Projects from the Udacity Robotics Software Engineer Nanodegree (2017), a two-term program covering perception, localization, path planning, control, computer vision, and deep learning for robotics through ROS.

  1. Follow Me — Trained a fully convolutional network (SegNet) and deployed it to a drone in Unity to follow a specific person in a crowd.

  2. Home Service Robot — Integrated autonomy modules to navigate a robot to predefined destinations in a known map.

  3. Search and Sample Return — Processed a rover camera feed to threshold ground from sky, accumulated frames, and transformed them into a navigable terrain map.

  4. Robotic Arm Deep Reinforcement Learning — Trained a convolutional DQN agent to control a two-joint robotic arm to reach a target object.