Adriano Garcia

I'm a fifth and final year Computer Science PhD candidate at Binghamton University, where I lead a team of graduate students under the supervision of professor Kanad Ghose. We are developing machine learning, deep learning, and computer vision perception algorithms to enable autonomous navigation of quadrotor drones and ground vehicle platforms.

Currently expanding the idea of CNN-based object detection to incorporate structural object detection using a YOLOv3 object detector. The general idea is to detect and locate navigational markers such as walls, doors, intersections, poster boards, etc. in order to extrapolate localization points that permit a quadrotor drone to localize itself and successfully navigate autonomously through indoor environments. Please see following video:

Throughout my research years I have worked on a variety of ground and aerial robotic projects. Some of these include:

• CNN Classification: A VGG-16 CNN classifier was trained to detect upcoming dead-ends and intersections.

  




• Indoor Drone Navigation with Dead-end Detection: Image line data extracted from the environment is used to detect front-facing windows, doors, and walls. Control mechanism appropriately stops and turns the drone before a frontal collision occurs.

  




• Outdoor quadrotor drone construction: Built two quadrotor drones for heavy-payload outdoor flights using the DJI Flamewheel 450 kit. This involved soldering motor controllers, power distribution boards, mounting and configuring on-board gyro-stabilized camera, mounting, wiring, and soldering DJI NAZA and Pixhawk Mini controller for manually remote controlled and GPS-based autonomous outdoor flights. (Click picture below for test flight video)

  




• Ground Explorer Drone Follower: An ER1 Robot Ground vehicle was programmed to follow a drone indoors. Template matching was used to match and track the drone.

  

Email: agarcia3@binghamton.edu

Linkedin: https://www.linkedin.com/in/adrianogarciacv

ResearchGate: Adriano Garcia on ResearchGate