A robot that picks up drinks and snacks on its own in front of a supermarket shelf and arranges them. It is still an unfamiliar scene, but a domestic startup has brought it to the stage of applying it to real industrial sites. Kairovada, a spinoff corporations that began in the research team of Kim Hee-won, a professor in the School of AI Software at Soongsil University.
Kairova is an AI robot corporations that designs both the software (SW), the "brain" of the robot, and the hardware (HW), its "body." Unlike the existing approach of developing only robot control software or producing only hardware, it provides an integrated robot solution that can operate immediately in real industrial settings.
The problem Kairova focuses on is a long-standing limit in the robot industry. If the location of an item changes even slightly, or in environments where people and obstacles move together, existing robots have difficulty operating properly. In spaces without fixed rules, such as supermarkets, logistics centers, and production sites, automation has progressed slowly.
The company chose to first train robots in virtual space and then apply the results to real robots. Using Nvidia's robot simulation platform "Isaac Sim," it completes thousands to tens of thousands of trial-and-error cycles in a virtual environment so the robot can carry out tasks immediately in the real world. This can significantly reduce time and expense in the physical experimentation process.
Kairova has already completed its proof-of-concept (PoC). It reproduced convenience store and supermarket shelves in a virtual environment and succeeded in reliably recognizing, picking up, and arranging drinks and snacks.
To that end, it also built its own dual-arm mobile robot for demonstration. Combining a mobile platform with two robot arms, it verified whether software and hardware operate together in an environment similar to an actual store.
Chief Executive Kim Hee-won said, "Our strength is that we directly designed the entire architecture, from robot control to sensor inputs and AI decisions," adding, "We conducted precise analyses of work procedures, logistics routes, and robot operating ranges and reflected them in the hardware design."
Kairova's goal does not end with convenience store shelf-stocking robots. It is currently advancing simulations that take into account supermarket and large logistics center environments. It is refining the technology so robots can make their own decisions and continue working even in logistics sites where travel routes are complex and obstacles change frequently.
Going forward, it plans to expand applications to manufacturing processes and service robots. The structure is not dependent on any specific robot maker, allowing it to apply software to existing robots or, if necessary, provide custom robot hardware as well.
Kairova's technical foundation has already been validated in academia. Chief Executive Kim Hee-won has presented numerous papers at global AI conferences, including the Computer Vision and Pattern Recognition conference (CVPR) and the Association for the Advancement of Artificial Intelligence (AAAI). In the international competition that tests robot manipulation abilities (Arnold Challenge), it won world No. 1 this year and placed third last year, proving competitiveness in real control technology.
The company currently has 16 researchers with master's and doctoral degrees, and talent with practical experience at Nvidia, Qualcomm, and the Korea Institute of Science and Technology (KIST) is leading development.
Alongside technology development, Kairova also emphasizes its social role. Supermarkets, logistics centers, and production sites suffer from chronic labor shortages. If robots can take on repetitive and strenuous tasks, they can expect two effects at once: easing labor shortages and improving working conditions.
Kim said, "Rather than replacing people, we want robots to be tools that take on work that is hard for people," adding, "Our goal is to build technology that can actually be used in the field."
Kairova moved into the Soongsil University Campus Town as part of the Seoul City Campus Town program. The Soongsil University Campus Town Promotion Team selects promising startups as resident corporations and provides startup space, practical training, business advancement support, and expert advice by field.