Comparison of Model Free and Model-Based
Learning-Informed Planning for PointGoal Navigation

Yimeng Li*

Arnab Debnath*

Gregory Stein

Jana Kosecka

George Mason University

In CoRL 2022 LHP Workshop

[Paper]

[Video]

[Code]

In the point goal navigation task, a robot is deployed in a previously-unseen environment and is required to navigate to a location within limited number of steps. Here we show a running example of navigating to the point goal with the a frontier-based approach. A frontier is a boundary between free and unknown space, as denoted by the green pixels. Yellow pixels are the selected frontier.

We compare the state-of-the-art Deep Reinforcement Learning based approaches with Partially Observable Markov Decision Process (POMDP) formulation of the point goal navigation problem. We adapt the (POMDP) sub-goal framework proposed by Stein et al. and modify the component that estimates frontier properties by using partial semantic maps of indoor scenes built from images' semantic segmentation. In addition to the well-known completeness of the model-based approach, we demonstrate that it is robust and efficient in that it leverages informative, learned properties of the frontiers compared to an optimistic frontier-based planner. We also demonstrate its data efficiency compared to the end-to-end deep reinforcement learning approaches. We compare our results against an optimistic planner, ANS and DD-PPO on Matterport3D dataset using the Habitat Simulator. We show comparable, though slightly worse performance than the SOTA DD-PPO approach, yet with far fewer data.

Overview

Learning over Subgoals Planner

Learning Module

Results

Citation

                        @inproceedings{
                            li2022comparison,
                            title={Comparison of Model Free and Model-Based Learning-Informed Planning for PointGoal Navigation},
                            author={Yimeng Li and Arnab Debnath and Gregory J. Stein and Jana Kosecka},
                            booktitle={CoRL 2022 Workshop on Learning, Perception, and Abstraction for Long-Horizon Planning},
                            year={2022},
                            url={https://openreview.net/forum?id=2s92OhjT4L}
                        }

Acknowledgements

We thank members of the GMU Vision and Robotics Lab and RAIL.
This webpage template was borrowed from some colorful folks.