ECE 250 Midterm October 28, 2016
SOLUTION
1. The discrete random variable X has the probabilities P(X = -2) = P(X = -1) = P(X = 1) = P(X = 2) = 1/4. The random variable Y is defined by Y = independent?
lxJ.
Are X andY uncorrelated? Are they
I Krkj-t-..,A,.;f;- tff- P(~.: n1y;: wr) =P~::. Yt) P( '(:= wr)
fOr ALL.. n) WI
An answer not supported by appropriate reasoning will not receive any credit.
-2 -\
I :J-..
Y= lZI XY- Prob. 2 -. 4'I+ I
-.1
I 2
l
E. fXiJ
t
1/4 1
/4
'/+
: : (-4)xP(X<"72)~(-')~
£LX) :: (- 2) P(X= -2.) +( -1) P(X=-1)+-( l)f[X=t) -f(2)P(K= 2 ) :.0
ELYj : \2) P< X= - 2)
=t
_·.
+(I) P(X=-1) 4-(1 )P(X;:I )+2 P(X= ;z)
ELXYJ = ~[2CJELY1
P( X= -;z., Y~ 2) "' ;f- j P(X.,-z)= ;( j P[ Y: 2) PC Z =- 2) P( '1:: 2) =- -ft;- =t PC X"= - 2 7 Y= z) 5vffcer..J~t- To ()&;It a. .. y ffJ•rdt vo. rc G\ \Jt.es X a"o{ Y
..
No
Uncorrelated? Independent? (Circle One)
=f
Yes
Page 2 of9
-· '
'
SOLUTION 2. The random variables X and Y have the joint characteristic function
1
1
2
2
xy(u, v) = -cos(u- v) + -cos(u + v).
'
Are X and Y independent?
An answer not supported by appropriate reasoning will not receive any credit.
CE x: (l<) ~_g(.v) =-
Co '7U. w~\1
= {f.I',.f (u1v)
= ±~ v.-v) + ~COS{?IN) ,..---.fy. o wr ''Us.t{vl" for..,Ja~ll
Independent? (Circle One)
No
Page 4 of 10
•
SOLUTION
3. The random variables X and Y are independent and identically distributed with probabilities 1 n 1 m P(X=n)=(-) and P(Y=m)= (-) , n,m= 1,2, ....
2
2
Evaluate the probabilityP(X ~ Y).
= ~~
P(X=
Vt)
~ == vn) ~
V\.Zwt
00
=
2:
V\
f(X~~) ~
Vl=-l
I
P(Y:: WI)
,..:J
"
)l.
JIYI
:£ P(Y= v.,}:: 2_(jj .....,::,
Wt:=l
::. 7Jl- (-5)V1
I
P(X ;>: Y)
~ Page 6 of 10
---
--------
~-----
-----
-
SOLUTION 4. The random variables Xand Y have the joint density
fx,Y (x,y) =
2, O~x~y~l { O h . , ot erwzse.
Determine the probability density of Z = X + Y. You must provide values of the density for all possible values of z.
An answer not supported by appropriate reasoning will not receive any credit.
}}-&,'l~J,)dtcdy
xfr~r
..
Page 8 of9
i
'
ECE 250 Final Exam December 8, 2016
SOLUTION
1. Consider the random variables
Here N 1 and N2 are independent Poisson variables
Uncorrelated? Independent? (Circle One)
8 Yes
Page 2 of 17
No
SOLUTION
2. The random variable X has a continuous probability density. The median of X is the value M that satisfies
Fx (M) = 1- Fx (M). Let W be the value that minimizes E[IX- WIJ. Prove that W = M.
Page 4 of 17
SOLUTION
3. N(t) is a Poisson process (independent increments, N(O) = 0, constant rate A.). Show that, for c > 0,
:i: ~ N;t)- AI~ e)~ o. 00
C h..e by'> [4ev
IVl~oa(t t
f(/ ~ - >.) ~,) ~
Va r ( ~- ;)
[HINT: Begin by evaluating the moments ofN(t).]
....
,
'
Page 6 of 17
----~-----~---'
SOLUTION 4. The non-random function g(t) is periodic with period T. That is g(t + T) = g(t). The random variable A is uniformly distributed on the interval [0, T]
fA(a)=
2_ { r' 0
O~a~T otherwise.
'
Consider the random process X(t) = g(t +A). Is X(t) wide sense stationary?
An answer not supported by appropriate reasoning will not receive any credit.
E[ XW] =- f"~(t-1-~'<){.',(,._)o(~ ;; 5~a.('t+ot)ol~ -~
~ Jv.te~val of q
~~;~og~ ;-;~':; LS
0
=
T 0
9.et g
t+T
r
= t fo(
t Llj(~>4(3 =
COM(jf;:
COMSi-
"
: . t 1)( tt-'i" >t< HA ~)cl"' = t- nt(tt-1:" ¥) ?ft+...)dol. 0 +oC
i;-14
-+f ~('t+(3)~(
5~ t ~::. t"+-d..
t+T
V
~=.
(J) d (3 .
~('C+~)~LP) r~
T
V
tf ?('C+~)~{f')o{p 0
= tarrcf,o.., W.S.S.?
No
(Circle One)
Page 8 of 17 I I________
f~r,odtc
fo., ction w1tn f~f'lo.l
CJ~ I y of
•
0\
77
T
SOLUTION 5. Consider the Poisson process N(t) (independent increments, N(O) = 0, constant rate A.). Denote by ~k the time between the (k- 1)st event and the k + n event. Let TM be the time until the M-th event
Th-e. f, »J~?
b~w-U.~? ~Jj Ct l.Qytt- ~V-tt~tfS lt ~ l Vl~p4Hcl4111t
M
TM=
IAt
k=l
Evaluate the mean and variance ofTM.
If
.ev-e,.,t5 are
t ..,cJ.ep- J r}te V a i't a VI r.l o1f o. ~ u VII l?
TJ..tt SOYI1
o { tk' llarz QlfCQ. ~
IE-[-TM_]_=_M_>_ _ _ _ _ _ _v_ar_[_TM_]_=_~_z._ _ _ _ _ _ L-
Page 10 of 17
___J\
SOLUTION 6. A zero-mean, white Gaussian process with constant power spectral density Po is passed through a linear, time-invariant filter
A ltnear
Y(t) = s:h(a)X(t-a)da. Here
h(t) =
sinm
--,
-00
< t < 00,
. Jrt
Evaluate the probability density ofY(t).
op.eratron OIA a ft"O uss r~sol'ts
GauGt;ra~ 1 n ct111o tker &aus~ta~ frrJCilSS . .
. . . Y (1)
tS,
a G-ao"'- var1 ct~le
An answer not supported by appropriate reasoning will not receive any credit.
+:rc·o C)J
t~-wr) :z
=
m: ~LYtt1] Va r[Yt-ln
J.r e;-::z.r:z.
___,___-==-= fvYJ ~ t-ra H s-1-or I ('..,___..,..-_ _
Hfiw)= r-Bct(~ )
=Po rtct( *1
K'
z
;:;
-::. { Po ) -'IT ~Wf. rr ')X ( W) 0 ) otJ..QrWI'-e, (r.-) ... J.,. l rr ;w~ ~ "211" ) d.w ~ Po ~/)1JTr
Pee
-
-'tT
l ) -n-~w~ n{ 0) CJTittlrwr s..e
=~
rr'Y;
~
:: Shlrf) X{t-o<)ol~ = o
E2 [ Yt-t)J
~
..
I.___ fv-(t)(-x)_=
...
~---~-~_:_ _ _ _ _ _ _ _---'1
_V_:Z_Tr_Po__
Page 12 of17
•
SOLUTION 7. N(t) is a Poisson process (independent increments, N(O) = 0, constant rate A).
P(N(t)- N(s) = n) = [A-(t ~s)f e-A,(t-s), n.
n =0,1, .... t ~
s.
Evaluate the conditional expectation E[N(t+r)IN(t)=m],
r~O,
m~O.
An answer not supported by appropriate reasoning will not receive any credit.
P( N('t~)=.., l N lt)=~ )::! P({Nl-l--+'L)-NHJJ-t-Nit)=- n fi.J lt)= Yn) N Lt-+'f )- Nl t)1 aKo/lJ lt}
a.~ 1Kd..e.~..fl.va ~tti.OK- ov-erlo.ppo .. r ~vt.T'.ervals
:. F({N(t+'i)-NLi--J}::Vl-n,) ~lt)=~) ::: p( {N (t+~)- N [-1:)~=- n- n, N ( t)=- ~)
-
? ::
Now E..( N l+-t't) l N [t):: V)'J]
e
=.
)
P( ~It) =-m)
P( ~(++'C)-IJ{t)= VJ-m) ~ p ~
zn
~~m~
( N l '1:-fJL) =V7 l N [i):::. m)
~
«»
:3>
2:.., P(NltH)-1./It> .. VI-In) ~ =="'~(k +-~"') f'( Nl-t¥f)- Nlt)::. P.) =
()()
(~
) ...
:: 2:_( /cz+m) 01:; ~tt::o
Jq. ..
e- ).1;"
= W1.,.. )if
+-r-)I_N-(t)_=_m_]=_V"'l_+_~_rr
IL-E-[N_(_t
_ _ _ _ _ _ _ _ _ _ _ ____.JI
Page 14 of 17
SOLUTION 8. The input, X(t), and output, Y(t), of a linear system are related via the differential
f-H Lt.V) =
equation
2.
I
d2 d d dt 2 Y(t) + 2 dt Y(t) + Y(t) = 2 dt X(t) + 2X(t).
-
r
:l(Jtv) +:Z(i~)-1-l,'~t.~):z.
:z_
.d . . h- l(l.+iW) . . If the mput 1s a zero-mean, w1 e sense statiOnary process w1t corre a.:tiO~n!.__--------..... 2 3 Rx( = (3/4)e -
r)
1rl,
5.z{w):
determine the correlation function of Y(t).
4 +W
..
An answer not supported by appropriate reasoning will not receive any credit.
)
-l'Lt
=-
I'-R-y( - r_)
_=
2.
e
-
4 (+{v2
4 4+t.V"-
- :2.l'Cl
B
z_e_-_''£_1-_e_2_(1_l_____________,l Page 16 of 17