Life Science Research

With GRATA (GraphRAG-based training and education system for robot-assisted medical procedures), adesso is collaborating with leading clinical and research partners to develop an adaptive training and education system for robot-assisted eye surgery. The system is designed to combine modern image processing, 3D simulation, GraphRAG technology and generative AI to make high-precision procedures safer and more efficient to practise.

Robot-assisted eye procedures are highly complex, and operating times and resources are limited. Traditional training methods reach their limits here. GRATA creates a realistic, risk-free training environment in which trainee surgeons can repeatedly practise all phases of an operation, analyse errors and specifically improve their learning progress. The platform is designed so that it can be extended to other surgical disciplines in the long term.

The GRATA project is developing an adaptive training system that supports trainee doctors from preparation through to the performance of robot-assisted eye surgery.

What GRATA will specifically deliver in training:

• 3D surgical simulation and image processing for realistic representation of eye surgery scenarios
• An intelligent knowledge database that semantically describes procedures, roles and instruments and makes them available in context
• Real-time analysis of sensor and robot data to detect errors early and generate targeted instructions
• Language model support that answers users’ questions about the procedure and provides situational assistance during training
• Personalised learning paths that automatically adapt content to the trainee’s level of experience

adesso brings its expertise at the interface of life sciences, AI and software engineering to the joint project and takes on key technological tasks:

• Development of a training platform demonstrator that manages and visualises training content textually and graphically
• Implementation of a no-code concept with graphical UIs as well as natural language and AR-based user interfaces
• Validation of the training system in realistic training and working environments
• Development and integration of AI technologies (GraphRAG, LLM fine-tuning, prompt engineering)
• Modelling of medical processes and implementation of semantic interfaces (e.g. SPARQL)
• Establishment of an intelligent monitoring system for real-time evaluation of training activities

With GRATA, adesso is establishing a reusable foundation for robotics and AI solutions: GraphRAG knowledge models, no-code interaction and real-time monitoring can be applied to further scenarios – from training systems for other robot-assisted procedures to assistance and training solutions for surgical teams and medtech manufacturers.

Furthermore, the technology unlocks potential in other sectors: for example, in the robotics training and assistance industry, in logistics for intelligent automation, or in education and training for complex simulation and learning environments. In this way, a research project is transformed into a modular toolkit for scalable robotics and AI solutions extending far beyond the medical field.

Under the leadership of the Technical University of Munich Hospital (TUM Klinikum) – Department of Ophthalmology, Medical Autonomy and Precision Surgery Lab (MAPS) – adesso is collaborating in an interdisciplinary consortium with strong partners from research and industry.

The combination of clinical excellence, specialised robotics and AI research, and adesso’s technological implementation expertise is creating a pioneering training system for robot-assisted eye surgery.