E2E3: LTE I/Q Post Processing

1) LTE I/Q Samples Dataset Experiment Scenario

• Experiment site: Lake Wheeler Field Lab (LWFL)

• BS tower setup (transmitter): tower height - 10 m, transmit power - 10 dBm, carrier frequency - 3,51 GHz, bandwidth - 1.4 MHz, antenna - dipole (RM-WB1)

• UAV setup (receiver): UAV height: 30 m, 50 m, 70 m, 90 m, 110 m, antenna - dipole (SA-1400-5900)

• Description

A fixed BS tower with a single dipole transmit antenna is deployed, and the flying UAV is equipped with a vertically oriented single dipole receiver antenna and a GPS receiver. LTE eNB is realized using the srsRAN open-source SDR software at the BS tower that transmits the LTE downlink signal. An SDR is mounted on the UAV to collect raw I/Q data samples. For both the tower and the UAV, a USRP B205mini from NI is used as an SDR. The UAV collects 20 ms segments of data out of every 100 ms. Out of that recording, extracts a full LTE frame that is of length 10 ms.

2) Post-processing Code for LTE I/Q Sample Dataset

• Post-processing code link: [download]

• Description

Run MATLAB 'processLTErecording_GPSlog.m' file with a proper file directory setup for LTE I/Q recordings mat files and a GPS logs csv file. The simulation output provides RSRP and RSRQ by using MATLAB LTE Toolbox with the corresponding GPS (longitude, latitude, altitude) and timestamps of the drone.

3) Representative Plots by LTE I/Q Sample Dataset

• Description

Run 'post_processing_plot_main.m' file after obtaining the output file from above Section 2). You can generate representative plots below.

• UAV altitude, speed, distance

You can generate altitude, UAV speed, and distance between the BS tower and the drone during a flight by GPS information.

• RSRP and UAV trajectory

 You can generate time vs RSRP and distance vs RSRP with fitting results by path loss models. In the 3D trajectory of the UAV, RSRP is mapped by the Jet colormap.

• Shadowing and correlation

 You can plot a distribution of a shadowing component by probability density function (PDF) with fitting by distribution functions.  The spatial correlation by horizontal distance can be generated and you can fit the correlation with fitting functions.  

• Resource grid and estimated channel

 You can also plot an LTE resource grid and an estimated channel by CRS while decoding I/Q samples.  Unlike other plots above, these results can be generated by 'processLTErecording_GPSlog.m' file. To plot them, you need to turn on the codes that generate those plots.