Learning Exploration Policies for Navigation
This work addresses the challenge of task-agnostic exploration for navigation agents, which is incremental as it builds on existing methods but focuses on a less-studied aspect.
The paper tackles the problem of autonomous exploration in complex 3D environments without task-specific rewards, proposing a learning-based approach that uses policies with spatial memory, bootstrapped with imitation learning and fine-tuned with coverage rewards, which outperforms classical geometry-based and generic learning-based methods.
Numerous past works have tackled the problem of task-driven navigation. But, how to effectively explore a new environment to enable a variety of down-stream tasks has received much less attention. In this work, we study how agents can autonomously explore realistic and complex 3D environments without the context of task-rewards. We propose a learning-based approach and investigate different policy architectures, reward functions, and training paradigms. We find that the use of policies with spatial memory that are bootstrapped with imitation learning and finally finetuned with coverage rewards derived purely from on-board sensors can be effective at exploring novel environments. We show that our learned exploration policies can explore better than classical approaches based on geometry alone and generic learning-based exploration techniques. Finally, we also show how such task-agnostic exploration can be used for down-stream tasks. Code and Videos are available at: https://sites.google.com/view/exploration-for-nav.