Reproduce “Neural Integral Control Barrier Functions for Agile Collision Avoidance” 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
Dasari Raj Vamsi, Vrushabh Zinage, Efstathios Bakolas, Rohan Chandra
The authors learn car dynamics, a nominal controller, and a safety boundary, and filter commands online. The boundary uses a finite library of simulated emergency maneuvers; actuator limits remain hard constraints, while learned safety conditions can be relaxed. On 200 held-out layouts at 4.0 m/s, filtering MPPI reduced unsafe episodes from 12 to zero, with success rates of 84.5% versus 84.0%.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.