{
  "schema_version": "paper_public_manifest_v1",
  "paper_id": "schrodinger-navigator_2026_06",
  "slug": "schrodinger-navigator",
  "title": "Can Navigation Robots Learn to Think Before They Walk? A Solution to the Occlusion Challenge?",
  "authors": [],
  "source": {
    "arxiv_id": "2512.21201v1",
    "pdf_url": "https://arxiv.org/pdf/2512.21201v1",
    "project_url": "",
    "github_url": "",
    "huggingface_url": "",
    "original_source": "https://arxiv.org/abs/2512.21201v1"
  },
  "site": {
    "post_url": "/posts/schrodinger-navigator",
    "canonical_url": "https://haiguangboy.com/posts/schrodinger-navigator",
    "cover_image": "https://static.haiguangboy.com/papers/schrodinger-navigator/cover.webp"
  },
  "taxonomy": {
    "domain": "embodied_ai",
    "track": "world_model",
    "tasks": [
      "embodied_ai",
      "world_model",
      "robotics",
      "state_prediction",
      "Embodied Intelligence",
      "Robot Navigation",
      "World Models",
      "Zero-Shot",
      "Quadruped Robots"
    ],
    "related_topics": [
      {
        "paper_id": "fast-wam_2026_07",
        "title": "No Need to Imagine the Future at Inference—Robots Still Hit 91.8%! Fast-WAM Debunks WAM's Core Assumption",
        "url": "https://haiguangboy.com/posts/fast-wam",
        "relation": "contrast",
        "summary": "Training video objectives matter more than imagining the future at test time",
        "strength": "strong"
      },
      {
        "paper_id": "d4rt_efficiently_reconstructing_dynamic_scenes_one_2026_06",
        "title": "D4RT_Efficiently_Reconstructing_Dynamic_Scenes_One",
        "url": "https://haiguangboy.com/posts/d4rt_efficiently_reconstructing_dynamic_scenes_one",
        "relation": "same_track",
        "summary": "ZSON's real bottleneck is occlusion and dynamics, not pure semantic understanding—the myopia of current visual input is the core gap, validating D4RT's use of a single feedforward interface to produce point clouds/depth/point trajectories/camera parameters (static + dynamic)",
        "strength": "strong"
      }
    ]
  },
  "ruling": {
    "importance_score": 3.0,
    "one_sentence": "Zero-shot navigation: Schrödinger's Navigator"
  },
  "asset_base_url": "https://static.haiguangboy.com/papers/schrodinger-navigator",
  "assets": [
    {
      "type": "pdf_screenshot",
      "object_key": "papers/schrodinger-navigator/page_01.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/schrodinger-navigator/page_01.webp",
      "role": "paper_first_page",
      "size_bytes": 157596
    },
    {
      "type": "pdf_screenshot",
      "object_key": "papers/schrodinger-navigator/key_figure.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/schrodinger-navigator/key_figure.webp",
      "role": "method_figure",
      "size_bytes": 180434
    },
    {
      "type": "cover_image",
      "object_key": "papers/schrodinger-navigator/cover.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/schrodinger-navigator/cover.webp",
      "role": "post_cover",
      "size_bytes": 71852
    },
    {
      "type": "public_brief",
      "object_key": "papers/schrodinger-navigator/public_brief.md",
      "content_type": "text/markdown; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/schrodinger-navigator/public_brief.md",
      "role": "public_brief",
      "size_bytes": 3365
    },
    {
      "type": "public_manifest",
      "object_key": "papers/schrodinger-navigator/public_manifest.json",
      "content_type": "application/json; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/schrodinger-navigator/public_manifest.json",
      "role": "public_manifest",
      "size_bytes": 4160
    }
  ],
  "created_at": "2026-07-18T08:59:17+08:00",
  "updated_at": "2026-09-02T10:55:06+08:00",
  "published_at": "2026-07-18T08:59:17+08:00",
  "analyst_take": {
    "type": "author_opinion",
    "text": "This paper directly clashes with WAM in the library—WAM claims 'future imagination at inference is useless; the key is video co-objectives during training,' while Schrödinger's Navigator argues 'in occluded/dynamic scenes, imagination at inference is the core gain.' Both papers have strong 3/3 consensus yet contradict each other, indicating this isn't either side's fault but rather scenario-dependent: the stronger the dynamic occlusion, the more valuable inference-time imagination; in static environments, training-time representations are the primary variable. Meanwhile, D4RT focuses on pixel-level reconstruction in dynamic scenes, and QWEN-RobotWorld builds state-transition world models—three papers approach the same direction from reconstruction/prediction/planning angles: dynamic perception is the next real frontier. For nav-brain's Orin NX real-time challenges, refer to π0.7's RTC asynchronous inference engineering solution."
  }
}