Transcript Final Presentation

Final presentation
Cellular Device Detection
Instructor :
Yossi Hipsh
Performed by : Ari Avitzur
Asher Pilai
Winter 2008/9
Semesterial Project
Project Objectives
Detecting a cellular device location in a given
room by its signal , using a phased-array
antenna.
• Our individual goal in this project was to
check whether the system's behavior is close
enough to theory - if not, then to do an
adjustments and to solve other problems in
the system, like reflections.
Inputs – general specifications
• 2 Multi-beam antenna at each side of the room for
full coverage.
• The antennas and the cell-phone are placed 1.14m
high from the floor in the room. The cell-phone is on
a shelf, with 1m length.
• The distance between the antennas and
cell-phone is 2m.
• We need to detect whether the cell-phone is on the
right side of the shelf, or on the left side. In other
word – our detection resolution is 0.5m .
System description
Amp
detector
Amp
detector
Antenna A
Cellphone
ADC
Amp
detector
Antenna B
Amp
detector
Computing
&
Display
Unit
Detection theory
• The value of
E
as function of X dimension is:
Detection theory(continue)
• To describe E as a function of the angle Θ , we
use Forier Transform:
rect. Forier  SINC
• In order to describe power as function of Θ ,
we use SINC .
2
Detection theory(continue)
9dB(deg ree)  ???
•Where:
2   3dB 
51*  51*0.17

 19.83
D
0.437
Detection theory(continue)
• The antenna is
designed that each
beam’s maximum
match the other
beam’s -9dB point.
• We need to
determine  in
order to get the right
graph.
The Antenna
Reflection problem and the
parabolic contour
• The parabola of the contour is determent that the
ratio between the a factor and the focal length
between the antenna and the parabola base would
1
a

be:
4f
The parabolic contour causes all the
waves that are parallel
to the antenna’s board to
be centered to the focus,
which is the antenna’s dipole.
The detection algorithm
• Calibrating the system: finding the gain of the
antenna Gr , for a given R . We will build a graph of
Gr as function of the angle θ.
• Measuring the power received by the antenna
measuring each beam separately.
• For two antennas – deciding where is the cell –
phone, by the antenna which both beams are not
less than -9dB than its maximum.
For one antenna – we need to find the ratio between
both beam’s power.
The detection algorithm
• Note: Finding the ratio between both beams' power
is intended for overcoming the dependence on the
distance R between the cell phone and the antenna.
• Finding the matching angle for the transmitted
power by using the calibration from step 1.
• We locate the cell-phone, using the θ angle from the
antenna.
Amplifier Measurements
• Amplifier – we have characterized the behavior in a
“Output Voltage Vs. Input voltage” graph:
amplifier characterize graph
Vout (effective) in Volts
2.5
2
1.5
1
0.5
0
0
0.2
0.4
0.6
0.8
Vin (effective) in Volts
1
• We don’t need to use the amplifier for noise
considerations - because the lowest power that
received in the antenna is much higher than the
thermal noise.
Power detector Measurements
• Power detector – we have characterized the behavior
in a “Output Voltage Vs. Input power” graph:
Power Detector
2.5
2
Vout
1.5
y = -0.022x + 1.4338
R² = 0.9958
1
0.5
Power in (dBm)
-30
-20
-10
0
0
• we haven't used the power detector because
we read the results directly from the scoop. Power
detector is necessary if the signal isn’t high enough,
or if we want to use the computing unit.
Antenna Measurements
• Antenna – we have characterized the behavior
in a “angle Vs. power” graph:
Power graph
power [dBm]
8
6
4
2
angle [degrees]
-40
right beam
0
-20
-2
-4
-6
0
20
40
left beam
Antenna Measurements
• We can see that the antenna behavior isn’t
like the theory. We would phase-shift (   13)
and amplify(+2dB) the right beam :
new
Fixed Power graph
power [dBm]
8
6
4
right beam
2
angle [degrees]
left beam
0
-60
-40
-20
0
-2
-4
20
40
Conclusions
• It is necessary to phase-shift and to amplify
the output, as shown in the previous slide.
• The phase-shift would perform with
transmission line . We would use the amplifier
for the right beam, but not for noise
considerations.
• With those adjustments above – the detection
is possible, even with one MBA.
Continued research
• As much as the angular gain is close to our
theory, we can satisfy with one antenna for
the detection or in contrary increase the
detection resolution.
• Detection in a higher dimension (2D, 3D).
• Considering more realistic noise model.
We would like to thank all those who helped
and a special thanks to Yossi Hipsh.