KIROS application
AI-supported robotic systems and simulation platform for autonomous applications in the healthcare industry.
Since February 18, 2025, it has been clear: The Hamburg Innovation and Development Bank has decided to fund our project. The total funding amounts to a non-repayable 1.9 million euros. The project aims to develop an AI-supported solution called KIROS, which makes the manual adaptation process of robots by the manufacturer superfluous. Currently, these customizations are often expensive and time-consuming, which discourages companies from using robots. KIROS uses reinforcement learning (RL) to enable robots to autonomously learn and apply new skills without the need for a manufacturer.
Problem description from the user‘s perspective: KIROS is a SaaS platform that enables users to train robots via a web interface and learn new skills with little programming knowledge. The training takes place in a simulated environment that enables the robot to act in the real world. The use of cobots, which can work in dynamic and unpredictable environments, is particularly important in the healthcare sector, where robots are increasingly being integrated.
Example of a planned solution: An example shows how a pharmacy can train a cobot, such as the UR5 from Universal Robots, to sort medicines without manufacturer assistance. The process includes:
scanning the environment and objects to create 3D models.
virtually training the robot in a simulated environment to learn specific tasks
testing and executing the tasks in the real environment after the training is completed.