Denoising Heat-inspired Diffusion with Insulators for Collision Free Motion Planning

TL;DR

This paper introduces a novel diffusion-based motion planning method that generates reachable goals and obstacle-free paths from a single visual input, enhancing robustness and multi-modality handling in robotics.

Contribution

It proposes a collision-avoiding diffusion kernel for training, enabling simultaneous goal generation and collision-free motion planning during inference from visual data.

Findings

Outperforms behavior-cloning and classical diffusion models in robustness.

Effective in multi-modal environments for goal reaching and obstacle avoidance.

Ensures collision-free trajectories in complex scenarios.

Abstract

Diffusion models have risen as a powerful tool in robotics due to their flexibility and multi-modality. While some of these methods effectively address complex problems, they often depend heavily on inference-time obstacle detection and require additional equipment. Addressing these challenges, we present a method that, during inference time, simultaneously generates only reachable goals and plans motions that avoid obstacles, all from a single visual input. Central to our approach is the novel use of a collision-avoiding diffusion kernel for training. Through evaluations against behavior-cloning and classical diffusion models, our framework has proven its robustness. It is particularly effective in multi-modal environments, navigating toward goals and avoiding unreachable ones blocked by obstacles, while ensuring collision avoidance. Project Website: https://sites.google.com/view/denoising-heat-inspired