Tasks remaining

[Preston-Walker]

HW for Rice

1. Find the values for the radio. Display all radio settings.
2. Making dipole antenna for 1485 Mhz, with a 50 Ohm characteristic impedance. Antenna will be either a half wavelength or a quarter wavelength

Calculated Values for Dipole Antennas:
We need 4 of them: 1 transmit, 2 receive, 1 spectrum analyzer
Full Wavelength = 0.20202020m or 20.202020cm which is about 7.95in
Half Wavelength = 0.1010101010m or 10.101010cm or 3.9767in
We are going to round up and make our antenna 8 inches because a dipole is supposed to have a bit extra length.
This comes from lambda = c/f (where c is the speed of light and f is 1485 megahertz))

[Gustavo-Oliveira]

June 9th, Hallway Tests to determine difference between channels and tx/rx and rx ports.
Test 1)
Channel A: RX-vertical
Channel B: RX-vertical
Results:
Ran it twice
Similar sinusoids, with similar amplitudes.
Similar amplitudes on Fourier, difference less than .5 (Not in DB)

Test 2)
Channel A: RX-vertical
Channel B: tx/rx-vertical
Results:
Big difference in sinusoids.
Fourier is complete noise Channel B tx/rx.
The RX port is flashing even though it is plugged into tx/rx port in channel B

Test 3)
Channel A: tx/rx-vertical
Channel B: tx/rx-vertical
Results:
Rx ports were flashing. Not the tx/rx ports.
A lot of noise on the sinusoids and the Fourier. You can't tell what the signal is. amplitude for both was around .20 compared to the 5 amplitude of the Rx ports.

Test 4)
Channel A: tx/rx-vertical
Channel B: rx-vertical
Results:
Channel has a sinusoid with a lot less amplitude. Its fourier is almost all noise. Channel B is working fine.

Summary:
When transmitting with two antennas have them both in RX because it will not register with a tx/rx.

[Gustavo-Oliveira]

June 9th Hallway tests to determine which ports work in combination with each other part 2
Test 5)
Channel A: tx/rx-vertical
Channel B: rx-vertical
Results:
Using a different command in the python API we were able to recognize slight improvement in the sinusoids. Also, there is a big difference with fourier. Instead of Channel B being sheer noise it is now half the amplitude of channel A's fourier.

Test 6)
Channel A: rx-vertical
Channel B: rx-horizontal
Results
Half amplitude on channel B. We were hoping for it to be close to 0.
Second result was a third amplitude

Test 7)
Channel A: rx-horizontal
Channel B: rx-vertical
Results
Actually worked as intended.
first result was 8 to .5 difference on the fourier
second result was 6 to .5 difference
third result was

[rskirkwood]

Take data in hallway, all the way across the hallway
Be in the middle of the hallway and have the receiver off of the floor
Do the test at night, when nobody is around or going to be around (try to minimize all changing variables)
Move the position of either the transmitter to get a "good" bad multipath channel,
Get two sets of data for every 10 degrees from 0 to 90 degrees (both with and without the hybrid coupler)

tx instructions:
Use Matlab (the while 1 loop is fine)
use a ts of 10e6

[Preston-Walker]

Questions to answer:

• Are our FFT charts using 512 for their Nfft?
• Why are they not coming across clearly even when we do cable tests? (They should look like a square wave in freq domain)
• Why does the signal strength taper off around sample 140?
• What can we do to fix the tapering behavior?
• Make thee csv reader better (support for nonexistent real/imaginary components)

[Gustavo-Oliveira]

Pruning
Get the energy of the signal
starting with bigger ones andworking to small
keep those that give 99 percent of total energy
maintain temporal seperation
one line 37 h = fftshift (ifftH
what is inverse abs of H

Save two cahnnels into 1 .mat file and call them H1 and H2 or Left and Right
Both H1 and H2 have to be the same length
add 0's to the small one to align to the big one

Name of the .mat file should have whether it is vertical/horizontal and rotation of transmitter
inside that file we have h1 and h2

csv read

When reading in both channels and computing both H's
Check to see if they line up

[Gustavo-Oliveira]

Plot fft

1. Raw
2. Initial Pruned
3. Additional Pruned

Have them all on same figure different plots. if you see a dramatic difference in frequency domain between second and third. Send Dr Rice an email. He will determine