I contracted to NASA for 18 months under both Oceaneering and Rob Platt Ph.D at Northeastern. During which I designed and implemented Valkyrie's perception system. I primarily focused on the robot's sense of pose, which was accomplished by sensor fusion of Inertial Measurement Units via Extended Kalman Filter. I also worked extensively on sensor selection, as well as data interpretation. Specifically, we developed a novel algorithm for the detection of grasp affordances, as well as point cloud compression algorithm based on SPIHT. Lastly, I developed tasks for the robot in simulation, largely in Gazebo, although a small amount of time was spent in V-REP. If you'd like to learn more about the robot feel free to read this article by IEEE Spectrum.