C-Free-Uniform: A Map-Conditioned Trajectory Sampler for Model Predictive Path Integral Control

Yukang
Cao
Rahul
Moorthy
O. Goktug
Poyrazoglu
Volkan
Isler
Robotics, Sensing and Networks Lab
University of Minnesota and The University of Texas at Austin
Code Hub Series Hub

Abstract

Trajectory sampling is a key component of sampling-based control mechanisms. Trajectory samplers rely on control input samplers, which generate control inputs u from a distribution p(u|x) where x is the current state. We introduce the notion of Free Configuration Space Uniformity (C-Free-Uniform for short) which has two key features: (i) the generated control input can be used to uniformly sample the free configuration space, and (ii) in contrast to previously introduced trajectory sampling mechanisms where the distribution p(u|x) is independent of the environment, C-Free-Uniform is explicitly conditioned on the current local map. Next, we integrate this sampler into a new Model Predictive Path Integral (MPPI) Controller, CFU-MPPI. Experiments show that CFU-MPPI outperforms existing methods in terms of success rate in challenging navigation tasks in cluttered polygonal environments while requiring a much smaller sampling budget.

Comparison of C-Free-Uniform, Neural C-Uniform, MPPI, and Log-MPPI in a cluttered environment

Slide-style overview figure: C-Free-Uniform keeps broad multimodal exploration while concentrating samples in the free space.

High-Speed Outdoor Demonstration: 2.5m/s with No Global Guidance

Navigating up to 2.5 m/s with no global guidance, a limited 3-meter lookahead, and 512 sampling budget.

Overview Video

BibTeX

@article{cao2025c,
  title={C-Free-Uniform: A Map-Conditioned Trajectory Sampler for Model Predictive Path Integral Control},
  author={Cao, Yukang and Moorthy, Rahul and Poyrazoglu, O Goktug and Isler, Volkan},
  journal={arXiv preprint arXiv:2510.16905},
  year={2025}
}