1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
|
rpem Scilab Group Scilab Function rpem
NAME
rpem - RPEM estimation
CALLING SEQUENCE
[w1,[v]]=rpem(w0,u0,y0,[lambda,[k,[c]]])
PARAMETERS
a,b,c : a=[a(1),...,a(n)], b=[b(1),...,b(n)], c=[c(1),...,c(n)]
w0 : list(theta,p,phi,psi,l) where:
theta
: [a,b,c] is a real vector of order 3*n
p : (3*n x 3*n) real matrix.
phi,psi,l
: real vector of dimension 3*n
During the first call on can take:
theta=phi=psi=l=0*ones(1,3*n). p=eye(3*n,3*n)
u0 : real vector of inputs (arbitrary size) (if no input
take u0=[ ]).
y0 : vector of outputs (same dimension as u0 if u0 is not
empty). (y0(1) is not used by rpem).
If the time domain is (t0,t0+k-1) the u0 vector contains
the inputs
u(t0),u(t0+1),..,u(t0+k-1) and y0 the outputs
y(t0),y(t0+1),..,y(t0+k-1)
DESCRIPTION
Recursive estimation of parameters in an ARMAX model. Uses
Ljung-Soderstrom recursive prediction error method. Model considered is
the following:
y(t)+a(1)*y(t-1)+...+a(n)*y(t-n)=
b(1)*u(t-1)+...+b(n)*u(t-n)+e(t)+c(1)*e(t-1)+...+c(n)*e(t-n)
The effect of this command is to update the estimation of unknown
parameter theta=[a,b,c] with
a=[a(1),...,a(n)], b=[b(1),...,b(n)], c=[c(1),...,c(n)].
OPTIONAL PARAMETERS
lambda : optional parameter (forgetting constant) choosed close to 1
as convergence occur:
lambda=[lambda0,alfa,beta] evolves according to :
lambda(t)=alfa*lambda(t-1)+beta
with lambda(0)=lambda0
k : contraction factor to be chosen close to 1 as convergence
occurs.
k=[k0,mu,nu] evolves according to:
k(t)=mu*k(t-1)+nu
with k(0)=k0.
c : large parameter.(c=1000 is the default value).
OUTPUT PARAMETERS:
w1: update for w0.
v: sum of squared prediction errors on u0, y0.(optional).
In particular w1(1) is the new estimate of theta. If a new sample u1, y1
is available the update is obtained by:
[w2,[v]]=rpem(w1,u1,y1,[lambda,[k,[c]]]). Arbitrary large series can thus
be treated.
|
