Digital Signal Processing Reference
In-Depth Information
BER
Available bandwidth fraction
F [%
]
5
6
6
1
6
7
7
2
7
6
8
0
8
3
8
6
Shannon limit
E
b
/N
c
≥ -1.59dB
F
available
≥ 58%
10
-1
1
2
1
hard decisions.
uncoded transmission.
3
10
−2
Coding gain ISSR (G
ISSR
)
2
10
−3
Turbo coding
rate=1/2, 5 iterations
2+3
10
−4
3
ISSR processing
no outer coder
10
−5
2+3
inner: ISSR
outer: Turbo coder
−2
−1
0
1
2
3
4
5
E
b
/N
c
[dB]
crossover ISSR/Turbo: approx. 0.3dB
Figure 3.7.
Simulated performance of issr as a function of the available band-
width. When issr is the inner coder of a Turbo coder, the overall
performance is within 0
.
4 dB of the Shannon limit.
Using the issr algorithm as the inner decoder, the performance of the com-
pound system comes within 0
.
4 dB of Shannon's theoretical limit. It is im-
portant to realize that, while coding techniques merely based on redundancy
cause a reduction in the effective throughput, the issr decoder does not affect
the data rate of the system.
The throughput of a wireless link is one way to measure its performance, how-
ever that's not the whole picture. In a practical implementation with limited
energy resources, it is important to properly distribute the available processing
power between issr and the outer coding mechanism. For example, in a chan-
nel that is uniformly contaminated by awgn noise, it does not make sense to
wasteenergyintheissr decoder. Also, Turbo coding costs around 10 times
more computing power than a viterbi fec decoder [Des03]. In a severe mul-
tipath environment, it is certainly worth the consideration to concatenate issr
with a less complex fec coding scheme. The reduction in power consumption
will more than compensate for the implementation loss that comes with a less
sophisticated error coding scheme.