Image Processing Reference
In-Depth Information
at different times (delay spread) and with different phase shifts at the receiver and overlap. his has
two consequences:
•
Overlapping signals can interfere constructively or destructively. Destructive interference
may lead to up to dB loss of received power. Such a situation is often called a deep fade.
•
Delay spread leads to inter-symbol interference, since signals belonging to neighbored
information symbols overlap at the receiver.
If the stations move relative to each other or to the environment, the number of paths and their phase
shifts vary in time. his results in a fast fluctuating signal strength at the receiver (called fast fading
or multipath fading). It is important to note that these fluctuations are much faster than those caused
by slow fading. Fast fading happens on the scale of milliseconds whereas slow fading happens at
scales of seconds or minutes. On the timescale of milliseconds, the mean signal strength is constant.
If the delay spread is small relative to the duration of a channel symbol, the channel is called non-
frequency-selective or flat, otherwise it is called frequency-selective.
These problems translate into bit errors or packet losses. Packet losses occur when the receiver fails
to acquire bit synchronization []. In case of bit errors, synchronization is successfully acquired,
but a number of channel symbols is decoded incorrectly. The bit error rate can, for example, be
reduced by using forward error correction (FEC) techniques [,]. [,].The statistical properties of bit
errors and packet losses were investigated in a number of studies [,,]. While the results are not
immediately comparable, certain trends show up in almost every study:
•
Both bit errors and packet losses are “bursty,” they occur in clusters with error-free periods
between the clusters. he distributions of the cluster lengths and the lengths of error-free
periodsotenhavealargecoeicientofvariationorevenseemtobeheavytailed.
•
Bit error rates depend on the modulation scheme, typically schemes with higher
bitrates/symbol rates exhibit higher error rates.
•
Wireless channel is much worse than wired channels, often bit error rates of
−
to
−
can be observed. Furthermore, the bit error rate can vary over several orders of
magnitudes within minutes.
Some knowledge about error generation patterns and error statistics can be helpful in designing more
robust protocols.
24.4.2 Wireless Transmission Techniques
A number of different transmission techniques have been developed to combat the impairments of
the wireless channel and to increase the reliability of data transmission.
Many types of WLANs (including IEEE .) rely on spread-spectrum techniques [], where a
narrowband information signal is spread to a wideband signal at the transmitter and despread back
to a narrowband signal at the receiver. By using a wideband signal, the effects of narrowband noise
or narrowband interference are reduced. The two most important spread-spectrum techniques are
direct sequence spread-spectrum (DSSS) and frequency hopping spread spectrum (FHSS).
In DSSS systems an information bit is multiplied (
XOR
ed) with a finite bipolar chip sequence such
that transmission takes place at the chip rate instead of the information bitrate. he chip rate is much
higher than the information rate and consequently requires more bandwidth; accordingly, the dura-
tionofachipismuchsmallerthanthedurationofausersymbol.hechiprateischosensuchthat
theaveragedelayspreadislargerthanthechipduration,thusthechannelisfrequency-selective.
Receivers can explore this in different ways. To explain the first one, let us assume that the receiver
receives a signal
S
from an LOS path and a delayed signal
S
′
from another path, such that the delay
difference (called lag) between
S
and
S
′
is more than the duration of a single chip. he chip sequences
