9 views

Original Title: lec16 Bocipherous

Uploaded by denmwa2030ce

- v67i01
- Control Systems - Wikibooks, Open Books for an Open World
- Sylllabus
- Doumiati11_OnboardRealTimeEstimationofVehicleLateralTireRoad.pdf
- sliding mode
- Anticolission System
- Mathematical Modeling of Dynamic Systems
- chapter 2
- 3 DOF Gyroscope Courseware Sample for MATLAB Users
- The Scalar Kalman Filter
- A Structural Equation Multidimensional Scaling Model for One-Mode Asymmetric Dissimilarity Data
- Modelling and Predictive Control of a New Injection System Forcompressed Natural Gas Engines
- 1-s2.0-S1474667015365460-main.pdf
- CDS-II (4)
- Fundamentals of Kalman Filtering
- 2_A Robust Real-Time Embedded Vision System on an Unmanned Rotorcraft for Ground Target Following
- IntroductoryCourse Trigonometry
- Editorial 2
- 04389045.pdf
- 4571 Ast With Figs

You are on page 1of 8

A state space model consists of two equations:

S

t+1

= FS

t

+Ge

t+1

, (1)

Z

t

= HS

t

+

t

(2)

where S

t

is a state vector of dimension m, Z

t

is the observed time series, F, G, H are

matrices of parameters, {e

t

} and {

t

} are iid random vectors satisfying

E(e

t

) = 0, E(

t

) = 0, Cov(e

t

) = Q, Cov(

t

) = R

and {e

t

} and {

t

} are independent. In the engineering literature, a state vector denotes

the unobservable vector that describes the status of the system. Thus, a state vector

can be thought of as a vector that contains the necessary information to predict the future

observations (i.e., minimum mean squared error forecasts).

Remark: In some applications, a state-space model is written as

S

t+1

= FS

t

+Ge

t

,

Z

t

= HS

t

+

t

.

Is there any dierence between the two parameterizations?

In this course, Z

t

is a scalar and F, G, H are constants. A general state space model in

fact allows for vector time series and time-varying parameters. Also, the independence

requirement between e

t

and

t

can be relaxed so that e

t+1

and

t

are correlated.

To appreciate the above state space model, we rst consider its relation with ARMA models.

The basic relations are

an ARMA model can be put into a state space form in innite many ways;

for a given state space model in (1)-(2), there is an ARMA model.

A. State space model to ARMA model:

The key here is the Cayley-Hamilton theorem, which says that for any m m matrix F

with characteristic equation

c() = |F I| =

m

+

1

m1

+

2

m2

+ +

m1

+

0

,

we have c(F) = 0. In other words, the matrix F satises its own characteristic equation,

i.e.

F

m

+

1

F

m1

+

2

F

m2

+ +

m1

F +

m

I = 0.

1

Next, from the state transition equation, we have

S

t

= S

t

S

t+1

= FS

t

+Ge

t+1

S

t+2

= F

2

S

t

+FGe

t+1

+Ge

t+2

S

t+3

= F

3

S

t

+F

2

Ge

t+1

+FGe

t+2

+Ge

t+3

.

.

. =

.

.

.

S

t+m

= F

m

S

t

+F

m1

Ge

t+1

+ +FGe

t+m1

+Ge

t+m

.

Multiplying the above equations by

m

,

m1

, ,

1

, 1, respectively, and summing, we

obtain

S

t+m

+

1

S

t+m1

+ +

m1

S

t+1

+

m

S

t

= Ge

t+m

+

1

e

t+m1

+ +

m1

e

t+1

. (3)

In the above, we have used the fact that c(F) = 0.

Finally, two cases are possible. First, assume that there is no observational noise, i.e.

t

= 0

for all t in (2). Then, by multiplying H from the left to equation (3) and using Z

t

= HS

t

,

we have

Z

t+m

+

1

Z

t+m1

+ +

m1

Z

t+1

+

m

Z

t

= a

t+m

1

a

t+m1

m1

a

t+1

,

where a

t

= HGe

t

. This is an ARMA(m, m 1) model.

The second possibility is that there is an observational noise. Then, the same argument

gives

(1 +

1

B + +

m

B

m

)(Z

t+m

t+m

) = (1

1

B

m1

B

m1

)a

t+m

.

By combining

t

with a

t

, the above equation is an ARMA(m, m) model.

B. ARMA model to state space model:

We begin the discussion with some simple examples. Three general approaches will be

given later.

Example 1: Consider the AR(2) model

Z

t

=

1

Z

t1

+

2

Z

t2

+a

t

.

For such an AR(2) process, to compute the forecasts, we need Z

t1

and Z

t2

. Therefore,

it is easily seen that

Z

t+1

Z

t

1

2

1 0

Z

t

Z

t1

1

0

a

t+1

and

Z

t

= [1, 0]S

t

2

where S

t

= (Z

t

, Z

t1

)

Example 2: Consider the MA(2) model

Z

t

= a

t

1

a

t1

2

a

t2

.

Method 1:

a

t

a

t1

0 0

1 0

a

t1

a

t2

1

0

a

t

Z

t

= [

1

,

2

]S

t

+a

t

.

Here the innovation a

t

shows up in both the state transition equation and the observation

equation. The state vector is of dimension 2.

Method 2: For an MA(2) model, we have

Z

t|t

= Z

t

Z

t+1|t

=

1

a

t

2

a

t1

Z

t+2|t

=

2

a

t

.

Let S

t

= (Z

t

,

1

a

t

2

a

t1

,

2

a

t

)

. Then,

S

t+1

=

0 1 0

0 0 1

0 0 0

S

t

+

a

t+1

and

Z

t

= [1, 0, 0]S

t

.

Here the state vector is of dimension 3, but there is no observational noise.

Exercise: Generalize the above result to an MA(q) model.

Next, we consider three general approaches.

Akaikes approach: For an ARMA(p, q) process, let m = max{p, q + 1},

i

= 0 for

i > p and

j

= 0 for j > q. Dene S

t

= (Z

t

, Z

t+1|t

, Z

t+2|t

, , Z

t+m1|t

)

where Z

t+|t

is the

conditional expectation of Z

t+

given

t

= {Z

t

, Z

t1

, }. By using the updating equation

of forecasts (recall what we discussed before)

Z

t+1

( 1) = Z

t

() +

1

a

t+1

,

it is easy to show that

S

t+1

= FS

t

+Ga

t+1

Z

t

= [1, 0, , 0]S

t

3

where

F =

0 1 0 0

0 0 1 0

.

.

.

.

.

.

m

m1

2

1

, G =

2

.

.

.

m1

.

The matrix F is call a companion matrix of the polynomial 1

1

B

m

B

m

.

Aokis Method: This is a two-step procedure. First, consider the MA(q) part. Letting

W

t

= a

t

1

a

t1

q

a

tq

, we have

a

t

a

t1

.

.

.

a

tq+1

0 0 0 0

1 0 0 0

.

.

.

0 0 1 0

a

t1

a

t2

.

.

.

a

tq

1

0

.

.

.

0

a

t

W

t

= [

1

,

2

, ,

q

]S

t

+a

t

.

In the next step, we use the usual AR(p) format for

Z

t

1

Z

t1

p

Z

tp

= W

t

.

Consequently, dene the state vector as

S

t

= (Z

t1

, Z

t2

, , Z

tp

, a

t1

, , a

tq

)

.

Then, we have

Z

t

Z

t1

.

.

.

Z

tp+1

a

t

a

t1

.

.

.

a

tq+1

1

2

p

1

1

q

1 0 0 0 0 0

.

.

.

.

.

.

0 1 0 0 0 0

0 0 0 0 0 0

0 0 0 1 0 0

.

.

. 0

0 0 0 0 1 0

Z

t1

Z

t2

.

.

.

Z

tp

a

t1

a

t2

.

.

.

a

tq

1

0

.

.

.

0

1

0

.

.

.

0

a

t

.

and

Z

t

= [

1

, ,

p

,

1

, ,

q

]S

t

+a

t

.

Third approach: The third method is used by some authors, e.g. Harvey and his asso-

ciates. Consider an ARMA(p, q) model

Z

t

=

p

i=1

i

Z

ti

+a

t

q

j=1

j

a

tj

.

4

Let m = max{p, q}. Dene

i

= 0 for i > p and

j

= 0 for j > q. The model can be

written as

Z

t

=

m

i=1

i

Z

ti

+a

t

m

i=1

i

a

ti

.

Using (B) = (B)/(B), we can obtain the -weights of the model by equating coecients

of B

j

in the equation

(1

1

B

m

B

m

) = (1

1

B

m

B

m

)(

0

+

1

B + +

m

B

m

+ ),

where

0

= 1. In particular, consider the coecient of B

m

, we have

m

=

m

0

m1

1

1

m1

+

m

.

Consequently,

m

=

m

i=1

mi

m

. (4)

Next, from the -weight representation

Z

t+mi

= a

t+mi

+

1

a

t+mi1

+

2

a

t+mi2

+ ,

we obtain

Z

t+mi|t

=

mi

a

t

+

mi+1

a

t1

+

mi+2

a

t2

+

Z

t+mi|t1

=

mi+1

a

t1

+

mi+2

a

t2

+ .

Consequently,

Z

t+mi|t

= Z

t+mi|t1

+

mi

a

t

, mi > 0. (5)

We are ready to setup a state space model. Dene S

t

= (Z

t|t1

, Z

t+1|t1

, , Z

t+m1|t1

)

.

Using Z

t

= Z

t|t1

+a

t

, the observational equation is

Z

t

= [1, 0, , 0]S

t

+a

t

.

The state-transition equation can be obtained by Equations (5) and (4). First, for the rst

m 1 elements of S

t+1

, Equation (5) applies. Second, for the last element of S

t+1

, the

model implies

Z

t+m|t

=

m

i=1

i

Z

t+mi|t

m

a

t

.

Using Equation (5), we have

Z

t+m|t

=

m

i=1

i

(Z

t+mi|t1

+

mi

a

t

)

m

a

t

=

m

i=1

i

Z

t+mi|t1

+ (

m

i=1

mi

m

)a

t

=

m

i=1

i

Z

t+mi|t1

+

m

a

t

,

5

where the last equality uses Equation (4). Consequently, the state-transition equation is

S

t+1

= FS

t

+Ga

t

where

F =

0 1 0 0

0 0 1 0

.

.

.

.

.

.

m

m1

2

1

, G =

3

.

.

.

.

Note that for this third state space model, the dimension of the state vector is m =

max{p, q}, which may be lower than that of the Akaikes approach. However, the innova-

tions to both the state-transition and observational equations are a

t

.

Kalman Filter

Kalman lter is a set of recursive equation that allows us to update the information in

a state space model. It basically decomposes an observation into conditional mean and

predictive residual sequentially. Thus, it has wide applications in statistical analysis.

The simplest way to derive the Kalman recursion is to use normality assumption. It should

be pointed out, however, that the recursion is a result of the least squares principle (or

projection) not normality. Thus, the recursion continues to hold for non-normal case.

The only dierence is that the solution obtained is only optimal within the class of linear

solutions. With normailty, the solution is optimal among all possible solutions (linear and

nonlinear).

Under normality, we have

that normal prior plus normal likelihood results in a normal posterior,

that if the random vector (X, Y ) are jointly normal

X

Y

N(

xx

xy

yx

yy

),

then the conditional distribution of X given Y = y is normal

X|Y = y N[

x

+

xy

1

yy

(y

y

),

xx

xy

1

yy

yx

].

Using these two results, we are ready to derive the Kalman lter. In what follows, let P

t+j|t

be the conditional covariance matrix of S

t+j

given {Z

t

, Z

t1

, } for j 0 and S

t+j|t

be

the conditional mean of S

t+j

given {Z

t

, Z

t1

, }.

First, by the state space model, we have

S

t+1|t

= FS

t|t

(6)

6

Z

t+1|t

= HS

t+1|t

(7)

P

t+1|t

= FP

t|t

F

+GQG

(8)

V

t+1|t

= HP

t+1|t

H

+R (9)

C

t+1|t

= HP

t+1|t

(10)

where V

t+1|t

is the conditional variance of Z

t+1

given {Z

t

, Z

t1

, } and C

t+1|t

denotes

the conditional covariance between Z

t+1

and S

t+1

. Next, consider the joint conditional

distribution between S

t+1

and Z

t+1

. The above results give

S

t+1

Z

t+1

t

N(

S

t+1|t

Z

t+1|t

P

t+1|t

P

t+1|t

H

HP

t+1|t

HP

t+1|t

H

+R

).

Finally, when Z

t+1

becomes available, we may use the property of nromality to update the

distribution of S

t+1

. More specically,

S

t+1|t+1

= S

t+1|t

+P

t+1|t

H

[HP

t+1|t

H

+R]

1

(Z

t+1

Z

t+1|t

) (11)

and

P

t+1|t+1

= P

t+1|t

P

t+1|t

H

[HP

t+1|t

H

+R]

1

HP

t+1|t

. (12)

Obviously,

r

t+1|t

= Z

t+1

Z

t+1|t

= Z

t+1

HS

t+1|t

is the predictive residual for time point t + 1. The updating equation in (11) says that

when the predictive residual r

t+1|t

is non-zero there is new information about the system so

that the state vector should be modied. The contribution of r

t+1|t

to the state vector, of

course, needs to be weighted by the variance of r

t+1|t

and the conditional covariance matrix

of S

t+1

.

In summary, the Kalman lter consists of the following equations:

Prediction: (6), (7), (8) and (9)

Updating: (11) and (12).

In practice, one starts with initial prior information S

0|0

and P

0|0

, then predicts Z

1|0

and

V

1|0

. Once the observation Z

1

is available, uses the updating equations to compute S

1|1

and

P

1|1

, which in turns serve as prior for the next observation. This is the Kalman recusion.

Applications of Kalman Filter

Kalman lter has many applications. They are often classied as ltering, prediction,

and smoothing. Let F

t1

be the information available at time t 1, i.e., F

t1

= -

led{Z

t1

, Z

t2

, . . .}.

Filtering: make inference on S

t

given F

t

.

Prediction: draw inference about S

t+h

with h > 0, given F

t

.

7

SMoothing: make inference about S

t

given the data F

T

, where T t is the sample

size.

We shall briey discuss some of the applications. A good reference is Chapter 11 of Tsay

(2005, 2nd ed.).

8

- v67i01Uploaded byjxieeducation
- Control Systems - Wikibooks, Open Books for an Open WorldUploaded byfwef
- SylllabusUploaded byThennarasu Ramachandran
- Doumiati11_OnboardRealTimeEstimationofVehicleLateralTireRoad.pdfUploaded byDanielaSalas23
- sliding modeUploaded byedvin_s667
- Anticolission SystemUploaded bychezz43
- Mathematical Modeling of Dynamic SystemsUploaded byAhmed Said
- chapter 2Uploaded byJose C Beraun Tapia
- 3 DOF Gyroscope Courseware Sample for MATLAB UsersUploaded byDragan Lazic
- The Scalar Kalman FilterUploaded byapi-3731916
- A Structural Equation Multidimensional Scaling Model for One-Mode Asymmetric Dissimilarity DataUploaded byChristian
- Modelling and Predictive Control of a New Injection System Forcompressed Natural Gas EnginesUploaded by刘文
- 1-s2.0-S1474667015365460-main.pdfUploaded byVignesh Ramakrishnan
- CDS-II (4)Uploaded byTalha Yasin
- Fundamentals of Kalman FilteringUploaded byknez_007
- 2_A Robust Real-Time Embedded Vision System on an Unmanned Rotorcraft for Ground Target FollowingUploaded byMeghraj Patil
- IntroductoryCourse TrigonometryUploaded byAzhar Qadri
- Editorial 2Uploaded byEnes Şavlı
- 04389045.pdfUploaded byImam Fauzi
- 4571 Ast With FigsUploaded byRodrigo Astroza
- Signals Analysis.pdfUploaded byaman raj gupta
- Masters Thesis (in preparation)Uploaded byadam5179
- Ch05_p97-133Uploaded bysonalz
- 5 (1)Uploaded byYahya Ben Walid
- Geostatistics Formula SheetUploaded byAnonymous cMBxGA
- Ch22HullOFOD8thEditionUploaded byJame
- Dynamic Bayesian Networks Representation, Inference and Learning - Kevin Patrick MurphyUploaded byarounek
- BALLUploaded byIng Alan Juarez
- a243895.pdfUploaded byAnonymous WkbmWCa8M
- Xsg Fpga DspUploaded byelamaran_vlsi

- KSDQMJ9KVB3Q9LB446339NCFLP4RNL.rtfUploaded bydenmwa2030ce
- OutUploaded bydenmwa2030ce
- outpt (1)Uploaded bydenmwa2030ce
- ғʋиUploaded bydenmwa2030ce
- ғʋиUploaded bydenmwa2030ce
- Bop PrimerUploaded bydenmwa2030ce
- On the PortaUploaded bydenmwa2030ce
- ProtofuUploaded bydenmwa2030ce
- rm AddsUploaded bydenmwa2030ce
- kit3 reliable saveUploaded bydenmwa2030ce
- PutsetUploaded bydenmwa2030ce
- OcmologyUploaded bydenmwa2030ce
- dump1Uploaded bydenmwa2030ce
- German WordsUploaded bydenmwa2030ce
- fred dryst sdhtUploaded bydenmwa2030ce
- Sample SenseUploaded bydenmwa2030ce
- The Med of Fitin - Wilm GuestUploaded bydenmwa2030ce
- PQ IMDb ExampleUploaded bydenmwa2030ce
- Sally LoquaciousUploaded bydenmwa2030ce
- Dwelling Ben Joe'sUploaded bydenmwa2030ce
- RogtenUploaded bydenmwa2030ce
- Pat StatUploaded bydenmwa2030ce
- RevieUploaded bydenmwa2030ce
- RevolvingsUploaded bydenmwa2030ce
- LicenseUploaded byanon-731958
- Octave tutorialUploaded byPablo Peñas
- Widen the Roof's ReadingUploaded bydenmwa2030ce
- Widen the Roof's ReadingUploaded bydenmwa2030ce
- SequinsUploaded bydenmwa2030ce

- Loss of E-Cadherin Promotes Metastasis via Multiple Downstream Transcriptional PathwaysUploaded bypanna1
- Gas Turbine(3)Uploaded bymuru009
- The Speaking ProcessUploaded byJames William
- HP2-896Uploaded byJulces Gutierrez Quintero
- cvcxUploaded byMuhammad Wasif
- memtech2013.pdfUploaded byAnonymous 7wk2G5BX
- Workplace Safety and Health (Confined Spaces) Regulations 2009Uploaded bypeterlimttk
- diffusion in solids 3Rd phase transport.pptxUploaded bygere
- Clin. Microbiol. Rev.-2008-Jacobsen-26-59 (1)Uploaded byJose Humberto
- ReferenceUploaded byKaness Mathz
- Sar Sep 05 Bga30Uploaded bybalkanmonitor
- 1450783 - Christian Demonology IVUploaded byMarioSamaniego
- Via ARTiGO A1250 Datasheet v160809Uploaded byJunior Cabada
- Sluzebnic Chrysostom EsUploaded byPavel Patricolo
- Students' Ideas of ScienceUploaded byDarius P. Tolentino
- TM_MHW_2010_50-60Hz (UMITA)Uploaded byRodrigo Neira De Fino
- 95 Scs 4 Sr. No. 9Uploaded byBhupender Ramchandani
- CartomancyUploaded bylucentum56
- article-miasms.pdfUploaded bymarkos
- IRJET-Waste Management System using AWSUploaded byIRJET Journal
- SIMATIC PCS7Uploaded by'-MubashirBaig
- Protection of Engine Enclosures Using Aerosol Generators-ForUploaded byuserscribd2011
- pone.0022577Uploaded bySiva Karthik
- Servomechanism DefectsUploaded byPandurang Nalawade
- Pocket Hole JoineryUploaded byJessica Sanders
- Design Data Sheet 30Uploaded bycool_rd
- Aerodynamic Drag of Heavy Vehicles (Class 7-8) Simulation and BenchmarkingUploaded byMarcusHuynh88
- Chem 101.1 Back Titration Formal ReportUploaded bytoopiar
- Annual_Report-MPL(12-13).pdfUploaded bySUKHSAGAR1969
- Sump Pump P-534 Service ManualUploaded bykaash028