This package contains all the software and results needed to reproduce the research conducted in Setting it right: Impact of User Settings on the Energy Consumption of Google Maps.
The package uses the included android-runner framework to perform measurements on the android device. This framework must first be setup correctly by following the steps documented in the README.md file in the android-runner directory.
While following the above instructions, the device's power_profile.xml should be extracted. This file needs to be copied to the experiment/ subdirectory also containing the DOffT/ to LOnT/ and Results/ subdirectories. The power_profile file already included there should be replaced with the profile extracted from your own device.
The package expects the following applications to be installed and operable on the device:
- Google Maps version 11.4.3 (released on 28.10.2021)
- Lockito - v3.2.1 (released on 29.09.2020) These applications can be installed directly from the app-store, or manually.
The GPS spoofing application Lockito should contain two, and only two routes.
- The first route should contain a route following some road navigable by car in Google Maps. The route length and traveling speed should be set such that the total drive time is longer than 10 minutes.
- The second route should be a static location set to the starting position of the first route. This location is used by Google Maps to determine the navigation.
Unaltered, the package will try to navigate towards "Veendam" from whichever starting location is configured in Lockito. While it is not strictly required to change this destination, it is recommended to set it according to the first route in Lockito. To change this destination, changes must be made in the file Scripts/after_launch.py on line 71:
device.shell('input text "Veendam"')In this line "Veendam" should be changed to the correct destination. This must be done in each Scripts/after_launch.py file in the subdirectories DOffT/ to LOnT/ in the experiment/ directory.
The package is set to run on a Rasberry Pi, mounted directly from the directory /home/pi/external_hdd/. Any changes made to this location requires changing the locations in the confic.json files found in each subdirectory in the experiment/" directory (save from the Results subdirectory). The systrace_path variable must be set according to the computer used to run the package, and the power_profile path must be set according to the path to the power_profile.xml file in the experiment directory.
Unaltered, this package contains the results reported on in the original experiment. When reproducing these results, the original results should be removed from the experiment/Results subdirectory. All files and subdirectories containing the treatment names such as DOffT or LonS must be removed.
Several steps must be taken to perform the experiments used in the research. The experiment itself is split up into an online and an offline part. This is due to the fact that for running the offline experiments, a set of manual steps are required different from the online experiments.
Use the following commands to start the virtual environment for android-runner (assuming the same file structure is used as discussed above):
cd /home/pi/external_hdd/android-runner
sudo python3 -m venv .venv
source .venv/bin/activate
cd ../You may replace the first cd command to wherever the android-runner package is located. Afterwards, the working directory should have been set to the directory this file is located in. Afterwards, the virtual environment should be running and the working directory should be the one containing this README file.
Before running the online experiments, the device should be freshly booted and both Google Maps and Lockito should be opened and given time to load into their starting screen once. Google maps should be closed afterwards, but Lockito should be kept running on its starting screen. Now the following commands can be run from the working directory this file is located in:
sudo ./runonline.shThis runs all the experiments for all the online treatments.
For running the experiments for the two offline treatments DOffT and LOffT several manual steps are required. To begin with, the device should be freshly booted. In Google Maps, an offline map containing the starting location, the entire Lockito route, and the set destination must be manually selected and downloaded. This needs to only be done once before running the offline experiments. After the map is downloaded, Wifi and Roaming must be turned off on the device to force Google Maps into offline mode. The offline map may now be tested to see if the correct navigation route is selected towards the given destination. If this is not the case, a larger offline map of the region may need to be downloaded.
If the offline map is correctly downloaded and the Wifi and Roaming on the device turned off, the following command can be used to start the offline experiments:
sudo ./runoffline.shAfter completing all the steps in running the experiment, the results folder should be populated with a new set of subdirectories consisting of the name of the treatment followed by the date and time of the experiment (E.g. DOffT_2022.01.16_123456).
The results from each experiment must be extracted before they can be processed in R. In the script gather_total_results.py, the list of directories should be changed to contain only the 6 subdirectories in the experiment/Results directory generated by the experiment previously executed. Afterwards, the script can be executed as follows:
cd experiment/Results
sudo python3 gather_total_results.py
cd ../../This should generate six files named total_ followed by a specific treatment name.
These files can now be processed in R to produce the results reported in our research by loading and running the file datanalysis.R in RStudio.