ECE 6151, Spring 2017 Lecture Notes: 1 Outline
ECE 6151, Spring 2017 Lecture Notes: 1 Outline
Lecture Notes
Shengli Zhou
January 20, 2017
1 Outline
Syllabus
Communication and spectrum
Channel models
Passband and baseband conversion
Complex random variables
Complex stochastic process
2 Spectrum
How many receivers are inside your iPhone?
Bluetooth, <10m, 20 kh/s, carrier frequency 2.4 GHz, bandwidth 1 MHz
GPS frequency. The GPS L1 band (1575.42 MHz), chip rate 1.023 MHz.
3G/4G/LTE, 1-3 km, 100-500 kb/s, data; 10-20 kb/s, voice
WiFi, < 100m, 6-54 Mb/s, 2.4 GHz, 5.2 GHz. bandwidth 20 MHz
Legacy systems: AM [535 1605]kHz, bandwidth 10 kHz. FM [88 108] MHz, bandwidth 200 kHz.
GSM: uplink 890-915 MHz, downlink 935-960 MHz. bandwidth 200 kHz.
3 Channel Model
x(t) () y(t)
1
Ambient noise
AWGN: (toy channel, idealized)
y(t) = x(t) + n(t)
One physical channel can be fit into each of the models (e.g., telephone)
I & Q components
xl (t) = xi (t) + jxq (t)
x(t) = xi (t) cos(2fc t) xq (t) sin(2fc t)
Passband to baseband
Complex envelope
x+ (t) = xl (t)ej2fc t
2
4.2 Bandpass systems
Passband vs baseband filtering
Why do we want baseband signals?
x(t) h(t) y(t)
5 Random variables
What could be random for a communication system?
What fully characterize a random variable?
mean: Z
E(X) = xfX (x)dx
0
second-order moment Z
E(X 2 ) = x2 fX (x)dx
0
Variance:
E([X E(X)]2 ) = E(X 2 ) E(2E(X)X) + [E(X)]2
= E(X 2 ) [E(X)]2
3
Gaussian
A particularly important pdf: Gaussian N (, 2 )
1 (x)2
fX (x) = e 22
2
where
E(X) = , var(X) = 2
We also have
E[(x )3 ] = 0, E[(x )4 ] = 3 4
Why Gaussian important?
Why important ?
Central limit theorem tells us that
n
X
Y = Xi
i=1
is Gaussian if n is large enough
We in this class focus on Gaussian and its variants.
Joint Gaussian
X and Y are independent and Gaussian distributed
f (x, y) = f (x)f (y)
Complex Gaussian
Complex Gaussian: Z = X + jY , X, Y are independent Gaussian with the same variance x2
f (Z) = f (X, Y ) = fX (x)fY (y)
(x x )2 (y y )2
1
= exp
2x y 2x2 2y2
z = x + jy
z2 = E[|z z |2 ] = E[(x x )2 + (y y )2 ] = 2x2
|z z |2
1
fZ (z) = exp
z2 z2
Rayleigh
X N (0, 2 ), y N (0, 2 ), X and Y independent
What is the pdf of R = X 2 + Y 2 ?
Z = X + jY = |Z|ej . What is the pdf of |Z|? What is the pdf of ?
X = R cos , Y = R sin
dx cos r sin dr
=
dy sin r cos d
dxdy = rdrd
1 x2 +y2 2 1 r22
e 2 dxdy = e 2 rdrd
2 2 2 2
1 r2
f (r, ) = e 2r
2 2
r r2
f (r) = 2 e 22
1
f () =
2
4
Uniform distribution
Uniform distribution X U (a, b) (
1
ba x [a, b]
fX (x) =
0 otherwise
Rician
x = cos(), y = sin()
2
r 2r cos( + ) + 2
r
f (r, ) = exp
2 2 2 2
Z 2
1
I0 (x) = ex cos d
2 0
2
r + 2
r
f (r) = 2 exp I0 (r/ 2 )
2 2
Chi-square
Pn
Central Chi-square: Y = Xi2 , where Xi N (0, 2 )
i=1
Pn
Non-central Chi-square: Y = i=1 Xn2 , where Xi N (i , 2 )
Collections
(x)2
Gaussian: fX (x) = 1 e 2 2
2
r2
r 2
Rayleigh: f (r) = 2 e
2
2 2
Rician: f (r) = r
2 exp r 2
+
2 I0 (r/ 2 )
5
Whiteboard pictures
6
7