We are very excited to share our latest technical paper, "Enhancing Robotic Autonomy via a Persistent Physics Layer".
This paper outlines the benefits of integrating a real-time, persistent physics engine operating in parallel with AI driving models.
Our Carbon framework is an ideal solution as a persistent physics engine, delivering a 0.5 to 1.0 second rolling predictive horizon, granting autonomous systems the ability to proactively improvise evasive actions. It also ensures occlusion immunity through a persistent digital double and calculates unknown payload mass in real-time using inverse dynamics.