Digital Signal Processing Reference
In-Depth Information
[17]. Channel autocorrelation is calculated using the channel estimates over the known
field of the transmitted data.
Instead of using channel estimates, the received signal can also be used directly in
estimating Doppler spread information. In [18], the Doppler frequency is extracted
from the samples of the received signal envelope. Doppler information is calculated as
a function of the squared deviation of the signal envelope. Similarly, in [19] the mobile
speed is estimated as a function of the deviation of the averaged signal envelope in flat
fading channels. For dispersive channels, pattern recognition, using the variation of
pattern mean, can be used to quantify the deviation of signal envelope. In [20], the fil-
tered received signal is used to calculate the channel autocorrelation values over each
slot. Then, the autocorrelation estimate is used for identification of high- and low-speed
mobiles. In [21], multiple antennas are exploited, where a linear relation between the
switching rate of the antenna branches and Doppler frequency is given. Also, the level
crossing rate of the average signal level has been used in estimating velocity [22, 23].
1.3.1.2 Frequency Selectivity Measure: Delay Spread
The multipath signals that reach the receiver have different delays as the paths that the
signals travel through have different lengths. When the relative path delays are on the
order of a symbol period or more, images of different transmitted symbols arrive at
the same time, causing intersymbol interference (ISI). Delay spread is one of the most
commonly used parameters that describes the time dispersiveness of the channel, and
it is related to frequency selectivity of the channel. The frequency selectivity can be
described in terms of coherence bandwidth, which is a measure of range of frequen-
cies over which the two frequency components have a strong correlation. The coherence
bandwidth is inversely proportional to the delay spread [24]. Figure 1.3 shows the effect
of time dispersion on channel frequency variation and channel frequency correlation, as
well as the corresponding power delay profiles.
1
2 Ta p
4 Ta p
8 Ta p
0.8
0.6
0.4
10
0
0.2
0
1
2
3
4
5
6
7
8
9
Separation in frequency (MHz)
0.8
2 Ta p
4 Ta p
8 Ta p
0.6
0.4
2 Ta p
4 Ta p
8 Ta p
0.2
10
1
0
0.5
1
1.5
2
2.5
3
3.5
4
2
4
6
8
10
12
14
16
18
20
Frequency (MHz)
Excess delay (nano sec)
(a)
(b)
FIgure 1.3
Illustration of the effect of time dispersion on channel frequency variation, chan-
nel frequency correlation, and delay spread. (a) Channel frequency variation for different delay
spread values. (b) Channel frequency correlation as a function of separation in frequency and the
corresponding power delay profiles.
Search WWH ::
Custom Search