Learning Language-Conditioned Traversability Representations for Adaptive Visual Navigation

Key points

• Traversability is essential for visual navigation but varies with robot capabilities and user preferences.1
• Moreover, viewpoint-dependent segmentation masks complicate asynchronous planning under perception latency.1
• To train the system, we develop a simulation-based data generation pipeline with controllable trajectories, producing observations paired with language instructions, traversability maps, goal locations, and diverse trajectories.1
• Evaluations on datasets from multiple sources demonstrate effective language-guided traversability segmentation and goal localization by the slow VLM, alongside adaptive pixel-space path planning by the fast planner.1

Sources (1)

• [1]Learning Language-Conditioned Traversability Representations for Adaptive Visual Navigation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 10:03 AM

  We present LaTraNav, a framework that learns language-conditioned traversability representations for adaptive visual navigation.

  Traversability is essential for visual navigation but varies with robot capabilities and user preferences.

Extractive summary: sentences quoted from the sources.