Low-Power Wireless
Radio Measurements, Analysis and Modeling
Introduction
Sensor networks
will be deployed in harsh environments from the
communication perspective, with significant multi-path effects.
In addition, the low power radios typically used in sensor networks do
not have sufficient frequency diversity to be resilient to multi-path
communication. Under these conditions, wireless communication is
known to be unpredictable and has been shown to vary drastically with
small spatial changes and on different time scales. Even though
most sensor network algorithms are designed to be adaptive to the
variations in the communication channel, there are several parameters
that need to be adjusted to the operating conditions in order to
improve performance. Furthermore, the real communication channels
are very difficult to model for the wide range of target environments
and the different type of radios, frequencies, and modulation schemes
in use. Thus, it is difficult to extensively test the algorithms
under development in simulations under realistic conditions.
Given the variability of the communication channel, and the difficulty
to model it accurately, it is essential to get quantitative data that may allow us
to better understand the channel characteristics in the target
deployment area.
Wireless Radio Measurement Tools
SCALE is a measurement tool
that I developed to study wireless communication channels with low
power radios in new environments. It facilitates the
characterization of the most basic communication metric from the
application point of view: packet delivery. The tool enables the
collection of packet delivery statistics using the same specific hardware platform and
in the same environment
intended for deployment. It also provides full parameter
configuration (including total number of packets, packet size,
inter-packet transmission frequency, RF power level, among others) and
a real-time visualization tool. The data gathered by SCALE allow protocol developers and
engineers to better estimate the appropriate density, system parameter
tuning constants, and expected performance of protocols and algorithms
(data capacity, convergence time, latency). The data was
collected in 3 different environments, with 2 different type of radios,
with 6 different power settings, 12 different packet sizes, and 4
different antennae heights. I used up to 16 nodes in our outdoor
experiments and up to 55 nodes in our indoor experiments. I
measured the packet delivery performance of 240 links for the outdoor
experiments and 2970 links for the indoors experiments. The
results of the measurements using SCALE
led to statistical analysis of the data and revealed some interesting
findings that constitute the basis of my thesis work.
Analysis and Modeling
In my thesis work,
I performed statistical analysis on data from a rich
set of links with different distances, directions, antennae elevations
from the ground, with or without line of sight ---conditions that we
expect to find in sensor network deployments---, and I found many
important lessons for network protocols and system design. My
goal was to provide sound foBy using these data I have
developed new radio propagation models usedundations for conclusions
drawn in my thesis by extracting relationships between location (e.g
distance) and communication properties (e.g. reception rate) using
non-parametric statistical techniques. The objective is to
provide a probability density function (PDF) that completely
characterizes the relationship. Furthermore, I studied individual link
properties and group link properties with their correlation with
respect to common transmitters, receivers and geometrical location.
Among the many interesting findings, I found
that there
is no clear
correlation between packet delivery and distance in an area of more
than 50% of the total communication range. In addition, I found
that temporal variations of packet delivery are not correlated with
distance from the transmitter or transmission power level, but to the
mean reception rate of each particular link. I also found that
the percentage of link asymmetries varies from 5% up to 30% in some
cases, and there was no obvious correlation between link asymmetries
and distance and/or transmission power levels. I provided
significant quantitative evidence that supports the commonly held
belief that link asymmetries are due to hardware calibration
differences. Another important observation is that the
distribution of lossy links can greatly affect routing algorithms based
on geometric concepts. For example, all local avoidance
approaches that reduce the routing problem to traversal on Gabriel or
local neighborhood graphs may no longer be applicable in
practice. Another, possibly more impacting ramification is that
no deterministic method can be used to guarantee packet delivery in
stateless routing protocols. This is justified by the small but
non-zero probability of having links with very small or close to zero reception
rate even at very small distances. Another conceptual change is
that there is a strong benefit of observing at least some percentage of
links on-line. This is because some of the most effective links
in terms of metrics of travel distance versus required number of
messages are links that have a reception rate between 40-60%. In
addition, it is perfectly possible to find high reception rate links
that are stable and highly symmetrical that cover medium to long
distances.
The existence of
superior nodes in terms of both transmitters and receivers capabilities
implies that fairness will
become one of the major issues for any routing, multicast, and
broadcasting approach, because all of these protocols have a tendency
to disproportionately use a subset of nodes. The statistically
demonstrated space correlation will also greatly impact the development
of routing protocols, as well as power management techniques. For
example, since nodes are naturally clustered in subsets that
efficiently communicate with each other and poorly with the rest of the
network, it will be important that power management strategies,
simultaneously turn down or up the majority of the nodes in one of such
subsets. Furthermore, clustering techniques might be even more
efficient than in networks modeled with other communication models.
As part of my thesis work, I also developed a series of wireless
network simulation environments that produce networks of an arbitrary
size and under arbitrary deployment rules with realistic communication
properties. For this task I used an iterative improvement-based
optimization procedure to generate instances of the network that are
statistically similar to empirically observed networks. I
evaluated the accuracy of the conclusions drawn using the proposed
model and
therefore comprehensiveness of the considered properties on a set of
standard communication tasks, such as connectivity maintenance and
routing. The simulation models have been fully implemeted in the EmStar simulator.
Sensor networks
will be deployed in harsh environments from the
communication perspective, with significant multi-path effects.
In addition, the low power radios typically used in sensor networks do
not have sufficient frequency diversity to be resilient to multi-path
communication. Under these conditions, wireless communication is
known to be unpredictable and has been shown to vary drastically with
small spatial changes and on different time scales. Even though
most sensor network algorithms are designed to be adaptive to the
variations in the communication channel, there are several parameters
that need to be adjusted to the operating conditions in order to
improve performance. Furthermore, the real communication channels
are very difficult to model for the wide range of target environments
and the different type of radios, frequencies, and modulation schemes
in use. Thus, it is difficult to extensively test the algorithms
under development in simulations under realistic conditions.
Given the variability of the communication channel, and the difficulty
to model it accurately, it is essential to get quantitative data that may allow us
to better understand the channel characteristics in the target
deployment area.
In my thesis work,
I performed statistical analysis on data from a rich
set of links with different distances, directions, antennae elevations
from the ground, with or without line of sight ---conditions that we
expect to find in sensor network deployments---, and I found many
important lessons for network protocols and system design. My
goal was to provide sound foBy using these data I have
developed new radio propagation models usedundations for conclusions
drawn in my thesis by extracting relationships between location (e.g
distance) and communication properties (e.g. reception rate) using
non-parametric statistical techniques. The objective is to
provide a probability density function (PDF) that completely
characterizes the relationship. Furthermore, I studied individual link
properties and group link properties with their correlation with
respect to common transmitters, receivers and geometrical location.
Among the many interesting findings, I found
that there
is no clear
correlation between packet delivery and distance in an area of more
than 50% of the total communication range. In addition, I found
that temporal variations of packet delivery are not correlated with
distance from the transmitter or transmission power level, but to the
mean reception rate of each particular link. I also found that
the percentage of link asymmetries varies from 5% up to 30% in some
cases, and there was no obvious correlation between link asymmetries
and distance and/or transmission power levels. I provided
significant quantitative evidence that supports the commonly held
belief that link asymmetries are due to hardware calibration
differences. Another important observation is that the
distribution of lossy links can greatly affect routing algorithms based
on geometric concepts. For example, all local avoidance
approaches that reduce the routing problem to traversal on Gabriel or
local neighborhood graphs may no longer be applicable in
practice. Another, possibly more impacting ramification is that
no deterministic method can be used to guarantee packet delivery in
stateless routing protocols. This is justified by the small but
non-zero probability of having links with very small or close to zero reception
rate even at very small distances. Another conceptual change is
that there is a strong benefit of observing at least some percentage of
links on-line. This is because some of the most effective links
in terms of metrics of travel distance versus required number of
messages are links that have a reception rate between 40-60%. In
addition, it is perfectly possible to find high reception rate links
that are stable and highly symmetrical that cover medium to long
distances. 