25-01-2012, 03:51 PM
See-Through Walls: Motion Tracking Using Variance-Based Radio Tomography Networks
[attachment=16600]
INTRODUCTION
THIS paper explores a method for tracking the location of a
person or object behind walls, without the need for an
electronic device to be attached to the target. The technology
is an extension of “radio tomographic imaging” [1], which is
so-called because of its analogy to medical tomographic
imaging methods. We call this extension variance-based radio
tomographic imaging (VRTI), since it uses the signal strength
variance caused by moving objects within a wireless
network.
RELATED RESEARCH
Previous work shows that changes in link path losses can be
used to accurately estimate an image of the attenuation
field, that is, a spatial plot of attenuation per unit area [1].
Experimental tests show that in an unobstructed area
surrounded by a network of nodes, the estimated image
displayed the positions of people in the area
VARIANCE-BASED RADIO TOMOGRAPHIC
IMAGING
In this section, we introduce and justify a model which
relates motion in spatial voxels to the variance of signal
strength measured on the links of a wireless network. In
particular, we justify the assumption of a linear model
when motion is sparse, and describe the limits on the
validity of such a model.
Measurement Model
The goal of a VRTI system is to use a vector s of RSS
variance measurements on M links in a wireless network to
determine an image vector x that describes the presence of
motion occurring within N voxels of a physical space. We
first describe the image vector, then specifically define RSS,
and discuss RSS variance.
[attachment=16600]
INTRODUCTION
THIS paper explores a method for tracking the location of a
person or object behind walls, without the need for an
electronic device to be attached to the target. The technology
is an extension of “radio tomographic imaging” [1], which is
so-called because of its analogy to medical tomographic
imaging methods. We call this extension variance-based radio
tomographic imaging (VRTI), since it uses the signal strength
variance caused by moving objects within a wireless
network.
RELATED RESEARCH
Previous work shows that changes in link path losses can be
used to accurately estimate an image of the attenuation
field, that is, a spatial plot of attenuation per unit area [1].
Experimental tests show that in an unobstructed area
surrounded by a network of nodes, the estimated image
displayed the positions of people in the area
VARIANCE-BASED RADIO TOMOGRAPHIC
IMAGING
In this section, we introduce and justify a model which
relates motion in spatial voxels to the variance of signal
strength measured on the links of a wireless network. In
particular, we justify the assumption of a linear model
when motion is sparse, and describe the limits on the
validity of such a model.
Measurement Model
The goal of a VRTI system is to use a vector s of RSS
variance measurements on M links in a wireless network to
determine an image vector x that describes the presence of
motion occurring within N voxels of a physical space. We
first describe the image vector, then specifically define RSS,
and discuss RSS variance.