I am a 5th year Ph.D. studying Robotics and Machine Learning student at Princeton University in the Department of Mechanical and Aerospace Engineering.

I am in the Intelligent Robot Motion Lab (IRoM Lab) advised by Anirudha Majumdar. With my research, I work towards guaranteeing safety of robotic agents when deployed in new settings.

Email: afarid@princeton.edu

Research Overview:

I am motivated by potentially non-linear and hybrid robotic systems that are equipped with rich sensors, in dynamic environments, and performing complex tasks. Using learning-based control, I focus on providing guarantees on the performance of such robotic systems when deployed in new environment settings. Specifically, I am developing and using existing tools in generalization theory to provide such guarantees. I concentrate on the out-of-distribution setting, where new environments are qualitatively different from the training environments. Current projects directions include:

  1. Out-of-distribution detection: determining if a robot is operating in an environment for which it is unprepared. With this knowledge, emergency maneuvers or halting motion can prevent crashes and guarantee safety.

  2. Out-of-distribution generalization: providing guarantees in environments which have loose resemblance to the training environments.

  3. Building and setting up hardware platforms on which to validate theoretical and simulated results: vision-based obstacle avoidance with a Parrot Swing drone (see above video), ground-based ball-catching vehicle, and an aerial manipulator (see images below).

Aerial Manipulator

Ball-catching Robot