These five highlights of geriatronics research from 2024 show the potential of care assistance robotics.
Research in the field of geriatronics has been underway at the TU Munich campus in Garmisch-Partenkirchen for over five years. First and foremost is the GARMI care assistance robot. It is set to develop into an all-in-one helper and later assist nursing staff. GARMI's repertoire already includes a lot, from simple pick-up and delivery services to possible physiotherapeutic use in the rehabilitation of patients and telemedical applications such as ultrasound.
If you want to know how much science there is in the individual capabilities of GARMI, take a look at the publications of the researchers involved. In the past six months of 2024, the presentations at the international robotics conference IROS, which took place in Abu Dhabi in mid-October, stand out in particular. They show the details of the overall development of the robotics all-rounder. The benchmark for this is always the mandatory criterion of safety and efficiency in use:
1. Robotic tasks in dynamic environments demand precision and safety
Researchers Seongjin Bien, Abdeldjallil Naceri, Luis Figueredo and Sami Haddadin have developed a new framework that enables robots to precisely avoid obstacles - even during ongoing tasks such as object handling or remote control. Called Generating Force Vectors from Projective Truncated Signed Distance Fields for Collision Avoidance and Haptic Feedback, this advanced approach combines 3D mapping with force-based feedback. This helps GARMI to recognize nearby objects and smoothly adapt its movements to them. By accurately detecting object shapes without complex processing, the system improves the robot's ability to perform tasks in crowded or dynamic environments. This technology has been successfully tested in real-world scenarios, including a teleoperation environment where users control robots remotely while receiving realistic haptic feedback. This development improves the performance of GARMI, especially for those tasks that require precision by doctors as remote users of GARMI. At the same time, it paves the way for safer, smarter robots in healthcare, industry and other areas.
2. How an upper limb exoskeleton enables the intuitive teleoperation of remote robots avatar
To provide optimal support for doctors and nurses in the future, researchers Moein Forouhar, Hamid Sadeghian, Daniel Pérez-Suay, Abdeldjallil Naceri, and Sami Haddadin dedicated themselves to the design of an exoskeleton that is intended to revolutionize the interaction of professionals with assistant robots such as GARMI. Titled A Tactile Lightweight Exoskeleton for Teleoperation: Design and Control Performance, the system not only offers potential wearability and tactile responsiveness, but also features a fully actuated shoulder joint as well as sensors that directly measure interaction forces. These advances allow users of the exoskeleton to control GARMI with natural, smooth movements. This ensures that the robot works safely and efficiently in tasks that require high precision. Advanced control strategies ensure that the exoskeleton adapts perfectly to the user's commands, increasing comfort and reducing effort. With its ability to integrate seamlessly into the health and care environment, this exoskeleton will make the operation of robots like GARMI easier and more intuitive than ever.
3. Human skills mapping and transfer to the robots
Zhelin Yang, Seongjin Bien, Simone Nertinger, Abdeldjallil Naceri and Sami Haddadin have set themselves the goal of teaching robots to move like humans in order to enable us to work better with them. Under the title An Optimization Based Scheme for Real-Time Transfer of Human Arm Motion to Robot Arm, they have developed a system that allows robots to mimic human movements in real time. This breakthrough enables the robot to replicate human arm movements with precision, making its actions intuitive and natural. By using advanced optimization techniques, the system ensures accurate motion transfer while maintaining human-like movement patterns. What's useful is that this technology seamlessly integrates with GARMI's state-of-the-art exoskeleton interface, which - as described in point 2 - enables doctors, nurses and other professionals to control the robot effortlessly. The combination of intuitive motion targeting and responsive exoskeleton control paves the way for more effective and user-friendly robotic care and service applications. This is yet another step towards usability.
4. Unreliable internet connectivity for safe remote control of the robot avatar?
The solution to the problem: To improve the control of robots with more intelligent communication systems, researchers Xiao Chen, Youssef Michel, Hamid Sadeghian, Abdeldjallil Naceri and Sami Haddadin have taken a new approach. Entitled Network-aware Shared Autonomy in Bilateral Teleoperation, they show how to improve the performance of remote-controlled robots such as GARMI, especially in environments with unstable Internet connections. Together, they have developed a system that adjusts the level of autonomy based on the quality of the communication network to ensure smoother and more efficient control, even in the face of delays or interruptions. The key: through the shared autonomy model, the system allows the robot to take over tasks autonomously when required, while the human controller still remains in control. This method has been confirmed by user studies to significantly increase performance and make teleoperation more reliable and effective, even under difficult network conditions. This breakthrough is definitely a step forward in making robotic care and service tasks more efficient and user-friendly in the future.
5. How digital twins enable robots to learn
Researchers face the constant challenge of teaching robots new skills, and not just in geriatronics. Dr. Abdeldjallil Naceri's workshop presentation at IROS 2024 was on how new skills for safe interaction can be transferred to robots via digital twins. For GARMI, for example, this means the ability to perform a task such as grasping with the necessary adaptability and safety, while its digital twin enables real-time monitoring and the learning of interaction strategies. In addition, GARMI uses a data recorder to securely manage patient data, optimize performance and, of course, take data protection into account. This innovative approach, entitled Variable Impedance Learning and Control: Navigating Challenges, Exploring Opportunities, and Shaping the Future, aims to achieve significant improvements in the care of seniors through advanced robotics and AI personalized, human-like assistance.