Vision-Based Indoor Localization of Nano Drone and its Applications
Under Review
S.Singh, F.Pocker, V.Fernandez, K. Arya
Abstract
Indoor localization of aerial vehicles are always a challenging and interesting research problem. There have been many
target recognition and locating technologies in recent years which uses the on-board camera of the drone or motion
capturing system to localize it. We propose here an affordable and accurate vision based indoor localization and
navigation model for Nano drone, in which a monocular camera and WhyCon marker are used for localization. An
overhead camera is used to track the marker on top of the drone. Using a modified ROS (Robot Operating System)
package of marker detection, the markers are detected and their respective positions in 3d space are computed. The
control system approach presented in this work is based on three parallel PID controller on an external control loop
based on the camera feedback to command the velocity of Nano drone in direction of its pitch, roll and throttle. The
applications such as Autonomous Path planning, Landing on a moving platform and Multi drone demonstrates the
feasibility of the proposed method, which opens new possibilities for the autonomous navigation of Nano-drones.
2019
Learning control system design using Nano drone in a PBL focused robotics competition
epiSTEME 8 Inter-national Conference to Review Research on Science, Technology and Mathematics Education; 111-120
F.Pocker, R.Madan, K.Arya
Abstract
Teaching advanced conceptual knowledge and practical skills in a hands-on-manner to a
large number of students is a challenge. The e-Yantra project hosted in IIT-Bombay, through
it's e-Yantra Robotics Competition(eYRC) for college students, teaches these skills scalably.
Participation is free and hardware is shipped to participants who are mentored constantly
throughout the competition. The 7th edition of the competition, eYRC-2018, had a theme (a
gamified problem statement), called “Hungry Bird,” that taught Marker-Based Localization,
Path planning using OMPL, and Waypoint Navigation using PID on a nano-drone using
open source platforms such as ROS and V-REP. This paper outlines how we optimally
designed and deployed these concepts as a series of tasks which eventually helped us to
quantify the learning outcomes among students. 832 students were assigned this theme, and
to achieve scale most of the tasks were automatically evaluated. Finally we illustrated how
we have achieved the effectiveness of the theme with task results and participant’s feedback.
This study and its outcomes are beneficial for academicians seeking to teach advanced
engineering skills at a large scale.
