Digital Signal Processing Reference
In-Depth Information
not be confused with channel estimation and equalization. Equalization is fo-
cussed on the compensation of the frequency response of the channel. Instead
of this, issr can decide to consider seriously corrupted frequency bands as
irreversibly lost and subsequently tries to reconstruct the original data using
the remaining, unaffected parts in the spectrum. On the other hand, issr is not
a holy grail system and cannot serve as a replacement for channel equaliza-
tion. issr is meant to be put in action when channel equalization fails due to
insufficient signal quality in some of the subbands. The ability to deliberately
ignore a specific part of the spectrum is also the biggest strength of issr.When
a certain subband suffers from destructive fading, equalization will boost the
gain for that particular band in an attempt to reduce isi. However, the useful
information within this band is often limited due to the background noise of
the wireless channel. A lot of noise is thus inevitably introduced by the channel
equalization filter.
The issr method is not willing to make this compromise between isi and back-
ground noise. It will simply reject those subbands with an insufficient contri-
bution to the overall signal quality (Figure 3.1). The rejection of subbands is
not only a useful feature for a channel that is plagued by frequency-selective
fading. The same feature can be thrown into the game to fight against interfer-
ence in some part of the signal spectrum. For issr, it is not important to know
exactly how a certain portion of the spectrum had gone missing. Rather the fact
that it is missing in the first place is a crucial step in the signal reconstruction
process of the issr method. Before jumping into the concrete specifics of issr,
it is important to fully recognize that issr is running at the client-side of the
wireless connection, hence it has the ability to react almost instantaneously to
changes and unexpected events which take place in the channel. If correctly
implemented, the performance of issr will be unmatched in comparison to
server-side modulation-aware coding techniques such as ofdm.
3.1
Principles of signal reconstruction
One of the target applications of the issr system described in this chapter
is the 802
.
11b wireless lan system, the oldest of the three Wi-Fi standards.
It was already mentioned in Section 2.6 that 802
.
11b data is encoded using
dsss, using the 11-bit Barker code or complementary code keying (cck)as
the frequency-spreading code [Con00]. Unfortunately, the effective spreading
factor of these codes is substantially limited as a result of the considerable
throughput of data (from 1 up to 11 Mbit/s) within a bandwidth of 22 MHz.
As a consequence, the promises made by the hologram labels on the package
of an 802
.
11b device are often not fulfilled by the device itself: frequency-
selective fading in the indoor environment or in-band interference originating
