Content pfp
Content
@
0 reply
0 recast
0 reaction
July pfp
July
@july
My favorite thing about robotics is how many disciplines it spans: - EE, power electronics - Custom PCB design/layout - Mechanical engineering, Mechatronics - If you're in aerospace, that's its own discipline - Same with Space systems - Systems engineering actually becomes important - Software: embedded software systems / design - Robotics algos: SLAM, motion planning, trajectory planning etc - Sensor implementations / understanding: sensor validation, bring up, etc - Real-time (hard / soft real time software systems) - EKF/UKF: Kalman filters, and filters for handling noise - OGFAI: search algos - Perception: edge perception models - Networking: routers, traffic handling, ip address assignement locally for multiple robots - Cloud infra: process sensor data, upload to cloud, explode the sensor data to make sense of what robots did - GCS: some sort of Ground Control Station (GCS) UI to actually understand what telemetry data they are doing in near real time - The new AI stuff: see below (ran out of room)
3 replies
9 recasts
49 reactions
July pfp
July
@july
And it keeps going: - SIL: software-in-the-loop testing of all the components - Testing pipeline: HIL (hardware-in-the-loop) testing - Simulation setup: vehicle dynamics, or robot dynamics modeling of the world - A DevOps / MLOps pipeline & infra that can handle all of the above specifically for robotics - Sensor fusion: deserves it own category - Deep learning / ML models running on the edge: - Safety and Reliability Engineering: esp for aerospace or autonomous vehicles, building systems with fail-safes, redundancy, and robustness in general is probably the longest - the new AI stuff is abundant, and least proven what the best path forward is, but the one thing that comes to mind is end-to-end multi-modal transformer models, that can do sim-to-real well. also deep RL, hiearchical RL is also interesting for safety critical systems - the list keeps going
1 reply
0 recast
20 reactions
Shaw pfp
Shaw
@shaws
WONDERFUL
0 reply
0 recast
0 reaction