shannon’s channel capacity - university of albertahcdc/library/mimochclass/channelcapacity.pdf ·...
TRANSCRIPT
Shannon’s Channel Capacity
Shannon derived the following capacity formula (1948) for an additivewhite Gaussian noise channel (AWGN):
C = W log2 (1 + S/N) [bits/second]
• W is the bandwidth of the channel in Hz
• S is the signal power in watts
• N is the total noise power of the channel watts
Channel Coding Theorem (CCT):The theorem has two parts.
1. Its direct part says that for rate R < C there exists a coding systemwith arbitrarily low block and bit error rates as we let the codelengthN →∞.
2. The converse part states that for R ≥ C the bit and block errorrates are strictly bounded away from zero for any coding system
The CCT therefore establishes rigid limits on the maximal supportabletransmission rate of an AWGN channel in terms of power and bandwidth.
Bandwidth Efficiency characterizes how efficiently a system uses itsallotted bandwidth and is defined as
η =Transmission Rate
Channel BandwidthW[bits/s/Hz].
From it we calculate the Shannon limit as
ηmax = log2
(1 +
S
N
)[bits/s/Hz]. (1)
Note: In order to calculate η, we must suitably define the channel bandwidth W . One com-monly used definition is the 99% bandwidth definition, i.e., W is defined such that 99%of the transmitted signal power falls within the band of width W .
EE 7950: Statistical Communication Theory 2
Average Signal Power S can be expressed as
S =kEb
T= REb,
• Eb is the energy per bit
• k is the number of bits transmitted per symbol
• T is the duration of a symbol
• R = k/T is the transmission rate of the system in bits/s.
• S/N is called the signal-to-noise ratio
• N = N0W is the total noise power
• N0 is the one-sided noise power spectral density
we obtain the Shannon limit in terms of the bit energy and noise powerspectral density, given by
ηmax = log2
(1 +
REb
N0W
).
This can be resolved to obtain the minimum bit energy required for reli-able transmission, called the Shannon bound:
Eb
N0≥ 2ηmax − 1
ηmax,
Fundamental limit: For infinite amounts of bandwidth, i.e., ηmax → 0,we obtain
Eb
N0≥ lim
ηmax→0
2ηmax − 1
ηmax= ln(2) = −1.59dB
This is the absolute minimum signal energy to noise power spectral den-sity ratio required to reliably transmit one bit of information.
EE 7950: Statistical Communication Theory 3
Normalized Capacity
The dependence on the arbitrary definition of the bandwidth W is notsatisfactory. We prefer to normalize our formulas per signal dimension.It is given by [5]. This is useful when the question of waveforms andpulse shaping is not a central issue, since it allows one to eliminate theseconsiderations by treating signal dimensions [2].
Cd =1
2log2
(1 + 2
RdEb
N0
)[bits/dimension]
Cc = log2
(1 +
REb
N0
)[bits/complex dimension]
Applying similar manipulations as above, we obtain the Shannon boundnormalized per dimension as
Eb
N0≥ 22Cd − 1
2Cd;
Eb
N0≥ 2Cc − 1
Cc.
System Performance Measure In order to compare different commu-nications systems we need a parameter expressing the performance level.It is the information bit error probability Pb and typically falls into therange 10−3 ≥ Pb ≥ 10−6.
EE 7950: Statistical Communication Theory 4
Examples:
Spectral Efficiencies versus power efficiencies of various coded and un-coded digital transmission systems, plotted against the theoretical limitsimposed by the discrete constellations.
Unachieva
ble
Region
QPSK
BPSK
8PSK
16QAMBTCM32QAM
Turbo65536
TCM16QAM
ConvCodes
TCM8PSK
214 Seqn.
214 Seqn.
(256)
(256)
(64)
(64)
(16)
(16)
(4)
(4)
32QAM
16PSK
BTCM64QAM
BTCM16QAM
TTCM
BTCBPSK
(2,1,14) CC
(4,1,14) CCTurbo65536
Cc [bits/complex dimension]
-1.59 0 2 4 6 8 10 12 14 150.1
0.2
0.3
0.40.5
1
2
5
10
EbN0
[dB]
BPSK
QPSK
8PSK
16QAM16PSK
ShannonBound
EE 7950: Statistical Communication Theory 5
Discrete Capacities
In the case of discrete constellations, Shannon’s formula needs to be al-tered. We begin with its basic formulation:
C = maxq
[H(Y )−H(Y |X)]
= maxq
−∫y
∑akq(ak)p(y|ak)log
q(a′k)∑a′k
p(y|a′k)
−∫y
∑akq(ak)p(y|ak) log (p(y|ak))
]
= maxq
∑akq(ak)
∫y
log
p(y|ak)∑a′kq(a′k)p(y|a′k)
where {ak} are the K discrete signal points, q(ak) is the probability withwhich ak is selected, and
p(y|ak) =1√
2πσ2exp
−(y − ak)2
2σ2
in the one-dimensional case, and
p(y|ak) =1
2πσ2 exp
−(y − ak)2
2σ2
in the complex case.
Symmetrical Capacity: When q(ak) = 1/K, then
C = log(K)− 1
K
∑ak
∫n
log
∑a′k
exp
−(n− a′k + ak)2 − n2
2σ2
This equation has to be evaluated numerically.
EE 7950: Statistical Communication Theory 6
Code Efficiency
Shannon et. al. [4] proved the following lower bound on the codeworderror probability PB:
PB > 2−N(Esp(R)+o(N)); Esp(R) = maxq
maxρ>1
(E0(q, ρ)− ρR)) .
The bound is plotted for rate R = 1/2 for BPSK modulation [3], togetherwith selected Turbo coding schemes and classic concatenated methods:
N=44816-state
N=36016-state
N=13344-state
N=133416-state
N=4484-state
N=204816-state
N=2040 concatenated (2,1,8) CCRS (255,223) code
N=2040 concatenated (2,1,6) CC +RS (255,223) code
N=1020016-state
N=1638416-state N=65536
16-state
N=4599 concatenated (2,1,8) CCRS (511,479) code
N=1024 block Turbo code using (32,26,4) BCH codes
UnachievableRegion
ShannonCapacity
10 100 1000 104 105 106−1
0
1
2
3
4
EbN0
[dB]
N
EE 7950: Statistical Communication Theory 7
Code Performance
The performance of various coding schemes is shown below (see [2]). Notethe steep behavior of the turbo codes known as the Turbo cliff.
0 1 2 3 4 5 6 7 8 9 1010−6
10−5
10−4
10−3
10−2
10−1
1
Bit Error Probability (BER)
EbN0
BPSK
CC; 4 states
CC; 16 states
CC; 64 states
CC; 214
states CC; 240 states
Turbo
Code
References
[1] R.G. Gallager, Information Theory and Reliable Communication, John Wiley & Sons,Inc., New-York, 1968.
[2] C. Schlegel, Trellis Coding, IEEE Press, Piscataway, NJ, 1997.[3] C. Schlegel and L.C. Perez, “On error bounds and turbo codes,”, IEEE Communications
Letters, Vol. 3, No. 7, July 1999.[4] C.E. Shannon, R.G. Gallager, and E.R. Berlekamp, “Lower bounds to error probabilities
for coding on discrete memoryless channels,” Inform. Contr., vol. 10, pt. I, pp. 65–103,1967, Also, Inform. Contr., vol. 10, pt. II, pp. 522-552, 1967.
[5] J.M. Wozencraft and I.M. Jacobs, Principles of Communication Engineering, John Wiley& Sons, Inc., New York, 1965, reprinted by Waveland Press, 1993.