Drift-Resistant Navigation World Model with Anchored Epipolar Guidance

arXiv 2026 61.1 method

TLDR

Proposes a drift-resistant navigation world model using anchor-guided rollout and epipolar geometry to reduce perceptual and geometric drift, improving long-horizon visual quality and planning.

Reasoning

The paper effectively addresses perceptual and geometric drift in navigation world models through a novel anchor-guided rollout with bidirectional epipolar constraints, demonstrating consistent improvements on four benchmarks. However, the approach is domain-specific to navigation and may need further validation across diverse environments.

Read-first score

Read-first score 61.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 45.

Recency 6%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,benchmark,experiment,metric,result

Citation impact 18%

85.7

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.85688937

Topical relevance 29%

64.3

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FoundationFrontierBridgeMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 342.

Keyword Scores

world model

generative world model

video world model

world dynamics prediction

world simulator

model-based reinforcement learning world model

interactive world model

Deep Analysis

Innovations

Anchor-guided rollout that predicts sparse future anchors as stable long-range targets before generating intermediate frames
Bidirectional epipolar geometry to provide geometric constraints for localizing content in intermediate frames
Joint mitigation of perceptual drift and geometric drift in navigation world models

Methodology

The method redesigns world-model prediction as an anchor-guided rollout. It first predicts sparse future anchors that serve as stable long-range targets, then generates intermediate frames within each chunk conditioned on both past context and future anchors. Bidirectional epipolar geometry is used to enforce geometric constraints on where corresponding content should appear in the intermediate frames.

Key Results

Experiments on four benchmarks show consistent improvements over strong baselines in long-horizon visual quality, geometric consistency, and multi-view coherence. These gains translate into improved downstream planning performance under the same planners.

Resources

Paper