- Can
I export Coyote data into my favorite mapping program?
- Yes.
Coyote supports a variety of post-processing software.
- Can
I use Coyote in GPS-based drive-studies?
- Yes.
Coyote is available with an optional, removable 12-channel GPS receiver and supporting software perfect for drive-studies.
- How
do I get the data into my PC?
- Use
the Coyote Data Logger software
to download the data off the CF storage card included with your
Coyote.
What
is the significance of Rayleigh Fading in coverage measurements
with 40 Lambda criteria?
There
are two types of fading that are relevant to performing coverage
measurements, fast fading, also know as Rayleigh fading, and
terrain based fading due to obstructions and propagation loss.
The goal of coverage measurement is to measure the local signal
strength in presence of terrain based fading.
Rayleigh fading is due to close in reflections. It produces drops in signal
strength that are only a fraction of a wavelength in size (often on the scale
of inches). These fades always exist and are only of interest in the sense
that coverage measurements must be made in a manner that rejects Rayleigh
fading, so that actuate measurements of terrain based fading can be made.
William C. Y. Lee derived a well accepted criteria for removing Rayleigh
fading and retaining slower terrain based fading, known as 40 Lamba averaging
[W.C. Y. Lee and Y. S. Yeh, “On the Estimation of the Second-Order Statistics
of Log Normal Fading in Mobile Radio Environment”, IEEE Trans. Comm., vol.
22, June 1974, pp. 869-873] (You can find an easier to read version of this
criteria in most mobile communications text books). In order to remove Rayleigh
fading, you should average data for a time period equal to the time it takes
to traverse 40 wavelengths in your measurement vehicle and you should have
no less than 36 to 50 points in that time. For the remainder of this text
we will assume 40 points to make the calculations easier.
The biggest misconception with applying the 40-Lamba criteria to coverage
measurements is that all measurements should be made with an averaging time
exactly equal to 40-Lamda. While 40-Lamda is the minimum filtering that needs
to be done to reject Rayleigh, it is not the most appropriate averaging time
for coverage analysis. In almost all coverage analysis, the 40-Lamda averaged
samples must be averaged further. Due to the accuracy and reporting rate
of position measurement devices (GPS for instance) and the real practical
problem of too much data this extra averaging is valid and necessary. Clearly,
the averages of averages is just a longer average, and the resulting data
no longer has an integration time of forty wavelengths.
That is not to say that the 40-Lamda criteria does not apply to coverage
analysis. In fact it is very important. From the criteria, the minimum sampling
rate is about once per wavelength at the fastest vehicle speed. At 800MHz,
for example, a wavelength is about a foot. Using the 40-Lamda criteria, an
initial sampling rate of 88 samples per second is needed to make measurements
with a maximum drive speed of 60 mile per hour.
It is important to realize that Rayleigh fading is a spatial phenomenon not
a time one. The averaging is actually intended to be done over a distance
of forty wavelengths. Since most people performing coverage analysis average
samples spatially after they are collected. That is to say, all the samples
within a grid on a map are averaged to produce one point. To insure rejection
of Rayleigh fading, the grid size must be larger than forty wavelengths.
This in general is not a problem. At 800MHz, forty wavelengths is about 40
feet. Generally much larger grid sizes are used.
A
Practical Example
A BVS Coyote receiver and Gator transmitter are going to be used to do pre-site
survey at 800MHz. The Gator transmitter is hoisted up on a crane and the
drivers are sent out to perform the drive study.
Since the Coyote receiver, in single channel mode, has an internal sampling
rate of 512 samples per second regardless of internal averaging, we can calculate
the maximum drive speed that is allowed. Since we want to have at least one
sample per wavelength, as per 40-Lamda, the maximum speed would be 1ft*512/sec
or 349 miles per hour! Obviously, the Coyote can be used for coverage analysis
at any speed.
Since 512 points per second is far too much data to be processed, further
data reduction is desired. The Coyote has internal filtering and decimation
and the output rate can be reduced to 1 report a second. Each report is the
average of all data received during that second, so we still have an initial
sampling rate of 512 samples per second. This setting is the most convenient
since the internal GPS receiver in the Coyote reports position once per second.
At this point we have statistically correct data samples each with a position
stamp.
The data should be spatially averaged in post processing. Since forty wavelengths
is only forty feet at this frequency, almost any convenient grid size will
do. Remember that the accuracy of GPS, with selective addition off, is only
about 30 meters, so grid sizes larger than that would be appropriate. In
expensive post processing software the grid size can be selected. When processing
data yourself, it is often convenient to make the grid based on Latitude
and Longitude from the GPS receiver. In the New York area, one degree of
latitude is 69 miles and one degree of longitude is 50 miles. Therefore,
averaging data with the first three decimal places the same would be an easy
way of getting an appropriate grid size. |
