Awesome World Model Hub Papers · Datasets · Projects
← Back to papers

Path Planning in Physically Viable World Models

arXiv 2026 39 method, system

TLDR

A physically viable world model using physics simulation on 3D scenes to evaluate robot navigation under terrain changes, tested with simulated flooding on a real field site.

Reasoning

Strengths include a novel integration of physics simulation with 3D reconstructions for long-horizon planning and real-world evaluation. Weaknesses are the limited scenario (flooding only) and lack of baseline comparisons in the abstract, with potential scalability concerns.

Read-first score

Read-first score 39, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 41.

Recency 6%
100

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

Methodology quality 18%
60

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

Topical relevance 29%
58.6

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%
30

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

Citation impact 18%
0

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 12%
0

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

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 231.

Keyword Scores

world model
9
world simulator
9
world dynamics prediction
8
interactive world model
7
generative world model
6
video world model
1
model-based reinforcement learning world model
1

Deep Analysis

Innovations

  • Physically viable world model that augments 3D Gaussian splat scenes with physics-based simulation to generate physically modified environments without recollecting data or rebuilding maps.
  • Terrain-aware planner that accounts for simulated physical events, obstacles, and deformations to evaluate route feasibility under future terrain changes.
  • What-if query evaluation for robot navigation that exposes long-horizon route failures and rerouting behavior not apparent in static reconstructions.

Methodology

The system augments reconstructed 3D Gaussian splat scenes with physics-based simulation to create physically modified versions of the environment. A terrain-aware planner then evaluates route feasibility under these simulated changes. The approach is tested on a real outdoor field site in Central Texas with simulated flooding at multiple severity levels, measuring route and mission feasibility as terrain deteriorates.

Key Results

Physically viable world models reveal long-horizon route failures and rerouting behavior that are not apparent when planning only on the original reconstructed environment, enabling pre-deployment evaluation of how future terrain changes affect route feasibility.

Tags