This directory contains the scripts that were used to evaluate the localization and the coverage methods that have been developed. It requires the data to be available in .csv format. In order to convert ROS bag files to .csv, you need to run for every topic that exists in the bag, the following:
rostopic echo -b filename.bag -p /topic/name > topicname.csv
Run the script with
python evaluation_localization.py <true_pose>.csv <amcl_pose>.csv
The first .csv file contains a recording of the /ground_truth/state (nav_msgs/Odometry) topic and the second a recording of /amcl_pose (geometry_msgs/PoseStamped). It calculated the Mean, Median, Min, Max, SSE, STD, RMSE of the localization error. Also, plots the true and the estimated trajectory of the drone with regards to time.
The iterator.sh file is running the above script for every .csv file that contains the essential files in the following paths: /type/speed/world/attempt, where:
typecan be meander, spiral or linespeedcan be slow, fast or normalworldcan be box, corridor or warehouseattemptcan be 1 up to 5
Run the script with
python evaluate_coverage.py <volume>.csv
The .csv file contains the recording of the /octomap_covered/volume (drone_gazebo/Float64Stamped) topic. The volume of the total environment is considered known and is defined as a variable inside the code. The output is a plot of the percentage covered with regards to time.
The compare_coverage.py can be used to compare the percentage covered when 2 different recordings are available.
Please report bugs and request features using the Issue Tracker.