Home
PRGFlyt is an AI based indoor quadrotor autonomy framework for accomplishing various navigation and interaction tasks designed for both research and pedagogical purposes. It is developed by the Perception and Robotics Group at the University of Maryland, College Park. PRGFlyt leverages the awesome work of the open-source and open-hardware community with the goal of giving back to the community. This framework implements a lot of state-of-the art research papers and also provides both hardware and software setup guidelines for various platforms. Common hardware and software issues are also presented so that it can save time for everyone.
- Nitin J. Sanket (nitinsan at terpmail dot umd dot edu), Fourth Year PhD Candidate
- Chahat Deep Singh (chahat at terpmail dot umd dot edu), Second Year PhD Student
All our hardware platforms are named after dog breeds with the names representing the size of the quadrotor. The following tables provides the name and size/weight of the quadrotor. Click on the name for visiting the corresponding page.
| Name | Motor to Motor Length | Max. Weight |
|---|---|---|
| PRGWhippet | 110 - 160 mm | 300 g |
| PRGCorgi | 180 - 210 mm | 700 g |
| PRGHusky | 250 - 360 mm | 1300 g |
| PRGLabrador | 450 - 500 mm | 2000 g |
| PRGMastiff | >= 650 mm | 3500 g |
Note that each quadrotor can be built with different autopilot softwares and each software has it's own advantages and disadvantages. The following table presents the suffix used for the particular autopilot software.
| Autopilot Software | Name Suffix |
|---|---|
| Ardupilot | Alpha or |
| Betaflight | Beta or |
| Parrot Bebop 2 | Gamma or |
If you like our work and use it in any of your projects please cite us as
@misc{prgflyt,
author = {Nitin J. Sanket, Chahat Deep Singh, Cornelia Ferm\"uller, Yiannis Aloimonos},
title = {PRGFlyt: AI based indoor quadrotor autonomy framework for navigation and interaction tasks},
year = {2019},
howpublished = {\url{https://github.com/prgumd/PRGFlyt/wiki}}
}
Citations for each algorithm is given in the particular repository.
PRGFlyt would not have been possible without the awesome open-source and open-hardware community, especially the ardupilot stack, betaflight stack, px4flow, openMV, Jevois camera and bebop_autonomy.