Rookie 2 is a next-generation service robot that evolved from the extensive operational experience and vast data accumulated in NAVER’s smart building. Validated through real-world deployment with thousands of employees, Rookie 2 delivers enhanced stability, AI-driven intelligence powered by the ARC platform, and global versatility proven through international testing and feedback.
Evolution of Robustness and Reliability
Leveraging years of accumulated operational data, Rookie 2 has achieved a new level of robustness and reliability. By integrating expanded sensors with both cloud and on-device localization, it operates reliably even around glass walls, drop-offs, and network shadow zones. Multiple safety mechanisms—including physical brakes and force-sensing bumpers—further ensure safe coexistence with people in complex indoor environments.
Human-Robot Interaction: Understanding People and Spaces
Rookie 2 goes beyond simple path navigation; it perceives and interprets both spatial context and human behavior to respond appropriately to each situation. Developed by NAVER Labs’ researchers in Korea and Europe, its foundation-model-based spatial and physical AI estimates human posture, direction, speed, and distance in real time and then incorporates these insights into navigation decisions. In narrow corridors, it recognizes the situation and slows down or yields; in dense areas such as elevators, it analyzes crowd flow and reroutes accordingly. These HRI principles extend beyond basic collision avoidance, enabling the robot to understand human behavior, anticipate intent, and react naturally within shared environments.
Scalability Across the Entire Building (Cloud Platform & OS)
Rookie 2 operates not as a standalone unit but as part of NAVER’s ARC ecosystem, seamlessly connected to the building’s systems and infrastructure. Proven in NAVER’s 1784 smart building, this platform manages fleets of robots through multi-robot control (ARC brain) and precise localization (ARC eye), supporting diverse operational scenarios at the building level. Built on the web platform-based ARC mind OS, it also enables the seamless development and deployment of new functions and services, expanding the smart building’s service ecosystem.
Global Standards
Rookie 2 meets product and communication certifications across different countries and can be customized to regional requirements, such as Japan’s disaster-response mode. Intuitive interfaces like the Emergency Stop button and hidden handle control also make it easy for local operators to manage.
Supporting multiple languages, including Korean, Japanese, and English, it also adjusts driving behavior and notification sounds according to cultural context. The modular exterior design allows easy replacement, while certified materials ensure durability and compliance with international safety standards.
A Reference Model for the Physical AI Ecosystem
“Physical AI” refers to a robot’s ability to understand spaces, interpret contexts, and act autonomously. NAVER has been advancing this strategy through three key differentiators: real-world data accumulated from large-scale testbeds, a cloud platform that connects robotic and spatial intelligence, and lightweight foundation models embedded into onboard AI. Rookie 2 embodies this approach as a reference model verified and refined in real-world environments. By opening its hardware, OS, and robot APIs, Rookie 2 establishes a foundation on which a wide variety of services and applications can be developed seamlessly.
Advanced Features for Global Smart Buildings
Rookie 2 has been completely redesigned to deliver the capabilities required in real smart-building operations. Extensive data and expertise gathered from NAVER 1784 are fully reflected in its latest features.
Expanded Vision & Sensing System
A multi-sensor configuration combining multiple cameras, LiDAR, and ToF sensors enables precise perception of people, obstacles, and terrain. Visual localization ensures exceptional precision, with a positional margin of error of less than 15 cm and a rotational margin of error of under 3 degrees.
Adaptive Navigation with Force Sensing
By combining force-sensing bumpers with fine-tuned force control, Rookie 2 responds smoothly to movement or contact with people and obstacles. This allows for natural, stable navigation even in tight or dynamic environments where close proximity to humans or obstacles is unavoidable.
Modular Design and Extended Payload Capacity
Interchangeable modules allow Rookie 2 to handle a wide range of tasks—from large-volume delivery to multi-point dispatch—while the expanded loading capacity enables the transport of larger quantities.
Human-Aware, Stable Navigation
Rookie 2 moves fluidly by understanding human walking directions and dynamically recalculating routes in real time. It can safely pass through narrow corridors with as little as 10 cm clearance on each side, maintaining both precision and stability.
Multi-Floor Localization
By combining camera-based visual localization with barometric sensors, Rookie 2 accurately identifies its position across multiple floors. Integration with elevators and gate systems enhances the stability and continuity of movement throughout the building.
Emergency Mode Linked to Building Infrastructure
In emergencies, Rookie 2 synchronizes with the building’s emergency-response systems to assist with evacuation guidance and coordinated actions. Even if the network connection is lost, it autonomously determines safe paths, avoids restricted zones, and halts in secure areas to signal alerts.
Cloud Robotics Platform for Smart Buildings
The ARC platform provides multi-robot control (ARC Brain) and high-precision localization (ARC Eye), enabling the unified orchestration of multiple, heterogeneous robots, including Rookie 2.
Autonomous Charging & Full IoT Integration
Rookie 2 autonomously docks for charging, maximizing uptime and operational efficiency. It is integrated with the building’s IoT and BMS systems, autonomously controlling key facilities such as lighting, access doors, and elevators for movement and service execution.
Team Commentary
Built upon NAVER’s smart building experience and Physical AI, Rookie 2 is ready to set a new global benchmark. It represents the collective expertise and shared principles of specialists across robotics, HRI, design, platform development, and service operations.
“The core philosophy of NAVER’s robotics is to create practical robots that care for people. We focus on functions that add value when performed by robots, and interactions that feel natural and familiar.”
— HRI (Human-Robot Interaction) Team
“This isn’t just a performance upgrade; it’s been re-engineered with robustness and reliability as its core values. It operates stably across different national environments and comes with high scalability and operational resilience.”
— Robotics Team
“Every detail of Rookie 2’s exterior carries intention. It delivers a refined and harmonious experience while maintaining strong functionality.”
— Robot Design Team
“The ARC system—the brain of robots within smart buildings—goes beyond NAVER’s own robots. It’s a platform built to empower any robot, anywhere in the world.”
— ARC brain Team
“We’re opening up a new path that overcomes the limits of speed and scalability in robot development.”
— ARC mind OS Team
“NAVER’s latest AI research is built on vast datasets, and Rookie 2 is the embodiment of that research. Its ability to understand spaces, perceive people, and make situational judgments is what enables true autonomy in complex real-world environments.”
— Vision AI Team