For a better view of my poster while you watch, open this link: https://sites.psu.edu/mcreu2020/files/formidable/2/Dalinas-Official-MCREU.pdf
To return to my MCREU project page and submit an evaluation for my project: https://sites.psu.edu/mcreu2020/2020/07/28/autonomous-light-seeking-robotic-system-inspired-by-dynamic-insect-vision/
Be sure to turn up your volume!
Dalina Thuy
Mi Tu,
Campus:
Harrisburg, University Park,
Anticipated
Graduation: May 2022,
Mentors:
Seth Wolpert (Harrisburg), Bo Cheng (University Park),
Project
Title: Autonomous Light-Seeking Robotic System Inspired by Dynamic Insect
Vision
ABSTRACT-
The project constructed an autonomous, mobile, light-seeking
robot with a vision method inspired by insects. The robot had to be capable of
adapting in unstructured environments, differentiating light intensities,
steering to the highest luminance areas, and continuously following an
erratically moving light source. These abilities were each measured through
various trials of courses. Light tracking robotics proves useful to
applications like farming, solar light tracking, and firefighting systems.
Vehicles that detect and track motion have applications like surveillance and
search and rescue. Vision is an essential sense for robots to understand and
interact with the world. Emulating insects could help advance robots, which
help people globally: elevating manufacturers, pioneering space exploration,
saving lives, and cleaning family homes. Humans have the spatial perception to
effortlessly analyze environments, things, and motion, but forming such
aptitudes in robots remains a work-in-progress. A literature study was
performed on certain insects owning some visual acuities that are superior to
humans, although having smaller, simpler anatomy. Interest was in the anatomy
and processing they own to see three-dimensionally and quickly. A comparison
between current visual technology and the structural organization of the
photoreceptors of adult insects was performed. The learned information was to
be inspirational. At the end of the project, the robot completed all its
planned tasks. Also, plentiful knowledge was learned in electrical engineering.
However, with the constraint of time and resources, satisfactory inclusion of
the insect principles was not achieved. In the next steps, that will be the
priority. Also, the robot will acquire more efficient and powerful mechanical
and electronic parts. Further in the future, the capacity for deep learning AI
may be added, to dive into the neural network of insects. This approach
supports the application of biological mechanics to advance technology and
supports robotic vision itself.
Acknowledgments
This research was funded by the
Penn State Multi-Campus Research Experience for Undergraduates (MC REU)
program. I would like to thank Penn State, Erin Hostetler and Cindy Reed for
this opportunity. I am grateful to Dr. Seth Wolpert and Dr. Bo Cheng for their
mentorship, flexibility, patience, and wisdom.