Piaggio Fast Forward (PFF) unveiled its next-generation hybrid navigation technology for its industry-leading PFF kilo® robot at MODEX 2026. The exclusive hybrid navigation technology features a production-ready physical AI model that significantly enhances a robot's ability to exhibit intelligent, intuitive, and dynamic person-following behaviors.

Utilizing a hybrid control framework that actively infers its leader's intent, kilo can now seamlessly toggle between AI models and algorithm-based control to handle dynamic forward and backward following motions with safety constraints.

The AI model, which bypasses the cloud and operates entirely on edge hardware, was trained using PFF's proprietary, in-house multi‑agent 3D kinematic capture data. Known as the "Smart Behaviors Dataset," these findings are the result of eight years of capturing human proxemic cues, behavioral patterns, social negotiation, and object handling. It is one of the world's largest known high-fidelity datasets of multi-person motion.

"We've built our technology around understanding human behavior in real environments, so our robots don't need to look human to feel intuitive," says Greg Lynn, Co-Founder and CEO of PFF. "They know when to wait, when to move, and how to navigate shared spaces in a way that feels familiar. Hybrid navigation is a big step forward in making the ‘decision making' of our robots appear even more seamless to the user. This combination allows our robots to blend predictable movements with faster, context responsive intelligence; defining a new standard for what collaborative and intuitive means in real world robotics."

kilo robots augment workers in environments where they are safely and productively deployed to transport goods. PFF's robots work with ease out-of-the-box and without the need for user training or changes to existing infrastructures. Familiar motions, such as intuitive forward and backward following, are what defines the robots' signature pedestrian etiquette.

With hybrid navigation, PFF's robots are now even better at interpreting social cues. In combining deterministic behaviors with their own AI models, these machines will take predictable, certifiable control logic to the next level.