We are a small team of four who learn engineering by building real projects together. We plan in weekly cycles, review every change before it merges and follow one rule above the others: whoever merges code can explain every line of it.
LidarSAT is GPS-denied drone navigation. GPS can be jammed, faked or blocked. Without it, a drone only knows roughly how it has moved since take-off.
Our approach uses the ground itself. The drone measures the shape of the terrain below it with a LiDAR and finds the same shape on the free national height maps governments publish, which tells it where it is. We compare a classic matching method with a small neural network trained only in simulation, then test both on real flights from our own low-cost drone.
Note
Development is private for now.
| Member | Focus |
|---|---|
| @Louisomeg | Project lead |
| @zaccesss | Repositories and tools, error measurement and the drone build |
| @mannycodes20 | Data pipeline and project documentation |
| @EB-Glitch08 | Simulator and matchers |
We are a closed team and do not take outside contributions while development is private. To reach us, contact any member through their GitHub profile.
