Driving · 2026-10-09

LiteNWM: Efficient Latent World Models for Onboard Visual Navigation in the Wild

Linkai Liu, Yuntian Zhang, Zhenshan Bing, Chen Chen, Lingjuan Lyu, Shangguang Wang, Mengwei Xu, Dongqi Cai

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

TL;DR

LiteNWM is a latent navigation world model that evaluates multiple candidate trajectories more efficiently by sharing visual encoding across candidates and jointly predicting action-conditioned future latent representations at multiple horizons (H1–H4). A learned proposal-aware scorer selects trajectories. The authors report 17.56% lower macro trajectory error vs NoMaD+NWM-XL offline, 128.00× end-to-end speedup on RTX 5090, and real-robot success improvement from 43.3% to 83.3% vs NoMaD.

Why it matters

The authors target the dominant cost of candidate-wise RGB rollouts in navigation world models by restructuring evaluation around shared encoding and decoder-free latent prediction. This supports future-aware trajectory selection for closed-loop onboard navigation, and the same frozen evaluator transfers across navigation policies and deploys on a physical robot.

Method

Shared frozen visual encoding (DINOv3) encodes history and goal once per state, reused across K=16 candidate trajectories.
Action-conditioned latent predictor jointly predicts candidate futures at multiple horizons H1–H4 without autoregressive feedback (pose-conditioned warping + factorized blocks).
A learned proposal-aware scorer ranks challengers using predicted outcomes relative to a retained incumbent; selection uses a learned acceptance gate.

Limitation

The authors note that NWM’s MPC/CEM and terminal-frame objective were designed for eight-step rollouts, which may limit their effectiveness when adapted to H=4 short-horizon trajectories.

Abstract (from arXiv)

Direct visual navigation policies generate trajectories efficiently but do not explicitly evaluate their future consequences. Generative navigation world models provide this foresight through visual rollouts, which are costly when evaluating multiple candidates. We present LiteNWM, a latent navigation world model that shares visual encoding across candidates and jointly predicts their action-conditioned future representations at multiple horizons, while a learned scorer uses these predictions to select trajectories. In offline evaluations on RECON, SCAND, and SACSoN, LiteNWM reduces macro-averaged trajectory error by 17.56% relative to NoMaD+NWM-XL and achieves a 128.00-fold end-to-end speedup on an RTX 5090. The same evaluator transfers from NoMaD to MBRA without proposer-specific retraining, reducing MBRA's macro-averaged trajectory error by 16.2%. In real-robot experiments in unseen indoor and outdoor environments, LiteNWM improves navigation success from 43.3% to 83.3% relative to NoMaD. These results demonstrate that LiteNWM can be deployed for future-aware planning and closed-loop navigation on a physical robot.

Publication record

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

First public release: 2026-10-08 · Indexed by Pub-AI: 2026-10-09 UTC

preprint