Reproduce “Value Function-Guided MPPI for Hierarchical Planning and Control in Off-Road Autonomous 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
Anoushka Alavilli, Raymond Song, Guanya Shi, Jeff Schneider
The authors introduce a planner using a reinforcement-learning controller’s value estimates to guide route selection. Model predictive path integral planning scores candidate trajectories by goal progress and negative value, without an explicit terrain cost map. In BeamNG, it reached the goal in 55 of 64 cases (86%), versus 35 (55%) for baseline MPPI; this simulation result does not establish physical-vehicle performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.