Scribd
Upload
47 views3 pages

Laboratory 10: Identification by The Least-Squares Method: Problem 1

The document describes three problems involving identifying unknown parameters in linear time-invariant (LTI) discrete-time systems using input-output data and the least-squares method. Problem 1 involves identifying coefficients in an autoregressive with exogenous inputs (ARX) model using given input-output sequences. Problem 2 involves identifying an unknown coefficient a in a given system equation using initial conditions and observed outputs for different inputs. Problem 3 similarly involves identifying unknown coefficients a and b in a given system equation using initial conditions, inputs, and observed outputs.

Uploaded by

Franco Claudio Antonio Porras Yarasca
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
47 views3 pages

Laboratory 10: Identification by The Least-Squares Method: Problem 1

The document describes three problems involving identifying unknown parameters in linear time-invariant (LTI) discrete-time systems using input-output data and the least-squares method. Problem 1 involves identifying coefficients in an autoregressive with exogenous inputs (ARX) model using given input-output sequences. Problem 2 involves identifying an unknown coefficient a in a given system equation using initial conditions and observed outputs for different inputs. Problem 3 similarly involves identifying unknown coefficients a and b in a given system equation using initial conditions, inputs, and observed outputs.

Uploaded by

Franco Claudio Antonio Porras Yarasca
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Laboratory 10:

Identification by the Least-Squares Method

Problem 1.
Unknown LTI system generates the following input-output sequences

u = [ 1.5; 0.8; 1; 1.2], y = [ 1; 1.5826; 0.92661; 0.5634]

(the first value of output y is the initial condition y(0) = 1).


Assuming that these data correspond to system of order n = 1 with noise, identify the optimal
estimates ✓ˆ1 , ✓ˆ2 for the coefficients of the corresponding ARX model

y(t) = ✓1 y(k 1) + ✓2 u(k 1) + d(k).

Problem 2.
A dynamical system is described by the equation

y(k + 2) = ay(k) + u(k + 1) + d(k),

where the coefficient a is unknown and d(k) is a unknown noise.


The system starts with the initial condition y(0) = 0, y(1) = 1. Inputs applied to the system
are: u(1) = 1, u(2) = 1, u(3) = 0. Observed outputs are y(2) = 3.77, y(3) = 1.35, y(4) = 6.8.

Problem 3.
Linear discrete-time system has the dynamics

y(k + 2) = au(k + 1) + bu(k) + d(k).

Here u(k) is the control input, d(k) is the random noise, and the coefficients a, b are unknown.
The initial condition of the system is y(0) = 0, y(1) = 0. Applying sequence of inputs u(0) =
1, u(1) = 1, u(2) = 2, u(3) = 3 we observe the outputs y(2) = 1.862, y(3) = 4.681, y(4) =
0.692.
Find the optimal estimates of the parameters â, b̂.

1
Procedure

① How prediction. 4 have


many you
② Sum of square
m s =

E, (ycH
-

(N)1*
S
& -> J= ly-dol1 (y-dElTy -00)
=

↑noun
*
o =

argmin 5- vo = 0
5py
Pinv(d)* Y
P0 y
-
o =
=

dyc/(
-o =

a dy
=

ston

-
3 t]
=

-[Ecop upinI
5 =(y(x) - y(H ly-101
=

ow
*
y(k) ay(k -2) u(k 1)
=
-

(g(t) 2)11y-pa)
1(Y(X)
U(t-1)- ay(z
g(t))
-
=

-
=

5
-

wa
=

ime

I ⑯

y =

[ce)-aux)]
re
g(3) -

~
-

y(x) au(k 1)
=
-

+ bu(z z) -

y -
g(t) y(t) au( 1) bu(kz)
=

- -
-

y
[ges]p-tailnaee]
=

You might also like