Reproduce “PerchRL: Vision-Based Agile Perching on Inclined Platforms under Rapid and Irregular Motion” 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

Zihong Lu, Zongzhuo Liu, Huaxu Li, Yitao Zeng, Jinqiang Cui, Jie Mei, Youmin Gong, U Kei Cheang and 1 more

PerchRL trains a quadrotor to perch on moving, tilted platforms despite intermittent visual measurements. It learns interception and alignment on varied paths, then fine-tunes with an extended Kalman filter estimate, a declining-reliability signal, and rewards for restoring visibility. In Hard-I trials, it succeeded 7 of 10 times on a 70° platform, showing feasibility in this test condition.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.