1
<?xml version="1.0" encoding="ISO-8859-1" standalone="no"?>
2
<!DOCTYPE MAN SYSTEM "../../manrev.dtd">
4
<LANGUAGE>eng</LANGUAGE>
6
<TYPE>Scilab Function</TYPE>
7
<DATE>April 1993</DATE>
8
<SHORT_DESCRIPTION name="armax1"> armax identification</SHORT_DESCRIPTION>
10
<CALLING_SEQUENCE_ITEM>[arc,resid]=armax1(r,s,q,y,u [,b0f]) </CALLING_SEQUENCE_ITEM>
15
<PARAM_NAME>y</PARAM_NAME>
17
<SP>: output signal</SP>
21
<PARAM_NAME>u</PARAM_NAME>
23
<SP>: input signal</SP>
27
<PARAM_NAME>r,s,q</PARAM_NAME>
29
<SP>: auto regression orders with r >=0, s >=-1.</SP>
33
<PARAM_NAME>b0f</PARAM_NAME>
35
<SP>: optional parameter. Its default value is 0 and it means that the coefficient b0 must be identified. if bof=1 the b0 is supposed to be zero and is not identified</SP>
39
<PARAM_NAME>arc</PARAM_NAME>
41
<SP>: is tlist with type "ar" and fields a, b, d, ny, nu, sig</SP>
44
<PARAM_NAME>a</PARAM_NAME>
46
<SP>: is the vector [1,a1,...,a_r]</SP>
50
<PARAM_NAME>b</PARAM_NAME>
52
<SP>: is the vector [b0,......,b_s]</SP>
56
<PARAM_NAME>d</PARAM_NAME>
58
<SP>: is the vector [1,d1,....,d_q]</SP>
62
<PARAM_NAME>sig</PARAM_NAME>
64
<SP>: resid=[ sig*echap(1),....,];</SP>
74
armax1 is used to identify the coefficients of a 1-dimensional
78
A(z^-1)y= B(z^-1)u + D(z^-1)sig*e(t)
79
e(t) is a 1-dimensional white noise with variance 1.
80
A(z)= 1+a1*z+...+a_r*z^r; ( r=0 => A(z)=1)
81
B(z)= b0+b1*z+...+b_s z^s ( s=-1 => B(z)=0)
82
D(z)= 1+d1*z+...+d_q*z^q ( q=0 => D(z)=1)
86
for the method, see Eykhoff in trends and progress in system identification) page 96.
88
z(t)=[y(t-1),..,y(t-r),u(t),...,
89
u(t-s),e(t-1),...,e(t-q)]
91
coef= [-a1,..,-ar,b0,...,b_s,d1,...,d_q]'
92
y(t)= coef'* z(t) + sig*e(t).</P>
94
a sequential version of the AR estimation where e(t-i) is replaced
95
by an estimated value is used (RLLS). With q=0 this method is exactly
96
a sequential version of armax</P>
98
<AUTHOR>J.-Ph.C; ; </AUTHOR>