Scribd
Upload
4 views

lecture14

Uploaded by

waleed ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
4 views

lecture14

Uploaded by

waleed ali
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 9

EE359 – Lecture 14

Outline
 Announcements:
 MT today, 2-4pm, 103 Hewlitt, pizza
afterwards
 HW posted today, due next Friday

 Discrete-Rate Adaptive
Modulation
 Introduction to MIMO
Communications
 MIMO Channel Decomposition
Review of Last
Lecture
 Introduction to adaptive modulation
 Varydifferent parameters of modulation
relative to fading
 Variable-rate variable-power MQAM
 Maximize average throughput by changing
rate and power 1
 Optimal power adaptation is water-filling

P()  1  1K   K K 0 0

 0
P  0 else 1
K
gk g
   adaptation:
power loss K=-1.5/ln(5BER).
 Optimal
R

rate
log   p( )d . Equals capacity with effective
B  
2
K K
Constellation
Restriction
 Restrict MD(g) to {M0=0,…,MN}.
 Let M(g)=g/gK*, where gK* is optimized
for max rate
 Set M (g) to max M : M  M(g)
D j j j
(conservative)
M(g)=g/g *
 Region
M 3boundaries are gj=MjgK*, j=0, K

M(g)=g/gK
…,N
M (g)
D
M
M 3
 Power control maintains target BER
2
M 2
M1 M1
Outage
0
g0 g1=M1gK* g2 g3 g
Power Adaptation and
Average Rate
 Poweradaptation: Fixed BER within
each region
 Es/N0=(Mj-1)/K
 Channel inversion within a region

Pj ( ) ( M j  1) /(K )  j    j 1 , j  0
 Requires 
P power
 0 increase when   1 increasing
M(g)

R N
 log 2 M j p ( j    j 1 )
B j 1
 Average Rate
Efficiency in Rayleigh
Fading
Spectral Efficiency (bps/Hz)

Average SNR (dB)


Multiple Input Multiple
Output (MIMO)Systems
 MIMO systems have multiple transmit
and receiver antennas (Mt at TX, Mr at
RX)
x=[x ,…,x ] y=[y ,…,y ]
1 Mt 1 Mr

TX power constraint:
E[xxH]=r=P/s2 y=Hx+n

 Inputdescribed by vector x of
dimension Mt
 Output described by vector y of
dimension Mr
 Channel described by M xM matrix
r t
MIMO Decomposition
 Decompose channel through
~ ~
transmit precoding (x=Vx) and
receiver shaping (y=U
~ ~y)
H
~
y=Hx+n y=S x+n
H=USV H
~y =s ~ ~
i ix+ni

 Leads to RHmin(Mt,Mr) independent


channels with gain si (ith singular
value of H) and AWGN

MIMO Fading Channel
Capacity
 Ifchannel H known, waterfill over
space (fixed power at each time
instant) or space-time
 Without transmitter channel
knowledge, capacity is based on an
outage probability
P is probability that channel capacity
out
given the channel realization is below the
transmission rate C

 Massive MIMO: random channel


Main Points
 Discretizing the constellation size in adaptive
MQMA results in negligible performance loss.
 Constellations cannot be updated faster than 10s to
100s of symbol times: OK for most dopplers.
 Estimation error/delay causes error floor
 MIMO systems exploit multiple antennas at
both TX and RX for capacity and/or diversity
gain
 With TX/RX CSI, decomposes into
independent channels
 Capacity of MIMO systems
 Static channel with TX/RX CSI: sum of capacity on each
spatial dimension
 Static channel without TX CSI: capacity metric is outage.
 Fading channel with TX/RX CSI: water-fill power over
space or space-time to achieve capacity

You might also like