Driver-WM: A Driver-Centric Traffic-Conditioned Latent World Model for In-Cabin Dynamics Rollout
arXiv:2605.05092v2 Announce Type: replace-cross Abstract: Safe L2/L3 driving automation requires anticipating human-in-the-loop reactions during shared-control transitions. While most driving world models forecast the external environment, in-cabin intelligence remains strictly recognition-oriented and lacks multi-step rollout capabilities for driver dynamics. We introduce Driver-WM, a driver-centric latent world model that rolls out in-cabin dynamics causally conditioned on out-cabin traffic context. This formulation unifies physical kinematics forecasting with auxiliary behavioral and emotio
The increasing sophistication of autonomous driving systems and the need for safer L2/L3 automation are pushing the boundaries of in-cabin human-machine interaction models.
This development allows for more robust anticipation of human behavior in partially automated driving, leading to safer and more reliable shared-control systems.
In-cabin intelligence shifts from mere recognition to predictive multi-step rollout capabilities for driver dynamics, integrating driver states with external traffic conditions.
- · Autonomous vehicle manufacturers
- · Automotive safety systems developers
- · AI model developers
- · Insurance companies
- · Manufacturers relying solely on external environment world models
Improved safety and reliability of L2/L3 autonomous vehicles through better human-in-the-loop predictions.
Reduced incidence of accidents and enhanced consumer trust in partially automated driving technologies.
Accelerated adoption of higher levels of driving automation as human interaction risks are better managed and mitigated.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI