Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации
A neuro-controller synthesis is performed on the basis of an autoregressive moving average model to solve a control problem for a light-armored vehicle armament guidance and stabilization system. An algorithm of NARMA-L2 controller synthesis for a given control object is described. NARMA-L2 controll...
Gespeichert in:
| Veröffentlicht in: | Електротехніка і електромеханіка |
|---|---|
| Datum: | 2011 |
| Hauptverfasser: | , , |
| Format: | Artikel |
| Sprache: | Russian |
| Veröffentlicht: |
Інститут технічних проблем магнетизму НАН України
2011
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/143554 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации / Б.И. Кузнецов, Т.Е. Василец, A.A Варфоломеев // Електротехніка і електромеханіка. — 2011. — № 4. — С. 41-46. — Бібліогр.: 7 назв. — рос. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-143554 |
|---|---|
| record_format |
dspace |
| spelling |
Кузнецов, Б.И. Василец, Т.Е. Варфоломеев, А.А. 2018-11-05T18:09:06Z 2018-11-05T18:09:06Z 2011 Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации / Б.И. Кузнецов, Т.Е. Василец, A.A Варфоломеев // Електротехніка і електромеханіка. — 2011. — № 4. — С. 41-46. — Бібліогр.: 7 назв. — рос. 2074-272X https://nasplib.isofts.kiev.ua/handle/123456789/143554 681.5.01.23 A neuro-controller synthesis is performed on the basis of an autoregressive moving average model to solve a control problem for a light-armored vehicle armament guidance and stabilization system. An algorithm of NARMA-L2 controller synthesis for a given control object is described. NARMA-L2 controller parameters that significantly affect the control quality are ascertained; the parameters values that provide the system’s preset performance quality ratings are specified. Computer simulation of the system is made. ru Інститут технічних проблем магнетизму НАН України Електротехніка і електромеханіка Електричні машини та апарати Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации NARMA-L2 controller synthesis for a guidance and stabilization system Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации |
| spellingShingle |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации Кузнецов, Б.И. Василец, Т.Е. Варфоломеев, А.А. Електричні машини та апарати |
| title_short |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации |
| title_full |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации |
| title_fullStr |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации |
| title_full_unstemmed |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации |
| title_sort |
синтез нейросетевого регулятора narma-l2 controller для системы наведения и стабилизации |
| author |
Кузнецов, Б.И. Василец, Т.Е. Варфоломеев, А.А. |
| author_facet |
Кузнецов, Б.И. Василец, Т.Е. Варфоломеев, А.А. |
| topic |
Електричні машини та апарати |
| topic_facet |
Електричні машини та апарати |
| publishDate |
2011 |
| language |
Russian |
| container_title |
Електротехніка і електромеханіка |
| publisher |
Інститут технічних проблем магнетизму НАН України |
| format |
Article |
| title_alt |
NARMA-L2 controller synthesis for a guidance and stabilization system |
| description |
A neuro-controller synthesis is performed on the basis of an autoregressive moving average model to solve a control problem for a light-armored vehicle armament guidance and stabilization system. An algorithm of NARMA-L2 controller synthesis for a given control object is described. NARMA-L2 controller parameters that significantly affect the control quality are ascertained; the parameters values that provide the system’s preset performance quality ratings are specified. Computer simulation of the system is made.
|
| issn |
2074-272X |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/143554 |
| citation_txt |
Синтез нейросетевого регулятора NARMA-L2 CONTROLLER для системы наведения и стабилизации / Б.И. Кузнецов, Т.Е. Василец, A.A Варфоломеев // Електротехніка і електромеханіка. — 2011. — № 4. — С. 41-46. — Бібліогр.: 7 назв. — рос. |
| work_keys_str_mv |
AT kuznecovbi sintezneirosetevogoregulâtoranarmal2controllerdlâsistemynavedeniâistabilizacii AT vasilecte sintezneirosetevogoregulâtoranarmal2controllerdlâsistemynavedeniâistabilizacii AT varfolomeevaa sintezneirosetevogoregulâtoranarmal2controllerdlâsistemynavedeniâistabilizacii AT kuznecovbi narmal2controllersynthesisforaguidanceandstabilizationsystem AT vasilecte narmal2controllersynthesisforaguidanceandstabilizationsystem AT varfolomeevaa narmal2controllersynthesisforaguidanceandstabilizationsystem |
| first_indexed |
2025-11-26T15:26:46Z |
| last_indexed |
2025-11-26T15:26:46Z |
| _version_ |
1850626437626724352 |
| fulltext |
ISSN 2074-272X. . 2011. 4 41
681.5.01.23
N RMA-L2 CONTROLLER
. , . , .
.
RMA-L2 Controller . , -
, -
. .
-
. -
N RMA-L2 Controller . , -
,
. .
. -
, -
, . -
-
, -
, .
-
-
, -
-
.
.
-
-
.
,
-
-
, -
, -
. -
,
.
" " -
1974 [1], 80-
-
. -
-
-
, [2].
90- -
-
. [3, 4].
-
–
-
.
,
, -
. . [5]. -
-
,
-
, -
.
-
,
.
.
N RMA-L2 Controller -
-
,
Neural Network
Toolbox MATLAB.MATLAB
: NN
Predictive Controller, -
N RMA-L2
Controller
Model Reference Controller. -
[6], -
-
NN Predictive Controller, -
.
,
. ,
,
. , -
. -
N RMA-L2 Controller.
42 ISSN 2074-272X. . 2011. 4
N RMA-L2 Controller,
-
.
N RMA-L2
Controller , -
toolbox/nnet/ nncontrol
SIMULINK: Nncontrolutil – ,
-
SIMULINK; Sfunxy2 – -
; Nnident.m – , -
(
-
,
Neural Network Toolbox MATLAB).
. 1 -
-
,
SIMULINK. -
Subsystem N RMA
– L2 Controller,
Random Reference, . -
, -
Kp, Kd ( -
-
),
Derivative Saturation.
[7].
f
1
s
In Out
Subsystem
Random
Reference1
Plant
Output
Reference
Control
Signal
f g
NARMA-L2 Control ler
Kd
Kp
du/dt
Derivative Saturation
. 1. N RMA-L2 Controller
: -
.
-
,
.
N RMA-L2 Controller. -
Plant Identification N RMA-L2, -
. 2.
,
SIMULINK. -
MATLAB
nnident.m.
. 2.
-
.
:
Size of Hidden Layer (N) –
;
Sampling Interval ( t) – -
-
;
No. Delayed Plant Inputs (Ni) – -
;
No. Delayed Plant Outputs (Nj) – -
;
Normalize Training Data. -
[0 1].
:
Training samples (NB) – -
( );
Maximum Plant Input ( max) - -
;
Minimum Plant Input ( min) – -
;
Maximum Interval Value (sec) (tmax) – -
;
Minimum Interval Value (sec) (tmin) – -
;
Limit Output Data. ,
,
2
:
Maximum Plant Output.
;
Minimum Plant Output.
;
Simulink Plant Model – Simulink
, -
ISSN 2074-272X. . 2011. 4 43
-
.
Browser -
,
, [7].
:
Training Epochs – ;
Training function – ;
Use Current Weights – , -
-
;
Use Validation/Testing Training –
, 25 %
-
.
Simulink gensim (netn)
. 3).
, -
, -
( . . 2): S = 8, -
Ni = 3
-
Nj = 4.
-
.
Simulink , . 4.
12
lz{3,2}1
11
iz{1,1}2
10
lz{6,5}
9
lz{5,4}
8
lz{4,3}
7
lz{2,1}
6
a{2}1
5
a{6}
4
a{4}1
3
a{3}1
2
a{1}1
1
y{1}1
y{1}
tansig1
tansig
Input 1
p{1}
purel in2
purel in1
purel in
netsum4
netsum3
netsum2
netsum1
netsum
w
p
z
dotprod43
w
p
z
dotprod42
w
p
z
dotprod41
w
p
z
dotprod40
w
p
z
dotprod39
w
p
z
dotprod38
w
p
z
dotprod37
w
p
z
dotprod36
w
p
z
dotprod35
w
p
z
dotprod34
w
p
z
dotprod33
w
p
z
dotprod32
w
p
z
dotprod31
w
p
z
dotprod30
w
p
z
dotprod29
w
p
z
dotprod28
w
p
z
dotprod18
w
p
z
dotprod17
w
p
z
dotprod16
w
p
z
dotprod15
bias
b{6}
bias
b{4}
bias
b{3}
bias
b{2}
bias
b{1}
a{5}
a{4}
a{3}
a{2}
a{1}
p{1} y {1}
Neural Network
Mux
Mux8
Mux
Mux7
Mux
Mux5
Mux
Mux4
Mux
Mux3
Mux
Mux2
a{5} a{6}
Layer 6
a{4} a{5}
Layer 5
a{3} a{4}
Layer 4
a{2} a{3}
Layer 3
a{1} a{2}
Layer 2
p{1} a{1}
Layer 1
weight
LW{6,5}
weight
LW{4,3}
weight
LW{3,2}
weight
LW{2,1}
weights
IW{6,5}(1,:)'
weights
IW{5,4}(1,:)'
weights
IW{4,3}(1,:)'
weights
IW{3,2}(8,:)'
weights
IW{3,2}(7,:)'1
weights
IW{3,2}(6,:)'1
weights
IW{3,2}(5,:)'1
weights
IW{3,2}(4,:)'1
weights
IW{3,2}(3,:)'1
weights
IW{3,2}(2,:)'1
weights
IW{3,2}(1,:)'1
weights
IW{2,1}(1,:)'
weights
IW{1,1}(8,:)'1
weights
IW{1,1}(7,:)'2
weights
IW{1,1}(6,:)'2
weights
IW{1,1}(5,:)'2
weights
IW{1,1}(4,:)'2
weights
IW{1,1}(3,:)'2
weights
IW{1,1}(2,:)'2
weights
IW{1,1}(1,:)'2
weight
IW{1,1}
TDL
Delays 5
TDL
Delays 4
TDL
Delays 3
TDL
Delays 2
TDL
Delays 1
a{5}
a{4}
a{3}
a{2}
a{1}
12
ad{3,2}1
11
pd{1,1}2
10
ad{6,5}
9
ad{5,4}
8
ad{4,3}
7
ad{2,1}
6
a{1}
5
a{5}
4
a{3}
3
a{2}
2
p{1}2
1
p{1}1
. 3. ,
gensim (netn)
. 4. netn N RMA-L2
, .
. 1
6 .
Nj
y(k), y(k 1),…, y(k Nj + 1) ( -
y(k), y(k 1), y(k 2), y(k 3), (Ni
1) u(k 1),…, u(k Ni + 1) (
u(k 1), u(k 2).
6 8 1
, , -
. : -
(tansig) – , -
(purelin) – , , -
.
netn -
:
netn.numInputs=2;
netn.numInputs=3;
netn.inputs{2}.size=netn.inputs{1}.size;
netn.inputs{2}.range=netn.inputs{1}.range;
netn.inputs{3}.range=minmax(ptr{3,1});
netn.biasConnect(5:6)=0;
netn.layers{5}.netInputFcn='netprod';
netn.inputConnect(3,2)=1;
netn.inputConnect(5,3)=1;
netn.layerConnect(6,2)=1;
netn.layerConnect(3,2)=0.
, -
. 5. 3
6 1
.
, , .
u(k). -
(purelin) -
netprod, -
-
.
y{1 }1
tan sig 1ta nsig
In put 1
p{1 }
pu rel in3
pu rel in2
p ure lin 1 pu rel in
n etsum 6
n etsum4
n etsum3net su m2
n etsum1ne tsu m
w
p
z
do tp rod 9
w
p
z
dot pro d8
w
p
z
dot pro d7
w
p
z
dot pro d6
w
p
z
dot pro d5
w
p
z
dot pro d4
w
p
z
dot pro d3
w
p
z
d otp rod 20
w
p
z
dot pro d2
w
p
z
do tprod1 9
w
p
z
do tpr od1 8
w
p
z
do tprod1 7
w
p
z
do tprod 16
w
p
z
do tprod1 5
w
p
z
do tprod1 4
w
p
z
do tprod1 3
w
p
z
do tprod1 2
w
p
z
do tprod1 1
w
p
z
do tprod1 0
w
p
z
dot pro d1
b ias
b {6}
b ias
b {5}
b ias
b {4}
b ias
b {3}
bia s
b{2 }bia s
b{1 }
M ux
Mux8
Mu x
Mu x6
M ux
Mu x5M ux
Mu x4M ux
Mu x3
M ux
Mux2
we igh ts
IW{6,5 }(1 ,:) '
we igh ts
IW{5, 4}(1 ,:)'we igh ts
IW{4, 3}(1 ,:)'
we igh ts
IW{ 3,2 }(8, :)'1
we igh ts
IW{ 3,2 }(7, :)'2
we igh ts
IW{ 3,2 }(6, :)'2
we igh ts
IW{ 3,2 }(5, :)'2
we igh ts
IW{ 3,2 }(4, :)'2
we igh ts
IW{ 3,2 }(3, :)'2
we igh ts
IW{ 3,2 }(2, :)'2
we igh ts
IW{ 3,2 }(1, :)'2
weig hts
IW{2 ,1}(1,: )'
weig hts
IW{1, 1}(8 ,:)'2
weig hts
IW{1, 1}(7 ,:)'1
weig hts
IW{1, 1}(6 ,:)'1
weig hts
IW{1, 1}(5 ,:)'1
weig hts
IW{1, 1}(4 ,:)'1
weig hts
IW{1, 1}(3 ,:)'1
weig hts
IW{1, 1}(2 ,:)'1
weig hts
IW{1, 1}(1 ,:)'1
44 ISSN 2074-272X. . 2011. 4
, -
.5, . 6.
y{1}
Input 3
p{3}
Input 2
p{2}
Input 1
p{1}
p{1}
p{2}
p{3}
y {1}
Neural Network
1
y{1}1
a{5}
a{4}
a{3}
a{2}
a{1}
a{2}
a{5}
a{6}
Layer 6
p{3}
a{4}
a{5}
Layer 5
a{3} a{4}
Layer 4
p{2} a{3}
Layer 3
a{1} a{2}
Layer 2
p{1} a{1}
Layer 1
a{5}
a{4}
a{3}
a{2}
a{1}
3
p{3}1
2
p{2}1
1
p{1}1
. 5. netn
. 6. netn N RMA-L2
. -
doubl , -
.
trainlm,
.
, -
-
. 7.
. 8.
, -
. 8.
IW{1,1}, IW{3,2},
IW{5,3}, LW{2,1}, LW{4,3}, LW{5,4}, LW{6,5},
LW{6,2} b{1}, b{2}, b{3}, b{4}
N RMA – L2 Controller Simulink.
Simulink
, .9. K
Matrix Gain : Matrix Gain
IW1_1=netn.IW{1,1};
Matrix Gain1IW3_2=netn.IW{3,2};
Matrix Gain2 LW2_1=netn.LW{2,1};
Matrix Gain3 LW2_1=netn.LW{4,3};
Matrix Gain5 LW4_3=netn.LW{6,2};
Matrix Gain8 LW6_5*LW5_4*IW5_3=netn.LW{6,5}*
*netn.LW{5,4}*netn.IW{5,3}.
. 7.
N RMA-L2 Controller
y{1}
tansi g1
tansig
Input 3
p{3}
Input 2
p{2}
Input 1
p{1}
purel in4
pureli n2
pureli n1
pureli n
net sum5
netsum 3
netsum 2
netsum1
netsum
netprod
w
p
z
dotprod8
w
p
z
dotprod43
w
p
z
dotprod42
w
p
z
dotprod41
w
p
z
dotprod40
w
p
z
dotprod39
w
p
z
dotprod38
w
p
z
dotprod37
w
p
z
dotprod36
w
p
z
dotprod35
w
p
z
dotprod34
w
p
z
dotprod33
w
p
z
dotprod32
w
p
z
dotprod31
w
p
z
dotprod30
w
p
z
dotprod29
w
p
z
dotprod28
w
p
z
dotprod20
w
p
z
dotprod19
w
p
z
dotprod18
w
p
z
dotprod17
w
p
z
dotprod16
bi as
b{4}bi as
b{3}
bi as
b{2}bi as
b{1}
Mux
M ux8
Mux
Mux7
Mux
M ux6
M ux
Mux5
M ux
Mux4
Mux
M ux3Mux
M ux2
M ux
Mux1
weights
IW{6, 5}(1,: )'
wei ghts
IW{6,2}(1,:) '
wei ghts
IW{5,4}(1, :)'
wei ghts
IW{5, 3}(1, :)'
wei ghts
IW{4,3}(1,: )'
wei ghts
IW{3,2}(8, :)'
wei ghts
I W{3, 2}(7,: )'1
wei ghts
I W{3, 2}(6,: )'1
wei ghts
I W{3, 2}(5,: )'1
wei ghts
I W{3, 2}(4,: )'1
wei ghts
I W{3, 2}(3,: )'1
wei ghts
I W{3, 2}(2,: )'1
wei ghts
I W{3, 2}(1,: )'1
weight s
IW{2,1}(1,:)'
wei ghts
IW{1, 1}(8,: )'1
wei ghts
IW{1, 1}(7,: )'2
wei ghts
IW{1, 1}(6,: )'2
wei ghts
IW{1, 1}(5,: )'2
wei ghts
IW{1, 1}(4,: )'2
wei ghts
IW{1, 1}(3,: )'2
wei ghts
IW{1, 1}(2,: )'2
wei ghts
IW{1, 1}(1,: )'2
ISSN 2074-272X. . 2011. 4 45
. 8.
RMA-L2 Controller
Constant value Constant
B1=netn.b{1}; B2=netn.b{2};
B3=netn.b{3}; B4=netn.b{4};
Discrete State Space ,
[6] NN Predictive Controller.
N RMA-L2 Controller,
NN Predictive Controller,
-
S.
S = 8-14
,
10 4-10 5.
1
Control Signal
tansig1
tansig
purelin1
purelin
+
netsum3
+
netsum2
+
netsum1
+
netsum
Zero-Order
Holdz
1
Unit Delay4
Switch3
Switch2
Switch1
Switch
Sum
Saturation1Product
signal1
signal2
K*u
Matrix
Gain8
K*u
Matrix
Gain5
K*u
Matrix
Gain3
K*u
Matrix
Gain2
K*u
Matrix
Gain1
K*u
Matrix
Gain
f(u)
Fcn3
f(u)
Fcn2
f(u)
Fcn1
f(u)
Fcn
y(n)=Cx(n)+Du(n)
x(n+1)=Ax(n)+Bu(n)
Discrete State-Space4
-C-
Constant7
B4
Constant6
B3
Constant5
-C-
Constant4
-C-
Constant3
-C-
Constant2
B2
Constant1
B1
Constant
2
Plant Output
1
Reference
. 9. N RMA-L2
NB -
t,
. -
: NB = 10000,
t = 0,001 c. t
,
. t
NB , ,
,
.
-
-
, . -
. -
. -
, – .
. :
tmin = 0,01 c, tmax = 0,1 c.
Ni
Nj
Ni = 1-4, Nj = 2-5.
N = 300,
, 300-600.
: S = 10, Ni = 1, Nj = 5; N = 300. ,
N RMA-L2 Controller -
. 7 8.
,
3,68 10 12, -
4 10 5. , ,
NN Predictive
Controller,
RMA-L2 Controller -
. . 10 -
-
.
,
, -
LW{6,5}, LW{5,4},
IW{5,3} 1-10 5,
LW{6,2} 1-10 1 ( ). -
, ,
Matrix Gain8 Matrix Gain5 ( . . 9). -
, Product
1. -
Control Signal +1 1
Saturation1 -
Maximum Plant Input = 1 Minimum Plant Input
= 1 ( . Plant Identification – NARMA-L2,
.2)). ,
0,
+27 27 ,
. -
0,006
12 .
46 ISSN 2074-272X. . 2011. 4
0 0.5 1 1.5 2
0
0.1
0.2
0.3
0.4
0.5
0.6
t, c
f, rad
1 2
)
1-
2-
0 0.5 1 1.5 2
-1
-0.5
0
0.5
1
t, c
Wm, c-1
) 0 0.5 1 1.5 2
-30
-20
-10
0
10
20
30
t, c
Ud, B
)
. 10. N RMA – L2 Controller:
) – ; ) – ; ) – U
RMA-L2 Controller -
-
-
. Simulink MATLAB -
-
N RMA-L2 Controller -
.
N RMA-L2 Controller, -
Neural Network
Toolbox MATLAB. -
, -
,
-
, -
.
.
1. Werbos P. J. Beyond regression: New tools for prediction
and analysis in the behavioral sciences. PhD Thesis, Harvard
University, Cambridge, MA. – 1974.
2. Narendra K.S., Parthasarathy K. Identification and control
of dynamical system using neural networks // IEEE Trans. Neu-
ral Networks. – 1990. – Vol.1. – 1. – P. 4-27.
3. ., ., C.A. -
// -
. – 1999. – 5. – . 2-6.
4. -
: ., .,
., ., . //
". – 2002. – 9. – . 47-52.
5. , . -
. – : , 2002. – 317 .
6. ., ., .
-
. // -
. – 2008. – 3. – . 27-32.
7. ., ., . -
//
. – 2008. – 2. – . 31-34.
Bibliography (transliterated): 1. Werbos P. J. Beyond regression: New
tools for prediction and analysis in the behavioral sciences. PhD Thesis,
Harvard University, Cambridge, MA. - 1974. 2. Narendra K.S., Par-
thasarathy K. Identification and control of dynamical system using neu-
ral networks // IEEE Trans. Neural Networks. - 1990. - Vol.1. - 1. - P.
4-27. 3. Klepikov V.B., Mahotilo K.V., Sergeev C.A. Primenenie meto-
dov nejronnyh setej i geneticheskih algoritmov v reshenii zadach uprav-
leniya `elektroprivodami// `Elektrotehnika. - 1999. - 5. - S. 2-6.
4. Nejro-fazzi regulyator dlya `elektroprivodov s proskal'zyvaniem:
Klepikov V.B., Klepikov A.V., Glebov O.Yu., Moiseenko P.L., Polyan-
skaya I.S. // V snik NTU "HP ". - 2002. - 9. - S. 47-52. 5. Rudenko
O.G, Bodyanskij E.V. Osnovy teorii iskusstvennyh nejronnyh setej. -
Har'kov: TELETEH, 2002. - 317 s. 6. Kuznecov B.I., Vasilec T.E., Var-
folomeev A.A. Sintez nejrokontrollera s predskazaniem dlya dvuhmass-
ovoj `elektromehanicheskoj sistemy. // Elektrotehn ka elektrome-
han ka. - 2008. - 3. - S. 27-32. 7. Kuznecov B.I., Vasilec T.E., Var-
folomeev A.A. Razrabotka nejrosetevoj sistemy navedeniya i stabilizacii
vooruzheniya legkobronirovannyh mashin // Elektrotehn ka
elektromehan ka. - 2008. - 2. - S. 31-34.
22.02.2011
, ., .,
, ., .,
61003, , . , 16
. (057) 733-79-59
, .
,
. . +197-39-54-34-42
Kuznetsov B.I., Vasilets T.E., Varfolomeev A.A.
NARMA-L2 controller synthesis for a guidance and
stabilization system.
A neuro-controller synthesis is performed on the basis of an
autoregressive moving average model to solve a control problem
for a light-armored vehicle armament guidance and stabilization
system. An algorithm of NARMA-L2 controller synthesis for a
given control object is described. NARMA-L2 controller pa-
rameters that significantly affect the control quality are ascer-
tained; the parameters values that provide the system’s preset
performance quality ratings are specified. Computer simulation
of the system is made.
Key words – neuro-controller, autoregressive moving average
model, neural guidance and stabilization, NARMA-L2
controller control system, synthesis.
|