AquaWAM: A Dynamics-aware World Action Model for Underwater Embodied Agents

2026-09-27

AI Summary

This study addresses the prediction inaccuracies of existing underwater agent models caused by neglecting passive dynamics such as inertia and buoyancy. We propose the first world action model tailored for underwater agents, which fuses multi-source sensor data—including Doppler Velocity Log (DVL) and Inertial Measurement Unit (IMU)—to explicitly model passive physical dynamics within a low-dimensional, compact navigation state space, replacing conventional high-dimensional visual observations. This approach substantially reduces computational overhead while enhancing system robustness. Experimental results demonstrate that the proposed model achieves a 72.6% success rate on USIM benchmark tasks and accelerates decision-making by 2.7×. Furthermore, it maintains high performance even under conditions of partial sensor failure.

Abstract

World Action Models (WAMs) are becoming increasingly important and useful for embodied intelligence, as they enable robots to anticipate the consequences of candidate actions before interacting with the physical environment. However, underwater robots are usually subject to passive dynamics, such as inertia, buoyancy, hydrodynamic drag, and persistent drift, which can continue to affect the vehicle even after an action is completed. Existing WAMs, which primarily predict action-conditioned visual observations, are not explicitly designed to capture such passive motion dynamics. In this paper, we present AquaWAM, the first World Action Model designed for underwater embodied agents. Instead of predicting future images, AquaWAM models both action-conditioned and passive physical dynamics, including the thruster dead band, the inertial glide that outlasts each command, and ambient currents. Specifically, it senses through the DVL, IMU, pressure sensor and joint encoders, while cameras supply only semantics for understanding goals and target pose. By modeling compact navigation states rather than high-dimensional visual observations, AquaWAM substantially reduces the model size and computational cost compared with conventional WAMs. Experimentally, AquaWAM achieves a 72.6% task success rate across 20 underwater tasks on the USIM benchmark, outperforming existing methods while making action decisions 2.7x faster than U0 on an NVIDIA Jetson AGX Orin. Our model also remains effective when some onboard sensor measurements are unavailable. For example, without DVL velocity measurements, our method still achieves a 61.6% success rate, compared with 39.4% for U0.

Problem

• World Action Model
• Underwater Embodied Agents
• Passive Dynamics
• Motion Prediction

Innovation

• World Action Model
• Underwater Embodied Agents
• Passive Dynamics Modeling
• Compact Navigation States
• Sensor Robustness