Experimental results for the global/local sensor fusion algorithm using numerical P systems and ROS

Published: 12-11-2018| Version 1 | DOI: 10.17632/2dhtz6y22p.1
Andrei Florea,
Catalin Buiu


This record presents experimental data that was recorded from ROS and processed using the statistical language R. The objective of the experiment was for a nano-quadcopter to detect the height of a previously unknown waypoint (Apriltag) and afterwards maintain a height as close as possible to that tag. Due to the lag of the fiducial tag detection algorithm, this algorithm uses the tag detection only at the beginning, for determining a waypoint and afterwards uses the local height sensor available onboard the quadcopter. The PDF files all contain plots that are produced by the attached R script (stats.r) by processing the three CSV files included in this record.


Steps to reproduce

All of the required software mentioned bellow is free and open-source. The statistical software R [1] is required In order to re-process the experiment data. After installing R, simply start the R executable from the command line, also specifying stats.r as a parameter. The script will output the four PDF files but can also be used to manually inspect specific values or plot them in a custom plot. The actual experiments were performed using the PeP numerical P system simulator [2] and the associated PeP robot controller for ROS [3]. The tags used for position, rotation and identification recording of each robot are AprilTags [4]. The crazyflie ROS driver is available at [5]. [1] https://www.r-project.org/ [2] https://github.com/andrei91ro/pep [3] https://github.com/andrei91ro/pep_controller [4] https://april.eecs.umich.edu/software/apriltag.html [5] https://github.com/whoenig/crazyflie_ros