new lawWith dry Disturb compensation controlled quentity controlled variable rjExpression formula be respectively
u j = W ^ j * · Θ ( x ) x + u j m m + r j - g j b j , j = 1 , 2 , 3
W ^ · j * = - e T PB * j x T Θ T ( x ) Γ
rj=-δjsgn(eTPB*j)
Wherein, represent roll frame, j=2 pitching frame during j=1, during j=3, represent orientation frame,It it is adaptive neural network The weight matrix of network,It is adaptive neural network ideal weight matrix W*The real-time estimated value of jth row,It is The basic function of adaptive neural network, n is the dimension of system model, and m is the dimension of system input, and l is adaptive neural network The nodes of hidden layer,For the state variable of system, bjPreferable control for the coefficient matrix B corresponding j row of framework input Coefficient processed, gjIt is the estimated value of system j framework inertia inertia, uj mmIt is the expectation input of system reference model j framework, with reference to mould Type is
x · m m = A m m x m m + Bu m m
Wherein,For the state equation matrix of reference model, WithIt is the matrix artificially designed according to three axle inertially stabilized platform dynamic performance requirements, k=1,2,3 Representing corresponding roll frame respectively, pitching frame and orientation frame, e is state variable x of current state amount x and reference modelmmDifference Value, matrix P is the positive definite symmetric solution of state equation,
PAmm+AmmTP=-Q
WhereinIt is positive definite symmetrical matrix,Being adaptive law gain matrix, l is adaptive neural network The nodes of hidden layer, B*jBeing the jth row of the coefficient matrix B of framework input, adaptive neural network estimation difference and the external world disturb The dynamic upper bound is
δj>|εjx+dj/bj|
Wherein, εjIt is the approximate error of adaptive neural network, djInterference for system j framework;
The three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net of the present invention, wherein said Extended state observer that step (2) builds and revised interference compensation controlled quentity controlled variable rjExpression formula be respectively
z ^ · 1 = A m z 1 - B W ^ · Θ ( x ) x + B r + z ^ 2 - 2 w 0 ( z ^ 1 - z 1 )
z ^ · 2 = - w 0 2 ( z ^ 1 - z 1 )
r j = - ( z ^ 2 ) j b j - ξ j sgn ( e T PB * j ) b j
Wherein, z1=e, z2=Δ=B ε x+Dd,WithIt is respectively state variable z1And z2Estimated value, w0> 0 be design Variable, can makeWithTight tracking e and Δ within the limited time, revised interference compensation controlled quentity controlled variable rjExpression formula For
r j = - ( z ^ 2 ) j + 3 b j - ξ j sgn ( e T PB * j ) b j
Wherein,Residual error B ε x and system interference Dd and disturbance estimated value z is estimated for system self-adaption neutral net2 Residual error
Corresponding system input is
u j = ( W ^ j T · Θ j ) T x + u j m - g j b j - ( z ^ 2 ) j + 3 b j - δ j sgn ( e T PB * j ) b j , j = 1 , 2 , 3 ,
Wherein,It is+ 3 components of jth.
Present invention advantage compared with prior art is:
(1) the present invention infinite extensive approximation capability by adaptive neural network, the feedback control of real-time estimating system Parameter, solves the feedback control parameters time-variant nonlinear problem caused owing to model parameter is non-linear, uncertain, makes actual mould Type approaches reference model, control method simple in construction, and capacity of resisting disturbance is strong;
(2) present invention is in the case of adaptive neural network ensures system stability, estimates further with expansion observer Counting and suppress in three axle inertially stabilized platform work process suffered disturbance, control accuracy is high, it is possible to meet three axle stable inertias The high accuracy demand for control of platform;
(3) present invention only requires according to the status information in three axle inertially stabilized platform work process, utilize Li Yapunuo Husband's function design adaptive neural network right value update matrix, can be with the weights of online updating adaptive neural network, it is not necessary to Any sample training, has data acquisition convenience, calculates simple advantage.
Accompanying drawing explanation
Fig. 1 is three axle inertially stabilized platform control flows;
Fig. 2 is that in flight experiment, three axle inertially stabilized platform pitch channels control effect;
Fig. 3 is that in flight experiment, three axle inertially stabilized platform roll passages control effect;
Fig. 4 is that in flight experiment, three axle inertially stabilized platform azimuthal channel control effect.
Detailed description of the invention
As it is shown in figure 1, the present invention's is implemented as follows
(1) build based on model reference adaptive neutral net
Based on Newton-Euler equation, the kinetics equation of three axle inertially stabilized platforms is expressed as
x · = A x + B u + D ( g + d )
Wherein, x=[θj ωj]T,H=03×3, F=(fjk), j, k=1,2,3,
U=[u1 u2 u3]T,G=[g1 g2 g3]T, d=[d1 d2 d3]T,
Wherein, represent roll frame, j=2 pitching frame during j=1, during j=3, represent orientation frame,State for system Variable, n=6 is the dimension of state variable, θjFor corresponding j framework angle, ωjFor corresponding j frame corners speed,For The coefficient matrix of state variable,For the coefficient matrix of framework input, m=3 is the dimension of system input, ujFor accordingly J frame voltage input, g is the estimated value of the gimbal moment of intertia, d be system framework interference, system interference, frame member be used to The perturbation value △ g of amountjWith the control input perturbation Δ b caused by measurement noisejConstitute, bjCoefficient matrix B phase for framework input Answering the preferable control coefrficient that j arranges, F is corresponding perfect condition variation coefficient matrix, wherein
f 11 = - K t K e N 2 J 2 R m , f 12 = - ( J p z - J p y - J a y ) ( ω i r z r c o s 2 θ p - ω i r y r sinθ p cosθ p ) J 2 ,
f 13 = J a z ω i p y p J 2 , f 21 = - ( J a x + J p x ) ω i r z r J 3 , f 23 = J a z ω i r x r cosθ p J 3 ,
f 22 = - ( K t K e N 2 J 3 R m + 2 J a z θ · p cosθ p sinθ p + ( J p z - J a y - J p y ) ( ω i r x r + 2 θ · p ) cosθ p sinθ p J 3 ) ,
f 31 = f 32 = 0 , f 33 = - K t K e N 2 J 1 R m ;
b ‾ 1 = NK t J 2 R m , b ‾ 2 = NK t J 3 R m , b ‾ 3 = NK t J 1 R m ;
g ‾ 1 = K t K e N 2 ω i b x p + N ( N - 1 ) R m J m ω · i b x p J 2 R m ,
g ‾ 2 = K t K e N 2 ω i b y r + N ( N - 1 ) R m J m ω · i b y t J 3 R m , g ‾ 3 = K t K e N 2 ω i b z a + N ( N - 1 ) R m J m ω · i b z a J 1 R m ;
d ‾ 1 = - ( J p z - J p y - J a y ) ω i r z r 2 sinθ p cosθ p + NT d m + T d p J 2 ,
d ‾ 2 = ( J a y + J p y ) sin 2 θ p + J p z cos 2 θ p - J r x + J r z J 3 ω i r z r ω i r x r + NT d m + T d r - [ ( J a y + J p y - J a z - J p z ) ω i r z r cosθ p sinθ p - sinθ p J a z θ · a ] ′ J 3 ,
d ‾ 3 = NT d m + T d a J 1
Wherein, N represents motor-driven ratio, KeRepresent back electromotive force constant, KtRepresent motor torque coefficient, RmRepresent motor Resistance, JmRepresent the rotary inertia of motor, Ja=diag (Jax,Jay,Jaz) it is that the rotary inertia of orientation frame is in orientation frame coordinate system Projection on x, y, z direction, Jp=diag (Jpx,Jpy,Jpz) it is that the rotary inertia of pitching frame is in pitching frame coordinate system x, y, z side Projection upwards, Jr=diag (Jrx,Jry,Jrz) be the rotary inertia of roll frame at roll frame in roll coordinate system x, y, z direction On projection,K=b, r, p, a are that the angular velocity in k system relative inertness space is in k system x, y, z side Downward mapping;It is respectively roll frame opposite base to exist relative to roll frame at roll coordinate system, pitching frame Pitching coordinate system, orientation frame relative to pitching frame at the angular velocity of azimuthal coordinates system, θrRepresent roll frame opposite base corner, θpTable Show the pitching frame corner relative to roll frame, θaRepresent the orientation frame corner relative to pitching frame, by the code-disc being arranged on gimbal axis Measurement obtains, TdmFor acting on the disturbance torque on motor, TdjFor acting on the disturbance torque on j framework, rotary inertia is
J1=Jaz+N2Jm,
J2=Jpx+Jax+N2Jm,
J3=Jry+(Jay+Jpy)cos2θp+(Jaz+Jpz)sin2θp+N2Jm
Design reference model
x · m m = A m m x m m + Bu m m
WhereinFor the state equation matrix of reference model,WithIt is the matrix artificially designed based on POLE PLACEMENT USING principle according to dynamic performance requirement,It is the control input of desired reference model,
Design of feedback controller
u j * = K j * T x + u j m m + r j - g j b j , j = 1 , 2 , 3
Wherein r=[r1 r2 r3]TIt is the controlled quentity controlled variable designed for compensating interference d, meets Feedback oscillator, then system
x · = ( A + B K ) x + Bu m m + B r + D d
If A+BK=Amm, then real system is infinite approaches desired reference model;
Therefore, feedback oscillator
K j * * = 1 b j ( h j * m m f j * m m - h j * f j * )
Due to hjk,fjk, k, j=1,2,3 is time-variant nonlinear function,Also it is time-variant nonlinear function, therefore, utilizes The infinite extensive approximation capability of adaptive neural network carrys out On-line Estimation time-variant nonlinear function
K j * * = W j * * · Θ ( x ) + ϵ j , j = 1 , 2 , 3
Wherein,It is the weight matrix of adaptive neural network,It it is adaptive neural network ideal weights square Battle arrayThe real-time estimated value of jth row,Being the basic function of adaptive neural network, l is that adaptive neural network is hidden Nodes containing layer, εjIt it is the approximate error of adaptive neural network;
Therefore, control based on model reference adaptive ANN Control input uj, adaptive neural network more new lawWith interference compensation controlled quentity controlled variable rjExpression formula be respectively
u j = W ^ j * · Θ ( x ) x + u j m m + r j - g j b j
W ^ · j * = - e T PB * j x T Θ T ( x ) Γ
rj=-δjsgn(eTPB*j)
Wherein, e is state variable x of current state amount x and expectational modelmmDifference, matrix P is the positive definite of state equation Symmetric solution,
PAmm+AmmTP=-Q
WhereinIt is positive definite symmetrical matrix,It is adaptive law gain matrix, B*jIt is the jth row of B matrix, The upper bound of adaptive neural network estimation difference and external disturbance is
δj>|εjx+dj/bj|
Wherein, εjIt is the approximate error of adaptive neural network, djInterference for system j framework;
(2) adaptive neural network is built
Interference compensation controlled quentity controlled variable r built for the upper bound of adaptive neural network estimation difference and external disturbancejTo being The impact of system control accuracy, builds extended state observer and estimates interference and suppress, it is achieved three axles under complex environment Inertially stabilized platform high accuracy controls,
Definition status variable z1=e, z2=Δ=B ε x+Dd,
Then three axle inertially stabilized platform error state equation are
z · 1 = A m z 1 - B W ^ · · Θ ( x ) x + B r + z 2
By z2Introduce error state equation, and
Wherein, s is the rate of change of Δ, and the extended state observer of structure is
z ^ · 1 = A m z 1 - B W ^ · Θ ( x ) x + B r + z ^ 2 - 2 w 0 ( z ^ 1 - z 1 )
z ^ · 2 = - w 0 2 ( z ^ 1 - z 1 )
Wherein,WithIt is respectively state variable z1And z2Estimated value, w0> 0 it is design variable;
By suitable Selecting All Parameters w0Can makeWithTight tracking e and Δ within the limited time,
Revised interference compensation controlled quentity controlled variable rjExpression formula be
r j = - ( z ^ 2 ) j b j - ξ j sgn ( e T PB * j ) b j
Wherein,Residual error B ε x and system interference Dd and disturbance estimated value z is estimated for system self-adaption neutral net2 Residual error
Corresponding system input is
u j = ( W ^ j T · Θ j ) T x + u j m - g j b j - ( z ^ 2 ) j + 3 b j - δ j sgn ( e T PB * j ) b j , j = 1 , 2 , 3
Wherein,It is+ 3 components of jth.
(3) flight example
In flight course, according to the angle information of high-precision attitude measuring unit, three axle inertially stabilized platform framework systems System is adjusted correspondingly, it is ensured that the remote sensing load optical axis is vertical over the ground, flight result such as Fig. 2, Fig. 3 and Fig. 4 institute of certain experiment Show.
Three axle inertially stabilized platforms achieve high-precision control, and the standard deviation of pitch channel is 0.0183 degree, and roll is led to The standard deviation in road is 0.0157, and the standard deviation of azimuthal channel is 0.0214.
The present invention three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net overcome existing The deficiency of control method, it is possible to achieve three axle inertially stabilized platforms high accuracy under complexity disturbs environment more controls.
The content not being described in detail in description of the invention belongs to prior art known to professional and technical personnel in the field.

Claims (3)

1. three axle Inertially-stabilizeplatform platform control method based on model reference adaptive neutral net, it is characterised in that realize Following steps:
(1) according to three axle inertially stabilized platform kinetic models, design a model reference adaptive neural network control method, for The three uncertain feedback control parameters time-varying characteristics caused of axle inertially stabilized platform model parameter, build adaptive neural network pair Feedback control matrix parameter carries out On-line Estimation, makes three axle inertially stabilized platform control accuracies approach expectational model control accuracy;
(2) the interference compensation control that the upper bound disturbed for adaptive neural network estimation difference and three axle inertially stabilized platforms builds The amount processed impact on system control accuracy, builds extended state observer and estimates interference and suppress, it is achieved complex environment Under three axle inertially stabilized platforms high accuracy control.
Three axle inertially stabilized platform controlling parties based on model reference adaptive neutral net the most according to claim 1 Method, it is characterised in that: described step (1) control based on model reference adaptive ANN Control input uj, self adaptation god Through network more new lawWith interference compensation controlled quentity controlled variable rjExpression formula be respectively
rj=-δjsgn(eTPB*j)
Wherein, represent roll frame, j=2 pitching frame during j=1, during j=3, represent orientation frame,It it is adaptive neural network Weight matrix,It is adaptive neural network ideal weight matrix W*The real-time estimated value of jth row,It is adaptive Answering the basic function of neutral net, m is the dimension of system input, and l is the nodes of adaptive neural network hidden layer, and n is system The dimension of model,For the state variable of system, bjThe preferable of coefficient matrix B corresponding j row for framework input controls system Number, gjIt is the estimated value of system j framework inertia inertia, uj mmBeing the expectation input of system reference model j framework, reference model is
Wherein,For the state equation matrix of reference model,WithIt is the matrix artificially designed according to three axle inertially stabilized platform dynamic performance requirements, k=1,2,3 points Not representing corresponding roll frame, pitching frame and orientation frame, e is state variable x of current state amount x and reference modelmmDifference, Matrix P is the positive definite symmetric solution of state equation,
PAmm+AmmTP=-Q
WhereinIt is positive definite symmetrical matrix,Being adaptive law gain matrix, l is that adaptive neural network implies The nodes of layer, B*jIt is the jth row of the coefficient matrix B of framework input, adaptive neural network estimation difference and external disturbance The upper bound is
δj>|εjx+dj/bj|
Wherein, εjIt is the approximate error of adaptive neural network, djInterference for system j framework.
Three axle inertially stabilized platform controlling parties based on model reference adaptive neutral net the most according to claim 1 Method, it is characterised in that: extended state observer that described step (2) builds and revised interference compensation controlled quentity controlled variable rjExpression Formula is respectively
Wherein, z1=e, z2=Δ=B ε x+Dd,WithIt is respectively state variable z1And z2Estimated value, w0> 0 it is design variable, Can makeWithTight tracking e and Δ within the limited time, revised interference compensation controlled quentity controlled variable rjExpression formula be
Wherein,Residual error B ε x and system interference Dd and disturbance estimated value z is estimated for system self-adaption neutral net2Residual Difference
Corresponding system input is
Wherein,It is+ 3 components of jth.
CN201610645413.4A 2016-08-08 2016-08-08 A kind of three axis Inertially-stabilizeplatform platform control methods based on model reference adaptive neural network Active CN106200383B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610645413.4A CN106200383B (en) 2016-08-08 2016-08-08 A kind of three axis Inertially-stabilizeplatform platform control methods based on model reference adaptive neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610645413.4A CN106200383B (en) 2016-08-08 2016-08-08 A kind of three axis Inertially-stabilizeplatform platform control methods based on model reference adaptive neural network

Publications (2)

Publication Number Publication Date
CN106200383A true CN106200383A (en) 2016-12-07
CN106200383B CN106200383B (en) 2019-10-18

Family

ID=57514611

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610645413.4A Active CN106200383B (en) 2016-08-08 2016-08-08 A kind of three axis Inertially-stabilizeplatform platform control methods based on model reference adaptive neural network

Country Status (1)

Country Link
CN (1) CN106200383B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107894713A (en) * 2017-10-20 2018-04-10 东南大学 A kind of high-accuracy control method without coding two axle inertially stabilized platforms of sensing
CN108469269A (en) * 2018-01-19 2018-08-31 天津大学 A kind of resonance point test system of wideband inertial reference stabilized platform

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1790197A (en) * 2005-12-22 2006-06-21 安徽工业大学 Simple method for neural network decoupling of multi-variable system based on model reference adaptive control
CN102053628A (en) * 2009-10-27 2011-05-11 北京航空航天大学 Neural network-based servo control system and method
CN102298315A (en) * 2011-06-21 2011-12-28 河海大学常州校区 Adaptive control system based on radial basis function (RBF) neural network sliding mode control for micro-electromechanical system (MEMS) gyroscope
CN102437816A (en) * 2011-10-25 2012-05-02 武汉鑫通科创科技发展有限公司 Adaptive motor motion control apparatus based on neural network
CN102508503A (en) * 2011-11-01 2012-06-20 北京航空航天大学 Compensation method based on generalized inner module for eccentric torque of three-shaft inertially stabilized platform
CN105785762A (en) * 2016-03-17 2016-07-20 北京航空航天大学 Bi-axis inertially-stabilized platform high-precision control method based on self-adaptive backstepping sliding mode

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1790197A (en) * 2005-12-22 2006-06-21 安徽工业大学 Simple method for neural network decoupling of multi-variable system based on model reference adaptive control
CN102053628A (en) * 2009-10-27 2011-05-11 北京航空航天大学 Neural network-based servo control system and method
CN102298315A (en) * 2011-06-21 2011-12-28 河海大学常州校区 Adaptive control system based on radial basis function (RBF) neural network sliding mode control for micro-electromechanical system (MEMS) gyroscope
CN102437816A (en) * 2011-10-25 2012-05-02 武汉鑫通科创科技发展有限公司 Adaptive motor motion control apparatus based on neural network
CN102508503A (en) * 2011-11-01 2012-06-20 北京航空航天大学 Compensation method based on generalized inner module for eccentric torque of three-shaft inertially stabilized platform
CN105785762A (en) * 2016-03-17 2016-07-20 北京航空航天大学 Bi-axis inertially-stabilized platform high-precision control method based on self-adaptive backstepping sliding mode

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
YINGZOU,等: "A compound control method based on the adaptive neural network and sliding mode control for inertial stable platform", 《NEUROCOMPUTING》 *
夏青元,等: "三轴式无人旋翼飞行器及自适应飞行控制系统设计", 《航空学报》 *
钟麦英,等: "基于PMI的三轴惯性稳定平台干扰力矩补偿方法研究", 《仪器仪表学报》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107894713A (en) * 2017-10-20 2018-04-10 东南大学 A kind of high-accuracy control method without coding two axle inertially stabilized platforms of sensing
CN107894713B (en) * 2017-10-20 2020-11-06 东南大学 High-precision control method for two-axis inertial stabilization platform without coding sensing
CN108469269A (en) * 2018-01-19 2018-08-31 天津大学 A kind of resonance point test system of wideband inertial reference stabilized platform

Also Published As

Publication number Publication date
CN106200383B (en) 2019-10-18

Similar Documents

Publication Publication Date Title
US20180292842A1 (en) Aircraft attitude control methods
CN105353615B (en) A kind of Active Fault-tolerant Control Method of the quadrotor based on sliding mode observer
CN103365296B (en) A kind of four rotor unmanned aircraft nonlinear object feedback flight control methods
Xu et al. Neural network based dynamic surface control of hypersonic flight dynamics using small-gain theorem
CN105138010B (en) A kind of distributed finite time tracking controller design method of Satellite Formation Flying
Friedland Control system design: an introduction to state-space methods
CN105911866B (en) The finite time full-order sliding mode control method of quadrotor unmanned vehicle
CN101937233B (en) Nonlinear self-adaption control method of near-space hypersonic vehicle
Hamel et al. Evolution of flight vehicle system identification
Kurnaz et al. Adaptive neuro-fuzzy inference system based autonomous flight control of unmanned air vehicles
CN101709975B (en) Estimation and compensation method for unbalanced moment of aerial remote sensing inertially stabilized platform
CN102073755B (en) Motion control simulation method for near-space hypersonic aircraft
CN102955477B (en) Attitude control system and control method of four-rotor aircraft
CN104165642B (en) Method for directly correcting and compensating course angle of navigation system
Stramigioli et al. A novel approach to haptic tele-operation of aerial robot vehicles
Zhang et al. NNFFC-adaptive output feedback trajectory tracking control for a surface ship at high speed
CN105353763B (en) A kind of noncooperative target spacecraft relative orbit posture finite-time control method
CN106985139B (en) Space robot active-disturbance-rejection coordination control method based on extended state observation and compensation
Liu et al. Tracking control of small-scale helicopters using explicit nonlinear MPC augmented with disturbance observers
CN102636995B (en) Method for controlling micro gyro based on radial basis function (RBF) neural network sliding mode
CN102591358B (en) Multi-UAV (unmanned aerial vehicle) dynamic formation control method
Shao et al. High-order ESO based output feedback dynamic surface control for quadrotors under position constraints and uncertainties
CN104898429A (en) Tri-rotor attitude control method based on auto-disturbance rejection controller
CN105329777B (en) Fuzzy control method for lifting bridge crane system with persistent disturbances
CN104022742A (en) Aircraft posture robust inversion fault tolerant control method based on neural network observer

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
GR01 Patent grant