Notable results out of academic and corporate research groups.
A new edge-based solution called Conductor enhances the planning algorithms of Connected Autonomous Vehicles by creating a unified world model. This method improves trajectory predictions while adhering to strict time constraints, demonstrating superior performance in simulations.
MOSAIC-SV is a new adaptive control system for aquatic robots that improves performance by continuously updating dynamics models. It has shown significant speed and accuracy improvements in both simulations and real-world trials.
This research presents a novel framework for improving failure detection in vision-language-action robotic systems. By leveraging unlabeled data and active learning, the method enhances detection accuracy during manipulation tasks.