- Ability to program in Python okayishly
- Finish what you take on (don't just start without finishing, you will thank yourself later)
- Don't give up, the world is grateful for what you have to bring to us with your newfound robotics knowledge. Don't give up.
- Machine Learning (Stanford CS229) ML is used a lot because machines can learn to perform tasks better than how we programmers can instruct them to
- Computer Vision (Carnegie Mellon 16-385) Most perception for robots revolves around computer vision, this is crash course to that
- Manipulation Estimation and Control (Carnegie Mellon (MEC)) Robots need to need to move, know where they are, and control themselves to keep up the good "motion", this is foundational to what comes next
- Robot Autonomy (Carnegie Mellon) Delves into all core aspects of robot autonomy for strong foundations in grasping, planning, mapping, decision-making etc.
- Convolutional Neural Networks for Visual Recognition (Stanford CS231n) Best course on modern day neural net based perception
- Statistical Methods in Robotics (Carnegie Mellon) Best introductory course on making robust intelligent robots that have to deal with uncertainty
- SLAM (Carnegie Mellon) Necessary to instil robots with reliable ability to localize themselves, and map their surroundings in an actionable way
- Geometric Computer Vision (Carnegie Mellon) Deals with projections, homography, epipolar geometry, stereoscopic vision, 3D reconstruction, Calibration
- Convex Optimization (Carnegie Mellon) Used to optimize motion of robots like Spot and Atlas from Boston Dynamics to give them smooth, fluid, steady, efficient gaits
- Deep Reinforcement Learning and Control (Carnegie Mellon) The future. That doesn't quite work yet. But worth studying nonetheless