Properties of the Fourier Transform:
x(t t0 ) X(f )ej2f t0
x(t) X(f )
x(t) ej2f0 t
X(f f0 ), if f0 is real
1
x(t) cos(2f0 t)
[X(f + f0 ) + X(f f0 )]
2
j
x(t) sin(2f0 t)
[X(f + f0 ) X(f f0 )]
2
d
x(t) j2f X(f )
dt
x(t) v(t) X(f )V (f )
x(t)v(t) X(f ) V (f )
Common Fourier Transform Pairs:
(t) 1
rect(t/T ) T sinc(f T )
1
[(f + f0 ) + (f f0 )]
cos(2f0 t)
2
j
[(f + f0 ) (f f0 )]
sin(2f0 t)
2
Autocorrelation function and spectral density for deterministic signals:
Z
Rx ( ) =
x(t)x(t + ) dt = x( ) x( ), for an energy signal
x (f )
Rx ( )
Gx (f )
|X(f )|2
Z T20
1
x(t)x(t + ) dt, for a power signal with period T0
T0 T20
X
|cn |2 (f nf0 )
Random Variables:
Z
FX (x)
p() d
mX
x p(x) dx
PX
x2 p(x) dx
2
X
PX m2X
WSS Random Processes:
mY
= mX H(0), when X(t) is filtered to get Y (t)
RY ( )
= RX ( ) Rh ( ), when X(t) is filtered to get Y (t)
GY (f )
GX (f )|H(f )|2 , when X(t) is filtered to get Y (t)
Sampling and Quantization:
Xs (f )
n
1 X
X f
Ts n=
Ts
SNRpeak
3L2
log2
1
2p
for p% peak distortion
Baseband Modulation:
R
W
(i) bi-NRZ: zeroes are represented by negative (V ) and ones are represented by positive (+V ) pulses
(ii) uni-NRZ: zeroes are represented by no pulse (0) and ones are represented by positive (+V ) pulses
(iii) bi/uni-RZ: same as for NRZ, but, pulses return to zero level at half-bit ( T2 ) duration
(iv) NRZ-M: zeroes maintain the previous voltage and ones toggle it
(v) Manchester coding (bi--L): ones are represented by half-bit wide pulses in the first half of the bit
duration and zeroes are represented by half-bit wide pulses in the second half of the bit duration
(vi) Duobinary & precoding:
bandwidth efficiency =
yk
xk + xk1 , for duobinary signaling with xk being pulse (1) voltages
wk
dk wk1 , for precoding with dk being binary (0 or 1) data
Baseband Demodulation:
Z
i2 (t) dt
Z
=
s(t)
1, and,
j (t)k (t) dt = 0 for an orthonormal basis
= a1 1 (t) + . . . + aN N (t), with an N -dim basis, where,
R
R
s(t)j (t) dt
, and, Kj = j2 (t) dt
=
Kj
Z
N
X
=
s2 (t) dt =
a2j is the waveform energy
aj
Es
(a2 )
02 ln( P
a1 + a2
P (a1 ) )
+
is the decision threshold; here, ai = si (t) h(t)|t=T
2
(a1 a2 )
!
Ed
, where, Ed is the energy in s1 (t) s2 (t)
Q
2N0
a2
a1
P (0)Q
+ P (1)Q
a1 a2
Q
2
s
E
(1
)
b
Q
N0
r
PB
PB
PB
PB
R
=
s1 (t)s2 (t) dt
is the correlation coefficient
Eb
h(t) = s1 (T t) s2 (T t) for the matched filter
The raised cosine (RC) spectrum with bandwidth W and symbol rate Rs = 2W0 has the characteristic:
1
for |f | < 2W0 W
2 |f |+W 2W0
for 2W0 W < |f | < W
H(f ) =
cos 4 W W0
0
for |f | > W
W W0
r =
, is the roll-off factor
W0
1
(1 + r)Rs
W =
2
The Range Equation:
EIRP = Pt Gt
Ae = Ap for an antenna with efficiency
Ls =
4d
Pr =
G=
M=
2
is the free space loss at distance d
EIRP Gr
is the power received
Ls Lo
4Ae
relates antenna directional gain to effective area
2
EIRP Gr
is the link margin (k = 1.38 1023 J/K)
(Eb /N0 )reqd RkT o Ls Lo
Noise Analysis:
o
o
N = kTeq
W relates noise power N to noise equivalent temperature Teq
and bandwidth W
To
To
, and,
290
(F 1) 290 relate noise figure to noise temperature
1+
F = L for a lossy line with loss factor L
= G1 G2 . . . Gn
F2 1 F3 1
Fn 1
= F1 +
+
+ ...
G1
G1 G2
G1 G2 . . . Gn1
To
T3o
Tno
= T1o + 2 +
+ ...
G1
G1 G2
G1 G2 . . . Gn1
Gcomp
Fcomp
o
Tcomp
o
Tsys
= TAo + TLo + LTRo is the system noise temperature
Satellite Repeaters:
PB Pu + Pd for a regenerative repeater
P
1
, where, PT = k Ak Pk is the total received power, and EIRPs is the variable gain
PT + kTs W
Pr
N0
=
ij
EIRPs j Ai Pi
for a non-regenerative repeater (Ai and j are U/L and D/L attenuations, resp.)
EIRPs j Ns + Ng
Eb
N0
1
=
ov
Eb
Ns
1
+
Eb
Ng
1
relates the overall SNR to U/L and D/L SNRs
d
M -PSK
r
si (t)
PE
PB
PB
PB
2ES
2i
cos 2fc t +
T
M
!
r
2ES
2Q
sin
N0
M
!
r
2Eb
Q
for BPSK and QPSK
N0
1
Eb
exp
for DPSK
2
N0
PE
for all other PSK
k
2
.
T
M -FSK
r
si (t)
PE
PE
PB
2ES
cos (2fi t)
T
!
r
ES
(M 1)Q
for coherent reception
N0
ES
M 1
exp
for non-coherent reception
2
2N0
M/2
PE
M 1
M
for coherent reception
2T
M
for non-coherent reception.
T
M -ASK
r
si (t)
2Ei
cos(2fc t)
T
2
.
T
Random Access
= G e2G , for Aloha with G = t
= G eG , for S-Aloha with G = t .
Spread Spectrum
GP GV GA
, where, GA = GV = 2.5, = 1.5, H0 = 1.55
(Eb /I0 )reqd H0
J0
2
J
= Jammer
2W
2
for
(Eb /J0 )
=
1
for
(
1
e
for
(Eb /J0 )
=
E
1 2Jb0
for
2e
0
PB,max
PSD when jamming bandwidth W with power J
(Eb /J0 ) > 2
for optimum partial band jamming
(Eb /J0 ) 2
(Eb /J0 ) > 2
(Eb /J0 ) 2
with optimum partial band jammer
The PDF of a Gaussian random variable with mean m and variance 2 is
1
(x m)2
p(x) =
exp
2 2
2
Bayes Rule:
P (A) =
P (A|Bi )P (Bi ).
Jensens Inequality:
PN
If ai are non-negative numbers with 1 ai = 1, then,
N
X
ai log(zi ) log
i=1
N
X
!
ai zi
i=1
Rx
If a(x) is a non-negative function of x with x12 a(x)dx = 1, then,
Z x2
Z x2
a(x) log(b(x)) dx log
a(x)b(x) dx .
x1
x1
Entropy and Information :
H(X)
H(X, Y )
H(X|Y )
1
p(xi )
xi
XX
1
p(xi , yj ) log
p(xi , yj )
xi yj
X
p(yj )H(X|Y = yj )
X
p(xi ) log
yj
p(yj )
yj
p(xi |yj ) log
xi
1
p(xi |yj )
!
1
dx
p(x)
H(X) H(X|Y ) = h(X) h(X|Y )
H(Y ) H(Y |X) = h(Y ) h(Y |X)
E[X Y ] for a binary source
M Z bi
X
E[(X Y )2 ] =
(x yi )2 p(x) dx for a continuous source
Z
h(X)
I(X; Y )
p(x) log
i=1
bi1
Hb (p) Hb (D) if D < p for a binary source.
2
1
R(D) =
log2
if D < 2 for a Gaussian source.
2
D
1
P
P
C = max I(X; Y ) = log2 1 +
= W log2 1 +
2
N
N0 W
p(x)
P
C =
log2 e
N0
p
= Pr[C Cp ], where, Cp is the p% outage capacity.
100
R(D)
EE 641: RF Wireless Communication Systems
List of Commonly Needed Expressions and Relationships
I. Fundamentals of RF Wireless Communication Systems
1. Spectral Efficiency spec = Data rate Rb (bits/sec) / Transmission bandwidth B (Hz)
2. Power Efficiency of modulation scheme, pow = Data rate Rb (bits/sec) / Radiated power Prad (Watts)
2. Power Efficiency of radiating antenna,
pow = Radiated power Prad (Watts) / Power drawn from source PDC (watts)
S
3. Shannons channel capacity
C = B log 2 1 + bits/sec
N
II. Electromagnetic Waves and Radiators
1. Maxwells Equations
( H )
( E )
E =
, H = J +
, .E = , .H = 0
t
t
2. Wave Equation
2 E + 2 E = 0,
2 H + 2 H = 0
3. Wave impedance of the medium
= / = 120 rel / rel
In free space, = 120 = 377
4. Propagation constant of the medium
= + j = j (2) ,
if = r + j i, then = + j = j [ r(1 j i/r ) ]
5. Phase velocity of the electromagnetic waves in the medium
vp (ohms) = 1/ () = c / (relrel) where c = 3108 m/sec
6. Poynting Vector
S=E H
7. Radiation Intensity due to a source at the origin
U(r, , ) = r2 . S (r, , )
8. Power radiated from a source at the origin
Prad = U ( , ) sin d d
=0 =0
9. Far Field condition
Conditions of Far Field :
R 2D2 /
R >> D
R >>
10. Fields due to a Hertzian Dipole of length z and current I, placed at origin along z axis.
E(r, , ) = Er(r, , ) ar + E (r, , ) a + 0 a
Er (r , , ) =
cos
I z 2
cos
e j r
j
2
2
( r )3
( r )
E (r , , ) =
sin j sin sin
j I z 2
e j r
+
+
4
( r ) 2 ( r )3
r
H (r , , ) =
j I z 2 j r sin sin
e
r ( r )2
4
Directivity
D( , ) =
3 2
sin
2
2
z
Rrad = 80 2
11. Dipoles: Short (Hertzian):
D = 1.5 = 1.76 dB
Half-wave:
D = 1.64 = 2.15 dB and Rrad = 73
Radiation resistance
III. Receiving Antenna Characteristics
2
PG
PG
t t
t t
A
=
Gr
eff
4 d 2
4 d 2 4
Pav
Aeff ( , ) @
Sinc ( , )
1. Friis Equation
Pr =
2. Effective Area
3. Reciprocity Theorem
G=
Aeff
4. Noise power available from a resistor
Pn,av = kTn,antB
5. Noise temperature of antenna
Tn,ant
1
=
sin d d D( , )TB ( , )
4 =0
=0
6. G/T Ratio G/T = [ 10 log10 Gant ] / Tant dB/K
IV. Physical Model of Wave Propagation
1. Reflection Coefficient
For E field parallel to ground
For E field in plane of incidence
Eref P
EincP
Eref
Einc
2 cos trans 1 cos inc
2 cos trans + 1 cos inc
2 cos inc 1 cos trans
2 cos inc + 1 cos trans
2. Transmission Coefficient
For E field parallel to ground
TP
EtransP
EincP
For E field in plane of incidence T
2 2 cos trans
2 cos trans + 1 cos inc
Etrans
22 cos inc
=
Einc 2 cosinc + 1 costrans
2
P
3. Power received under free-space propagation (Friis equation) : rec = Gt Gr
Ptr
4 R
4. Power received due to perfectly reflecting ground with antennas at heights ht and hr :
2
Prec
ht hr
= Gt Gr
Ptr
4R 4R
5. Normalized diffraction parameter
=h
2(d1 + d 2 )
d 1 d 2
6. Excess Path Loss due to diffraction from single knife edge
V. Empirical Models of Wave propagation
1. Delisle model of path loss in urban environment
4
2
17 r ( f /1MHz )
4.27
10
for hmob < 10m
hbs2 hmob
L=
4
2
4.27 1016 r ( f /1MHz )
for hmob > 10m
hbs2 hmob
2. Ikegamis model of excess path loss between two edges separated by ds :
( f /1MHz ) {(ho hmob ) /1m}2
L=
186 (d s /1m)
3. Okamura Hata model for VHF/UHF (150 MHz to 1 GHz)
L(dB) = 69.55 + 26.16log10 ( fc /1MHz ) 13.82log10 hbs a(hmob ) + (44.9 6.55log10 hbs ) log10 r C
Where
8.29 [log10 (1.54 hmob /1m) 1.1
for l arg e city and f c 300MHz
a(hmob ) =
3.2 [log10 (11.75 hmob /1m) 4.97
for l arg e city and f c > 300MHz
1.1 log [( f /1MHz ) 0.7] (h /1m) 1.56 log [( f /1MHz) 0.8]
for small city
10
c
mob
10
c
and C =
5.4 + 2 [log10 ( f c / 28MHz )]2
40.94 + 4.78 [log ( f /1MHz )]2 18.33 log ( f /1MHz )
10
c
10
c
for Urban area
for Sunurban area
for Open area
VI. Statistical Model of Wave Propagation
1. Rayleigh density function for received signal amplitude
y<0
0
2
y
f Y ( y) = y
exp
y>0
2
2
2. Exponential density function for received signal power
1
f P ( p) =
exp p / 2 2 u ( p)
2
2
3. Rician distribution in the presence of a strong signal
fX(x) = ( x / R2 ) exp [ - (x2 + A2)/ 2R2 ] . Io ( x A / R2 )
4. Error Probability in the absence of Fading
ES
S 1
1
PE = erfc
= erfc
2
N 2
NO
5. Error Probability in the presence of fading for Rayleigh-distributed signal
PE =
1
(S / N )
1
2
1 + (S / N )
6. Diversity Gain
G=
( S / N ) for diversity system
( S / N ) for sin gle channel
7. Doppler Shift Frequency
f D = f carr
v cos
c
VII. Channel Characterization
1. Given the power delay profile P(),
Average delay < TD >=
0
P( )
P( )d
0
[ (t ) T
Power delay spread D (t ) =
(t ) ] P( ) d
P( ) d
0
2. R.M.S. Delay spread
rms = D
or, Multipath spread TMUL = 2
3. Coherence time Doppler spread relationship:
Tcoh 1 / 2fD
4. Coherence Bandwidth r.m.s. Delay Spread Relationship
Bcor
1
2 rms
VIII. Multiple Access and Cellular Systems
1. Minimum signal-to-interference Power Ratio (for hexagonal cells with -th power law)
R
S
[3N c R] / 2
=
I min ( N c 1)
2. Erlangs B formula for the probability of call blocking with N available duplex channels, as a
function of total caller traffic intensity of U erlangs:
UN
P[ Blocking ] =
N
Um
N!
m =0 m !
ACI =
3. Adjacent channel interference ratio :
sig
( f ) | H BP ( f f ) |2 df
sig
( f ) | H BP ( f ) |2 df
IX. Noise and Interference
1. Power spectral density of noise at the output of a noiseless linear filter with frequency
response H(jf) excited at its input with a random signal of power spectral density Sx(jf) :
Sy(f) = |H(jf)|2 Sx(f)
2. Noise bandwidth (or noise-equivalent bandwidth) of a filter with frequency response H(jf):
Beq =
| H ( jf ) |
N o df
No
2. Noise figure of a linear system:
S /N
F = in in
Sout / N out Tn ,source =Tref
3. Equivalent noise temperature Teq = (F 1) Tref
or, Noise figure F = 1 + (Teq/Tref)
1
Tp
4. Noise figure of a passive filter at temperature Tp: F = 1 +
1
Gav Tref
5. Combined noise temperature of a cascade of n linear systems, Ttot = T1 +
Or, combined noise figure, Ftot = F1 +
F 1
F2 1
+ 3
+ ....
Gav,1 Gav,1Gav,2
X. Nonlinear and Intermodulation Distortion
For a memory-less nonlinear system with transfer characteristic
y = a0 + a1 x + a2 x2 + a3 x3 + ignorable higher-order terms
T3
T2
+
+ ....
Gav,1 Gav ,1Gav ,2
1. When excited with a harmonic signal of amplitude Vamp,
Gain compression
3
Gnonlin a1 + 34 a3Vamp
=
Glin
a1
Second-harmonic generation
Vamp|@2f = a1 Vamp2
2. When excited with two harmonic signals of frequencies f1 and f2, with equal amplitudes Vamp
Amplitude of intermodulation signal (at each of 2f2 f1 and 2f1 f2) : VIMD = a3 Vamp3
Intermodulation power ratio: IMPR VIMD2 / (a1Vamp)2 = (a3 / a12) Vamp4
Third-order intercept (TOI) point, referred to the input:
PTOI = Input Power ( a1Vamp)2 |@IMPR=1 = 2a13 / 3a3
3. Dynamic range
DR = [ PTOI / No ]2/3
