Rapid Exploration for Open-World Navigation with Latent Goal Models

Oral Talk at Conference on Robot Learning (CoRL) 2021

London, UK


Oral Talk at Workshop on Never-Ending Reinforcement Learning (NERL) at ICLR 2021

Abstract

We describe a robotic learning system for autonomous exploration and navigation in diverse, open-world environments. At the core of our method is a learned latent variable model of distances and actions, along with a non-parametric topological memory. We use an information bottleneck to regularize the learned policy, giving us (i) a compact visual representation of goals, (ii) improved generalization capabilities, and (iii) a mechanism for sampling feasible goals for exploration. Trained on a large offline dataset of prior experience, the model acquires a representation of visual goals that is robust to task-irrelevant distractors. We demonstrate our method on a mobile ground robot in open-world exploration scenarios. Given an image of a goal that is up to 80 meters away, our method leverages its representation to explore and discover the goal in under 20 minutes, even amidst previously-unseen obstacles and weather conditions.


Summary Video

Idea


Graphical Model of Goals, Actions and Distances

Our model uses images of goals and current observations to obtain a latent state-goal representation that summarizes the goal for the purpose of prediction the action and temporal distance to goal.

Exploring Open-World Environments with RECON

Combining the latent goal model with the topological graph, RECON can quickly discover user-specified goals in new environments and navigate to them reliably. Our method consists of three stages:

Example Environments

Robustness to Distractors

RECON plans over a compressed representation that ignores distractors in the environment, while the learned policy is reactive. It can explore a non-stationary environment, successfully discovering and navigating to the visually-specified goal. The learned representation and topological graph are robust to visual distractors, allowing RECON to reliably navigate to the goal under previously unseen obstacles and a variety of lighting and weather conditions.

BibTeX

@inproceedings{

shah2021rapid,

title={{Rapid Exploration for Open-World Navigation with Latent Goal Models}},

author={Dhruv Shah and Benjamin Eysenbach and Nicholas Rhinehart and Sergey Levine},

booktitle={5th Annual Conference on Robot Learning },

year={2021},

url={https://openreview.net/forum?id=d_SWJhyKfVw}

}