How to Command Robots: A Multi-Robot Intelligence System

Robots are rapidly expanding beyond factory equipment into everyday life and service domains. At Naver 1784, around 100 autonomous robots per day move throughout the building to carry out delivery services, and an operating model in which humans and robots coexist in the same space is actually running.

At this stage, the key is not the performance of individual robots, but a central system that manages a large number of robots simultaneously. What Team Naver calls a multi-robot intelligence system or multi-robot orchestration system monitors and controls the movements of robots, coordinates tasks, and reflects the latest algorithms and map data in real time to raise overall efficiency. By processing computation centrally through a cloud-based structure, the burden on robots can be reduced, energy efficiency and operational efficiency can be increased, and sensor and computing costs can also be lowered.

This need is also evident in large-scale mobile robot services such as robotaxis. It is emphasized that to safely operate a large number of vehicles on the road, the intelligence of a single vehicle alone is not enough, and central control and integrated optimization must work together. As the scope of operation expands—such as airports, hospitals, complex shopping malls, and smart cities—the importance of such control systems grows even further.

At the center of this is ARC (AI-Robot-Cloud). Robots at 1784 are connected to the cloud via 5G to perform delivery, lunch box, and beverage services, and they move based on digital twin-based location information and optimal route updates. It also links with building infrastructure such as robot ports and door control to support movement.

The second pillar is a robot-dedicated OS. ARC mind is introduced as the world's first web platform-based robot OS that allows web developers to implement services such as ordering, payment, and recognition through robot-dedicated web APIs without needing a deep understanding of robot hardware. When a single service is commonly applied across multiple robots, the robot service ecosystem can expand rapidly like smartphone apps, and Team Naver envisions a future in which the control system, robot OS, digital twin, and cloud are combined into a single platform, allowing robots to operate naturally across an entire city.