Volary trains people the way it builds: mission first. The programs form one funnel, from a low-pressure reading group to a student shipping an autonomous UAV payload from zero.
Nobody is expected to start at the deep end. Each stage feeds the next, and dropping in at the right level is the point.
Bi-weekly, 60 minutes, coffee and one idea. The lowest-pressure way in.
A gentle on-ramp for a student with limited background and old hardware.
Eleven sessions from teardown to a flying autonomous payload.
A real question on the platform, and a place on the team.
Organized around two chains, action and perception, and one outcome: robot system integration. Wire those chains together and the same skill transfers to other robots, not just UAVs.
Teardown, Linux, and ROS with RViz. Learn the machine before automating it.
Depth-camera data streams, coordinate transforms, and visual-inertial odometry.
Mapping, trajectory generation, and EGO-Planner running in simulation.
PX4 in software-in-the-loop, then a payload the student designs, prints, and defends.
The material is built around a role, not a person: any student who joins the UAV direction. That keeps it reusable when people rotate in and out.
The 12-week version exists for the hardest case: a motivated student with a limited background and only legacy hardware, an old Jetson Nano and a first-generation depth camera. It has three honest goals, and failing one still counts as a result worth writing down.
Be able to describe the full autonomous-UAV technology chain in plain terms.
Get a minimal EGO-Planner running on an ordinary x86 laptop.
Benchmark what old devices can and cannot do, and record it for the next person.
Six sessions, once a fortnight, sixty minutes each. No prerequisites. It routes straight into the handbook, the first lectures, and the bootcamp, so anyone curious can find the depth that fits them.
Ask about joiningVolary is looking for students who want to build, and for partners who want to run a track at their own university. Both start the same way.