Driving · 2026-11-03

Toward Uncertainty-Aware, Physics-Grounded Human World Models

Yufei Zhang, Zijun Cui

arXiv:2611.03428PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

The authors connect two published components to form an uncertainty-aware, physics-grounded human world model: KNOWN estimates SMPL body state along with aleatoric and epistemic uncertainty in real time (11.6 ms/frame), and PhysMoP predicts future body configuration by integrating Euler–Lagrange dynamics with learned terms. They define an interface to couple per-frame perception beliefs as the input history, propose three uncertainty-injection mechanisms, and specify an evaluation protocol on in-the-wild video.

Why it matters

For robots acting around people, the paper argues that a usable human world model needs both (i) a belief over body state and its uncertainty and (ii) a dynamics model that propagates that state forward. The authors propose how perception uncertainty can enter a physics-based motion predictor via initial-condition sampling, aleatoric covariance-weighted smoothing, and epistemic gating, plus a protocol that evaluates calibration under estimated inputs.

Method

  • KNOWN provides per-frame mean/covariance over SMPL pose and shape; aleatoric uncertainty from observation covariance and epistemic uncertainty from low-density training regions; it runs in 11.6 ms/frame.
  • PhysMoP predicts future body configuration by embedding Euler–Lagrange equations with learned inertia/Coriolis/forces and a fusion network combining a physics rollout with a data branch.
  • The proposed coupling sends KNOWN’s per-frame belief as a T=25 history to PhysMoP, using mechanisms: initial-condition sampling, covariance-weighted smoothing (aleatoric), and epistemic gating (minority threshold).

Limitation

KNOWN’s epistemic term is induced by the predicted state distribution and validated empirically, not derived from a posterior over network weights. KNOWN provides no metric root translation, and the protocol assumes a known camera pose. Neither component is conditioned on the robot’s actions, so the composite forecasts rather than answers counterfactuals; PhysMoP’s fusion weights are not yet informed by the perception uncertainty itself.

Abstract (from arXiv)

A robot that acts around people needs a world model of the human: a belief over the person's body state and a dynamics model that predicts where that state goes next. This abstract argues that two published components form exactly these two halves, and specifies how to couple them. KNOWN [1] reconstructs the 3D human body while estimating both aleatoric and epistemic uncertainty in real time (11.6 ms per frame); its aleatoric term identifies unreliable observations, its epistemic term identifies inputs from low-density regions of the training data, and refinement driven by the epistemic term reduces the error on such inputs from 81.3 to 70.3 mm. PhysMoP [2] predicts future body configuration by solving the Euler-Lagrange equations of the body with learned dynamics terms, reducing the 80 ms prediction error from 9.6 to 2.1 mm on Human3.6M and from 12.1 to 0.7 mm on 3DPW. We define the interface between the two, propose three mechanisms by which KNOWN's uncertainty enters PhysMoP, and specify an evaluation protocol on in-the-wild video. No new experimental results are reported here.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First public release: 2026-11-02 · Indexed by Pub-AI: 2026-11-03 UTC

preprint

Toward Uncertainty-Aware, Physics-Grounded Human World Models

Yufei Zhang, Zijun Cui

Preprint posted: 2026-11-02 · Original record ↗

arxiv
2611.03428
Explicit source versions (1)
  • v1 · 2026-11-02
  1. Preprint available

    Source date: 2026-11-02 · First detected: 2026-11-03 UTC · Historical baseline

    Source evidence (2)
Abstract · arxiv

A robot that acts around people needs a world model of the human: a belief over the person's body state and a dynamics model that predicts where that state goes next. This abstract argues that two published components form exactly these two halves, and specifies how to couple them. KNOWN [1] reconstructs the 3D human body while estimating both aleatoric and epistemic uncertainty in real time (11.6 ms per frame); its aleatoric term identifies unreliable observations, its epistemic term identifies inputs from low-density regions of the training data, and refinement driven by the epistemic term reduces the error on such inputs from 81.3 to 70.3 mm. PhysMoP [2] predicts future body configuration by solving the Euler-Lagrange equations of the body with learned dynamics terms, reducing the 80 ms prediction error from 9.6 to 2.1 mm on Human3.6M and from 12.1 to 0.7 mm on 3DPW. We define the interface between the two, propose three mechanisms by which KNOWN's uncertainty enters PhysMoP, and specify an evaluation protocol on in-the-wild video. No new experimental results are reported here.