We extend the Rhoda Direct Video-Action (DVA) model to combine manipulation and mobility in one policy, enabling it to learn new mobile tasks from as little as 15 minutes of additional robot data.
Read articleOur team focuses on deploying robotic systems into the real world. We build general purpose foundation models that can adapt to the variability of commercial and industrial environments.
FutureVision brings the capability to handle real world industrial tasks autonomously. Our Direct Video Action model unlocks generalization that conventional VLA pipelines struggle to achieve. Rhoda works not just in labs but in factories, warehouses, and real production environments.
We first pre-train our model with hundreds of millions of videos, giving it a strong prior on motion, physics, and dynamics. We then post-train on robot data with actions to teach the model to learn specific tasks. This combination of pre- and post-training creates generalist robot policies that can work and adapt in dynamic environments.
We explore a new paradigm for scaling robot intelligence with web-scale video pre-training. We’ve published an in-depth research blog to discuss the novel architectural designs and the benefits of such an approach.
We work with a variety of customers across verticals in automotive, manufacturing, logistics, and ecommerce. If you’re interested in working with us at your facility, reach out here.
Strength: Custom actuators enable 25kg rated, 40kg peak
Safety: Wheel-base, brakes in every actuator, safety-rated vision
Reliability: 3 years of continuous operation at rated payload
AI Control: High component stiffness, linear response
