Lifelong Topological Visual Navigation
This work addresses navigation failures in lifelong visual navigation for agents in dynamic environments, though it is incremental as it builds on existing topological methods.
The paper tackles the problem of spurious or missing edges in topological navigation graphs, which often cause navigation failures, by proposing a sampling-based graph building method and lifelong maintenance strategies, resulting in sparser graphs with higher navigation performance and successful real-world fine-tuning.
Commonly, learning-based topological navigation approaches produce a local policy while preserving some loose connectivity of the space through a topological map. Nevertheless, spurious or missing edges in the topological graph often lead to navigation failure. In this work, we propose a sampling-based graph building method, which results in sparser graphs yet with higher navigation performance compared to baseline methods. We also propose graph maintenance strategies that eliminate spurious edges and expand the graph as needed, which improves lifelong navigation performance. Unlike controllers that learn from fixed training environments, we show that our model can be fine-tuned using only a small number of collected trajectory images from a real-world environment where the agent is deployed. We demonstrate successful navigation after fine-tuning on real-world environments, and notably show significant navigation improvements over time by applying our lifelong graph maintenance strategies.