Professional Documents
Culture Documents
An appraisal of propagation
modelling methods for point-point
and point-multipoint microwave
networks in the bands 2.3GHz
to 42GHz
Solutions in Radiocommunications
Page 2 of 8
Introduction
Context
Putting the point to point and point to multi-point link scenario into context we see a network topology something
like that shown in the following view.
Overview
Finally I will also look at some of the constraints on the network deployment covering in turn the modelling of interference and the achievement of coexistence between networks
and specifically between networks of differing technologies.
As a round up and by way of conclusion I will also look at
probability in planning and will draw attention to the main
probability distributions and how they affect what we as
planners and integrators can deliver by way of a service to
the end user or subscriber.
P COM
P COM
P COM
P C OM
Path Geometries
P COM
P COM
P COM
P C OM
P C OM
P COM
P COM
Fibre Backbone
Network Management
Software
ogy now allows the profile to be assembled with significantly enhanced data and millions of profiles can be analysed in seconds.
Environmental Models
The Digital Elevation Model
The DEM is a digital terrain model describing ground
heights and a buildings elevation model combined. It describes the maximum or canopy height at any point on the
ground. It is described generally by a matrix of points in
the x and y or Eastings and Northings directions with the
axes aligned to a chosen coordinates system. The matrix
has a given resolution. For planning mobile systems and
for microwave systems where every path will be surveyed
a resolution giving a height point every 50 metres is usable. For PMP networks at 28GHz we need to achieve a
resolution of nearer 1 metre to position nodes and subscribers more precisely.
Page Page
3 of 8 3 of 88
Rooftop Definition
One highly useful data file is that defining the roof outlines.
This can in turn be further expanded to describe urban, suburban or postal zone rooftops hence allowing filtering and
more especially relocation of errored terminals. An example
of this file is shown below.
Given this definition we can immediately suspect the business subscribers in the roads and in parks and take corrective action. Such a file also allows the reporting of rooftops
covered or other definitive performance metric.
Whilst the DEM is in fact all that is needed to permit planning to proceed, we do need to overlay the rather bland
model with something that is meaningful to recipients of the
planning work. Typically this is an image of some form.
Page 4 of 8
Page Page
4 of 84 of 8
Page 5 of 8
Page Page
5 of 84 of 8
Note that there is a cross over in frequency where the predominant mechanism rain or atmospheric multi-path and
super-refraction takes over. Calculation of both and assessment of correlation between the two allow a final path
availability to be reached.
Environmental Multi-path
At lower frequencies where we need to work over an obstructed path there will be significant multi-path but here the
reflections will be from the buildings in the vicinity of the
terminals. In modelling this it is useful to consider the scenario as two vectors one representing the direct ray and
the other the vector sum of the multi-path at any point in
time.
This resultant obeys a Ricean distribution. The more in the
clutter that the terminals are, the higher the multi-path and
the higher the standard deviation of the distribution. Referring to ITU-R PN.1057 we see that typical standard deviations lie in the region of 4.5-7.5dB. Given that we want to be
n standard deviations confident of receiving a usable signal,
we can calculate the fade margin needed to achieve this.
Multi-path
Direct
Resultant
Page 6 of 8
Page Page
6 of 82 of 8
-44dBm
Margin
14dB
-79dBm
-83dBm New Noise Floor
-97dBm Noise Floor
Frequency
B2
C/I Methods
B1
F1
Overlap of B1
Rise in Noise Floor =
10LOG(10^-97 + 10^(-79+10LOG0.4))=-83dBm
Threshold Impairment = -83 97 = 14dB
Frequency
N=0
-44dBm
Margin
Basic C/I
-79dBm
N=0
F2
Conclusions
In setting out to design and analyse a network plan in the
2.3GHz to 42GHz region we must understand that there are
several probability distributions in play.
The first is the basic ability to predict the performance of a
path. The environmental model that we use will be errored.
Provided that we know that error we can determine the certainty of the path being as we describe.
The second is the physics of the path and the fact that various mechanisms affect the propagation. Knowing something about where the path is and what factors affect it allows us to predict how the path, so modelled will perform
and what availability it will provide.
The third is that unwanted signals will be received which
degrade the wanted link performance for small percentages
of time. This interference degrades further the availability.
The fourth has not so far been discussed and covers the
scenario that given a diffracted path heavy in reflections, we
will experience a standing wave pattern as we change the
location of one or other of the terminals over a few 10s of
wavelengths. Practically signal will vary significantly across
the user premises giving a locations variability around the
terminals. This will in turn require a further margin to yield a
percentage probability of establishing a viable link on installation at the optimum installation location.
Finally, assuming high multi-path, there will be delay spread.
Depending on the nature of this, our received Bit Error Rate
will degrade through inter-symbol interference. Whilst a
propagation and planning issue, this is beyond the scope of
this paper.
Page 7 of 8
Page Page
7 of 82 of 8
on
modelling
methods
visit
ATDI Ltd
Kingsland Court - Three Bridges Road
Crawley - West Sussex - RH10 1HL - UK
Tel. +44 (0)1293 522052 / Fax. +44 (0)1293 522521
World wide web page: http://www.atdi.co.uk
E-mail: enquiries@atdi.co.uk
Advanced Topographic Development and Images Limited
Company Registration Number 3198786
Solutions in Radiocommunications