Reproduce “JEPTERA: A Platform Agnostic JEPA-Assisted Model Predictive Planner for Urban UAV Navigation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Gibran Fathoni Belva, Mansur Arief, Ahmad Alfan Alfian Irfan
JEPTERA combines a learned predictor of UAV motion with model-predictive control, safety checks, and PID execution for simulated landing navigation. LiDAR separately tracks obstacles; the predictor estimates responses to velocity commands. Across three scenarios, the complete system with safety reached and landed without collision in 28 of 30 runs, versus 27 of 30 for MPC–PID with safety. Tests were simulated.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.