In free time, created pitch to director of engineering for the use of ML in company projects.
Evaluated applications with AI group and created a detailed literature review of current state of the art
Created presentations to be used in pitch
Deep Learning for Canadarm2 Mission Analysis
Created the first deep learning project to accelerate Canadarm2 mission planning.
Problem Tackled
Current worst-case mission planning operations require running many time-consuming, simulations of Canadarm2 dynamics with various payloads to identify suitable speeds to maneuver the robotic arm
Results of Successful Project
Models predicts accurate estimates of the optimal rate to move the arm, given mass properties and arm commands.
Predicted reduction in mission analysis time from 2-3 weeks down to 1-2 days.
This is the first project to use deep learning models to aid in mission planning, and will also be a demonstration to show the potential of using deep learning methods in future programs
Dragonfly NASA Tipping Point Project
Worked with mechanical, electrical, controls, and software engineers on the end effector of the robotic arm for the NASA tipping point Dragonfly project, an on-orbit satellite self-assembly project.
Communicated to setup electrical testbed and wrote scripts to test if the end effector met the mechanical requirements.
Proposed the methods learned to project manager, for a similar, separate project.