Code that creates the analyses the data and figures of the article:
Gkanias, E., Mitchell, R., Stankiewicz, J., Khan, S.R., Mitra, S., and Webb, B. (2023). Celestial compass sensor mimics the insect eye for navigation under cloudy and occluded skies. Communications Engineering.
Open a terminal, clone the project, and navigate to its working directory:
git clone https://github.com/InsectRobotics/CelestialCompassAnalysis.git
cd CelestialCompassAnalysis
The processed data can be downloaded from here and they should be placed in the csv directory. Alternatively, they can be generated by following the instructions below.
If you have access to the ROS-bag files produced by the robot (upon request from the authors), proceed with step 1a, otherwise proceed with step 1b.
Copy the sardinia_data and south_africa_data directories
to the working directory, navigate to the templates directory (e.g., cd templates), and run the
script that creates the raw dataset:
python create_csv.py -t raw
The ROS-bag files contain a lot of information, including low resolution videos captured by the fish-eye camera
during the experiments that increased the size of the datasets to almost 100 GB. The above script creates a CSV
file (named raw_dataset.csv) in the csv directory, which includes only the data used in the article
without any further processing and reduces the size of the datasets to 677.7 MB. It also creates a directory in
named sessions in the csv directory, which contains the high-resolution fish-eye images captured at
beginning of each session (400.1 MB).
Navigate to the csv directory and download the raw dataset from the above link. Alternatively, run the below lines from the working directory:
cd csv
wget https://datashare.ed.ac.uk/bitstream/handle/10283/7116/raw_dataset.csv?sequence=9&isAllowed=n
wget https://datashare.ed.ac.uk/bitstream/handle/10283/7116/sessions.zip?sequence=5&isAllowed=n
cs ..
This will download the CSV file (raw_dataset.csv) and fish-eye images (sessions.zip)
in the csv directory. These contain all the important data from the sardinia_data and
south_africa_data directories (which are excluded from the dataset because of their large size).
Finally, extract the sessions.zip in the csv directory.
Providing the raw_dataset.csv as input, should allow the following commands to work without problems.
The option -o [output_path] allows for a different output path.
To create the pooled data from the above CSV file, run:
python create_csv.py -t pooled -i DATASET_DIR/raw_dataset.csv
which will create another CSV file in the csv directory (named pooled_dataset.csv).
The -i DATASET_DIR/raw_dataset.csv part is optional, and if you followed step 1a or 1b, it
shouldn't be necessary.
The errors for all the sessions and models can be calculated by running:
python create_csv.py -t error
This CSV file is saved by default in csv/error_dataset.csv and needs as input the
csv/pooled_dataset.csv file. If an alternative name was chosen for this file, it needs
to be specified through the -i [input_path] as before.
To generate the plots for the article, the above datasets need to be generated first. If
alternatives names or paths were used for the generated dataset files they can be set as
inputs using the -i [input_file] as an option. The plots can be generated in order
of appearance in the article by running:
python plot_csv.py -f 2d -o png
python plot_csv.py -f 3a -o png
python plot_csv.py -f 3b -o png
python plot_csv.py -f 4a -o png
python plot_csv.py -f 4b -o png
python plot_csv.py -f 4c -o png
python plot_csv.py -f 5a -o png
python plot_csv.py -f 5b -o png
python plot_csv.py -f 5c -o png
python plot_csv.py -f 6a -o png
python plot_csv.py -f 6b -o png
python plot_csv.py -f 6c -o png
python plot_csv.py -f 7 -o png
python plot_csv.py -f 9h -o png
python plot_csv.py -f S2a -o png
python plot_csv.py -f S2b -o png
python plot_csv.py -f S2c -o png
python plot_csv.py -f S2d -o png
python plot_csv.py -f S3a -o png
python plot_csv.py -f S3b -o png
python plot_csv.py -f S4b -o png
python plot_csv.py -f S4c -o png
Note that the option -o [output_file] can be used either to specify the output file path
or the file extension (supported extensions are png, jpeg, jpg, svg,
pdf).
If you have any issues installing or using the package, you can report it here.
The code was written by Evripidis Gkanias and Robert Mitchell.
Copyright © 2023, Insect Robotics Group, Institute of Perception, Action and Behaviour, School of Informatics, the University of Edinburgh.