Robust decision-making improves robot safety using latent space disturbances
Modeling Latent Disturbances for Robust Decision-Making in World Models
RoboticsArtificial IntelligenceMachine Learning
Summary
Robots use world models to make decisions based on simplified internal maps of their environment. The paper shows how to handle unexpected changes or errors in these internal maps by carefully modeling possible disturbances in the robot’s thinking process. This lets the robot plan safer and more reliable actions, even when things don’t go exactly as expected. The authors tested their approach in simulations and on a real robot arm, reducing failures significantly.
What this means in practice
- •For robotics engineers: Enable robots to plan actions that withstand worst-case internal model errors for safer real-world operations.
- •For robotic manipulation teams: Improve control of robot arms by reducing failures caused by unexpected changes in internal world representations.
Authors
Junwon Seo, Andrea Bajcsy
Abstract
In this paper, we study robust decision-making in the latent space of world models (WMs). Robust optimization is a mathematical framework where, given explicitly specified dynamics and physically meaningful disturbances, a robot can select actions that remain effective even under worst-case disturbances. However, applying this principle to the learned latent space of WMs introduces a fundamental challenge: because WMs have fully learned state spaces and dynamics inferred from high-dimensional observations, it is unclear how to define latent-space disturbances that faithfully represent uncertainty in the underlying system. Our key idea is to model a latent-space disturbance as a perturbation to the learned latent dynamics that induces pessimistic but plausible transitions. Specifically, we construct a set of plausible latent dynamics by combining a dynamics-aware similarity metric that captures plausible transitions with out-of-distribution detection that excludes implausible latent states. We calibrate this uncertainty set over latent dynamics using conformal prediction, ensuring that WM imaginations induced by the latent disturbance remain plausible without becoming overly pessimistic. We then jointly optimize robust robot actions and the worst-case latent disturbances through game-theoretic optimization. We leverage this latent-space robust optimization to robustify policy steering, considering two paradigms: latent safety filtering and sample-and-verify steering of a generative control policy. Our controlled simulation experiments show that our latent disturbance enables robust decision-making directly in WM latent spaces, and hardware experiments with a Franka manipulator show that modeling latent disturbances enables robust policy steering, reducing failures by 70% in safety filtering and 54% in sampling-based policy steering. Project website: https://junwon.me/LatentDisturbance/.