Indoor mobile classification and coverage analysis
Abstract: As a broader market of high bitrate services is introduced within the mobile
radio network, an increasing fraction of usage will be indoor. This
drastically changes the demands on the network infrastructure, which will
need to handle both indoor and outdoor users at higher bitrates. Since
indoor users in general experience worse signal conditions than clients
located outdoors, differences between usage in typical indoor (obstructed
by building material) and outdoor (unobstructed clear sky view)
environments become highly interesting. Therefore, methods to classify
measurement samples as typical indoor are needed.
A study of present indoor and outdoor signal coverage has been performed.
Data collected by TEMS measurement tools was used as a training set for
discriminant function classification of clients as experiencing typical
indoor or outdoor signal conditions, respectively. It was concluded that the
best variable for classification is RxLev (Received signal level) for cells
within the macro cell layer, normalized with respect to an approximated
client to antenna tower distance.
Using the identified classification procedure, data samples of commercial
traffic obtained by RPMO (Real-time Performance Monitoring) of two different
networks - one within the center of a large city and one much less dense -
were analyzed. In both regions, typical indoor environment usage was found
to represent 40% of the traffic. This excludes in-building locations where
signal conditions are more outdoor-like, such as high structures or
buildings with large sections of glass. The total current use of mobiles
from within buildings in the center of the investigated city was estimated
to be about 60%.
Studying pico (indoor) cell classification results more closely, three
typical groups of building types were identified among the studied data:
Concrete, glass and high buildings (ordered by descending fraction of
typical indoor environment classification). Clients within concrete
structures were classified as experiencing a typical indoor environment,
while clients in high buildings were found to experience signal conditions
very similar to those experienced by a typical outdoor client. Usage
conditions in buildings with large sections of glass in their exterior are
in general in-between the two extremes. Also, micro cells were verified as
handling mainly typical outdoor traffic.
The average improvement in typical indoor environment signal strength
resulting from reinforcement by pico cell installation was found to be 12dB.
It was also predicted that average RxLev in the network, given the indoor
usage of today, could be increased by up to about 4dB by improving coverage
in typical indoor areas by solutions similar to pico cells (from the current
8% up to 100%). Conversely, the current network average RxLev, keeping the
indoor infrastructure at the level of today, would be reduced by 5dB if
typical indoor traffic were to increase from 40% to 70%, which is the
current proportion of in-building pico cell data classified as experiencing
a typical indoor scenario.
It was also concluded that in areas with large cells, a denser macro network
could improve coverage in typical indoor environments more efficiently than
specific indoor solutions.
CLICK HERE TO DOWNLOAD THE WHOLE ESSAY. (in PDF format)