• Random telegraph process: A random telegraph process Y (t), t ≥ 0, assumes

values of +1 and −1 with Y (0) = +1 with probability 1/2 and −1 with

probability 1/2, and
Y (t) changes polarities with each event of a Poisson process with rate λ > 0
Sample path:
y(t)

+1

t
0 x1 x2 x3 x4 x5
−1

EE 278B: Random Processes 6 – 23

Mean and Autocorrelation Functions

• For a random vector X the first and second order moments are
◦ mean µ = E(X)
◦ correlation matrix RX = E(XXT )
• For a random process X(t) the first and second order moments are
◦ mean function: µX (t) = E(X(t)) for t ∈ T
! "
◦ autocorrelation function: RX (t1, t2) = E X(t1)X(t2) for t1, t2 ∈ T
• The autocovariance function of a random process is defined as
#! "! "$
CX (t1, t2) = E X(t1) − E(X(t1)) X(t2) − E(X(t2))
The autocovariance function can be expressed using the mean and
autocorrelation functions as
CX (t1, t2) = RX (t1, t2) − µX (t1)µX (t2)

EE 278B: Random Processes 6 – 24

 Examples

• IID process:
µX (n) = E(X1) %
E(X12) n1 = n2
RX (n1, n2) = E(Xn1 Xn2 ) =
(E(X1))2 n1 ̸= n2
• Random phase signal process:
& 2π
α
µX (t) = E(α cos(ωt + Θ)) = cos(ωt + θ) dθ = 0
0 2π
! "
RX (t1, t2) = E X(t1)X(t2)
& 2π 2
α
= cos(ωt1 + θ) cos(ωt2 + θ) dθ
0 2π
2π
α2 #
&
$
= cos(ω(t1 + t2) + 2θ) + cos(ω(t1 − t2)) dθ
0 4π
α2
= cos(ω(t1 − t2))
2

EE 278B: Random Processes 6 – 25

• Random walk:
n
'( ) n
(
µX (n) = E Zi = 0 = 0
i=1 i=1

RX (n1, n2) = E(Xn1 Xn2 )

= E [Xn1 (Xn2 − Xn1 + Xn1 )]
= E(Xn21 ) = n1 assuming n2 ≥ n1
= min{n1 , n2} in general

• Poisson process:
µN (t) = λt
RN (t1, t2) = E(N (t1)N (t2))
= E [N (t1)(N (t2) − N (t1) + N (t1))]
= λt1 × λ(t2 − t1) + λt1 + λ2 t21 = λt1 + λ2t1t2 assuming t2 ≥ t1
= λ min{t1, t2} + λ2t1t2

EE 278B: Random Processes 6 – 26

 Gaussian Random Processes

• A Gaussian random process (GRP) is a random process X(t) such that

[X(t1), X(t2), . . . , X(tn) ]T
is a GRV for all t1, t2, . . . , tn ∈ T
• Since the joint pdf for a GRV is specified by its mean and covariance matrix, a
GRP is specified by its mean µX (t) and autocorrelation RX (t1, t2) functions
• Example: The discrete time WGN process is a GRP

EE 278B: Random Processes 6 – 27

Gauss-Markov Process

• Let Zn , n ≥ 1, be a WGN process, i.e., an IID process with Z1 ∼ N (0, N )

The Gauss-Markov process is a first-order autoregressive process defined by
X1 = Z1
Xn = αXn−1 + Zn , n > 1,
where |α| < 1
• This process is a GRP, since X1 = Z1 and Xk = αXk−1 + Zk where Z1, Z2, . . .
are i.i.d. N (0, N ),
⎡ ⎤ ⎡ ⎤⎡ ⎤
X1 1 0 ··· 0 0 Z1
⎢ X2 ⎥ ⎢ α
⎢ ⎥ ⎢ . 1 · · · 0 0 ⎥ ⎢ Z2 ⎥
⎢ X3 ⎥ = ⎢ . .. . . . .. .. ⎥ ⎢ ⎥
⎥ ⎢ Z3 ⎥
⎢ . ⎥ ⎢ n−2
αn−3 · · · 1 0 ⎦ ⎣ .. ⎦
⎥⎢ ⎥
⎣ . ⎦ ⎣α
Xn αn−1 αn−2 · · · α 1 Zn
is a linear transformation of a GRV and is therefore a GRV
• Clearly, the Gauss-Markov process is Markov. It is not, however, an independent
increment process

EE 278B: Random Processes 6 – 28

• Mean and covariance functions:
µX (n) = E(Xn) = E(αXn−1 + Zn)
= α E(Xn−1) + E(Zn) = α E(Xn−1) = αn−1 E(Z1) = 0
To find the autocorrelation function, for n2 > n1 we write
n2 −n
(1−1
n2 −n1
X n2 = α X n1 + αiZn2−i
i=0

Thus
RX (n1, n2) = E(Xn1 Xn2 ) = αn2−n1 E(Xn21 ) + 0 ,
since Xn1 and Zn2−i are independent, zero mean for 0 ≤ i ≤ n2 − n1 − 1
Next, to find E(Xn21 ), consider
E(X12) = N
E(Xn21 ) = E (αXn1−1 + Zn1 )2 = α2 E(Xn21−1 ) + N
# $

Thus
1 − α2n1
E(Xn21 ) = N
1 − α2

EE 278B: Random Processes 6 – 29

Finally the autocorrelation function is

1 − α2 min{n1,n2}
RX (n1, n2) = α|n2−n1| N
1 − α2
• Estimation of Gauss-Markov process: Suppose we observe a noisy version of the
Gauss-Markov process,
Yn = X n + W n ,
where Wn is a WGN process independent of Zn with average power Q
We can use the Kalman filter from Lecture Notes 4 to estimate Xi+1 from Y i
as follows:
Initialization:
X̂1|0 = 0
2
σ1|0 =N

EE 278B: Random Processes 6 – 30

 Update: For i = 2, 3, . . . ,
X̂i+1|i = αX̂i|i−1 + ki(Yi − X̂i|i−1),
2 2
σi+1|i = α(α − ki)σi|i−1 + N, where
2
ασi|i−1
ki = 2
σi|i−1 +Q

Substituting from the ki equation into the MSE update equation, we obtain

2
α2Qσi|i−1
2
σi+1|i = 2 + N,
σi|i−1 +Q
This is a Riccati recursion (a quadratic recursion in the MSE) and has a steady
state solution:
α2Qσ 2
σ2 = 2 +N
σ +Q
Solving this quadratic equation, we obtain
0
2 N − (1 − α2)Q + 4N Q + (N − (1 − α2)Q)2
σ =
2

EE 278B: Random Processes 6 – 31

The Kalman gain ki converges to

0
−N − (1 − α2)Q + 4N Q + (N − (1 − α2)Q)2
k=
2αQ
and the steady-state Kalman filter is
X̂i+1|i = αX̂i|i−1 + k(Yi − X̂i|i−1)

EE 278B: Random Processes 6 – 32