RF Sensor Geolocation Software

This page explains the step by step process that was followed to post-process the measurements collected using Keysight RF sensors and the RF Sensor Geolocation Software. For a detailed case study using this software, refer to Case Study: Tracking Drones, and some further details can be found in the following workshop paper:

• U. Bhattacherjee, E. Ozturk, O. Ozdemir, I. Guvenc, M. L. Sichitiu, and H. Dai, "Experimental Study of Outdoor UAV Localization and Tracking using Passive RF Sensing", in Proc. ACM WiNTECH Workshop, Oct. 2021.

An example of data-processing for Inspiron Drone is explained there in detail. The location comparison between the collected and GPS data has been plotted. For easy interpretation, CDF curves of error are included. You can also find the error with respect to time there. In this document, you can learn about the tools used in the processing, how to observe the measurements closely, how to export those, and how to post-process. There are other few tools as well which can be used to explore more about the collected data and to modify those. Those tools are discussed in short at the end of this document.

Data Processing

1. Starting with Geo-location Software (Sensor Management Tool)

Suppose, you have the data collected in any field experiment saved in your computer in a zip file. Sensor Management Tool (SMT) or Geo-location software can restore those data and help you to interpret those with a nice user interface. After opening the tool, load the collected data using the below options:

Tools (Upper left options) DatabaseRestore Database (Choose you file and open)

You can manage sensors from the tab Geolocation Sensors. For better location estimation, three sensors should be used. Also, if the UAV location is out of the triangle made by three sensors, that particular location may not be detected. Hence, you may need to select or deselect sensors for better estimation. You can manage that from this tab.

The historic Measurement tab lists all the data collected and loaded from the zip file on your computer. Each measurement shows the time, frequency, and correlation plots in the right-hand tab. You can select a few measurements (not all) by pressing Shift + Ctrl + down arrow (Shift + Ctrl + ). Press “Display” now to estimate the locations based on only those data. P.S. Keysight has three different estimation algorithms - TDOA, RSS, and hybrid. Among those TDOA performs the best estimation.

This tool is beneficial to select useful measurements and discard redundant data. Since the Keysight RF sensors can detect any RF signals, it may capture some nearby RF signals which are redundant for your experiment. For example, in Dix Park experiment, there were some nearby WiFi signals and UAV controller signals were detected sometimes.

There are multiple ways of separating data. In Dix park experiment, the controller had a different time and frequency plots than those of the UAV. There was a sudden change in power for controller signals. The other way of differentiating data is to look into the timestamps of each measurement. One can also get insightful information from the center frequency.

2. Exporting estimated location in Google Earth

The tool named Keysight N6854 to KML helps a user to export all the estimated data and project it into Google Earth (Install ‘Google Earth Pro’ free from the internet). After selecting the useful measurements from SMT, mention those IDs in the N6854 to KML tool. You can enable only the required sensors which you want to use for location estimation calculation. You can also adjust some parameters like algorithm, Tentagram size, and resolution, etc. To see the estimated path of the UAV in Google Earth, enable the ‘Estimated position path’.

After selecting all the parameters, hit ‘Refresh’ and wait. If Google Earth Pro is already installed on your computer, it will automatically open Google Earth and show the results. In Google Earth, you can adjust your view. For my experiment, I wanted to check the path what the UAV followed. One can also import GPS data into Google Earth and compare the estimated path with the path from GPS data.

When you have a large number of data, the plot in Google Earth become a little congested. You can adjust the number of plots from the bar in the left upper corner. One can also see the animation of estimated positions of the UAV with respect to time. The exact location of the UAV can be found in the bottom right of the screen in Longitude and Latitude form. These values are in degree decimal and you can adjust that from ToolsOptions.

3. Exporting estimated location in CSV file

The measured data can also be exported into a CSV file for post-processing with the help of N6854 Export Tool. The features of this tool are almost the same as those of N6854 to KML tool. Mention the measurement IDs, algorithm, sensors, and other required parameters and then hit ‘Export’. It will ask you to mention the path of the file. When you are done with that hit ‘Enter’.

The useful parameters in the CSV file are:

1. Measurement ID

2. Center Frequency

3. Latitude

4. Longitude

5. RHO (Correlation coefficient)

6. CEP (Circular Error Probability)

7. Timestamp

The Latitude and Longitudes are in degree decimal format. For plotting it in cartesian coordinate (for better intuitions), a function is used to convert the Latitude and Longitude into X-Y coordinate. P.S. The function accepts radian inputs. The timestamp is in UTC time and you can adjust the time according to your requirement.