Mechanics and Control of Human-Robot Interaction
The healthcare sector is under pressure worldwide due to shortages in nursing and care-support staff. My team aims to develop a humanoid robotic platform that can learn to collaborate with human caregivers, or even autonomously take over specific nursing tasks, to ensure a workable healthcare sector in the future.
Background
I am Mark Vlutters, and I am taking the lead in this research on Mechanics and Control of Human-Robot Interaction. I obtained my Bachelor (2010), Master (2012), and PhD (2017) degrees in Biomedical Engineering at the University of Twente, all three with distinction (cum laude). After my PhD I worked in business for several years, developing body motion tracking software based on video images and machine learning. I am now an assistant professor at the Biomechanical Engineering group.
The shortage of qualified personnel in the health sector became painfully clear during the recent COVID-19 pandemic. More shortages are expected in the future due to the general ageing of the world's population. Therefore, we aim to develop a robotic system that can support healthcare staff by collaborating with humans on specific tasks, or even fully take over those tasks autonomously. We are planning to do so by using machine learning techniques to have the robot learn during operation from its environment, and humans in that environment.
In the years to come
In this line of research, I and my team aim to develop a humanoid robot that can support healthcare staff in a variety of tasks. The overall objectives are to have the robot:
1) collaborate with nursing staff on a task. That is, both robot and human can simultaneously work towards the same objective. This could be through direct physical interaction, as well as through tele-manipulation. In the former, both human and robot may physically interact to complete a task, such as moving a patient from one bed to another together. In the latter, a human may remotely control the robot, while the robot itself also maintains some form of autonomy on the task.
2) take over the task from a human. The robot is able to autonomously execute a variety of supporting tasks, such as gathering materials from hospital storage. The robot may require a demonstration to learn a specific task, but will require minimal human input during execution of the learned task.
In the coming period, the goal is to have the robot learn a variety of motion tasks using human demonstration and translate the learned skills to new tasks and scenarios in a healthcare setting, all while maintaining safety for the environment and the robot itself. Be care-ful!
Education
Current related master courses in which I am involved are Learning and Adaptive Control as well as Human Movement Control. These courses are, given in the master programmes Mechanical Engineering, Systems and Control and Biomedical Engineering. Furthermore, I am also involved in the BioRobotics minor, in which various bachelor programmes participate to build their own robot.
Collaborations
The main platform in this research will be a unique humanoid robot system that was realised through a collaboration with Halodi Robotics and OttoBock. Halodi Robotics has developed a robot that is safe for interaction with the environment, while OttoBock has created a set of human-like hands that are integrated into the robotic device to allow dexterous grasping and interaction with the environment.