This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 8 Bivariate Processes 8.1 Timedomain Quantities Suppose we have N observations recorded on two variables, ( x 1 , y 1 ) , . . . , ( x N , y N ) . This can be thought of as a finite realization of a discrete bivariate stochastic process ( X t , Y t ). Definition 8.1.1: The crosscovariance function is defined as γ XY ( t, τ ) = cov( X t , Y t + τ ) . (8.1) For stationary processes we may abbreviate this to γ XY ( τ ) = γ XY ( t, τ ) . (8.2) Proposition 8.1: γ XY ( τ ) = γ Y X ( τ ) . Definition 8.1.2: We define the crosscorrelation function as ρ XY ( τ ) = γ XY ( τ ) radicalbig γ XX (0) γ Y Y (0) (8.3) Suppose that { X t } and { Y t } are both formed from the same purely random process { Z t } with mean 0 and variance σ 2 Z by X t = Z t , Y t = 0 . 5 · ( Z t 1 + Z t 2 ) . Then after some calculations ρ XY ( τ ) = braceleftBigg 1 √ 2 , τ = 1 , 2 , otherwise 8.2 Estimation Basically uses obvious quantities similar to quantities in Section 2.7 which are sensibleBasically uses obvious quantities similar to quantities in Section 2....
View
Full Document
 Spring '09
 jsdkasj
 Signal Processing, Probability theory, Stochastic process, Autocorrelation, Stationary process, bivariate stochastic process

Click to edit the document details