Explore / App / Jev Drone

Jev Drone

Self-hostTerminal

Choose a detour, climb, brake or target reacquisition while code handles flight control.

by admin · listed 7 days ago

⚡ What System1 does here

In the main experiment, ordinary code converts simulated depth and segmentation buffers into a scene. Jev chooses a maneuver with Choice, assesses risk with Score and judges whether the target is genuinely lost with Noul. Flight control, guidance and safety reflexes remain in code and can override its advice.

About

A MuJoCo research demo in which a simulated Skydio X2 follows a ground rover through five obstacle types. It includes flight-control and perception code, an experiment write-up, a course recording and screenshots, plus an incomplete tunnel experiment. Evidence and limits: All evidence is simulated; the model does not consume images. The highlighted success is a single run, and an earlier multi-seed comparison showed no advantage. Different climb options also confound the baseline comparison. The tunnel experiment remains incomplete. Tactical calls are separate from flight control; this does not establish real-drone reliability. This review checked public documentation, code and media; it did not rerun simulations or test physical hardware. Original project images and demonstrations (with attribution):

Simulated drone climbing over a barrier ↗

Simulated drone following a ground rover ↗

Simulated drone passing a sliding gate ↗

Jev Drone simulated course demonstration ↗

Sources: https://github.com/RomanSlack/jev-drone/blob/main/README.md https://github.com/RomanSlack/jev-drone/blob/main/tactics.py https://github.com/RomanSlack/jev-drone/blob/main/site/app/page.tsx https://github.com/RomanSlack/jev-drone/blob/main/site/lib.ts https://jev-drone.vercel.app/

🧪 Community test reports

0 reports · 0 passed on the real model

No one has tested this yet. Be the first.

💬 Discussions

New post →

No discussions yet.