Reproduce “Imitation Learning of Avoidance Modes for Efficient Multimodal Model Predictive Control” 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
Changyu Lee, Byung-Kwan Lee, Kanghoon Yoon
SideNet-MPC replaces exhaustive obstacle-side choices with a learned proposal while retaining trajectory optimization and the expert’s two-candidate selection rule. The network uses pairwise obstacle gaps to propose a passing mode; the optimizer evaluates trajectories and supplies controls. In simulation, it succeeds on 90/100 runs versus 97/100 for the expert. Computation averages 18.6 ms per 0.2 s control period on one thread.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.