Helga the Robot in Trial Operation on the Children's Ward

Enlarge

Sarah (name changed) is sitting upright on a bed in Pediatric Surgery Ward 2 at Hauner Children’s Hospital, looking at a screen. “Finally, I’ll measure how your eyes react. If you’re wearing glasses, please take them off now,” says Helga in a clear voice. Helga is a robot and has been part of the pediatric surgery ward team since this spring. In the future, she could assist the ward team in monitoring children with mild traumatic brain injuries.

The project is led by Prof. Dr. Jan Gödeke, Deputy Director of the Department of Pediatric Surgery at Dr. von Hauner Children’s Hospital and Professor of Digitalization in Pediatric Surgery. The Pediatric Surgical Center for Robot-Assisted Surgery at LMU Medical Center has been in operation since 2023, and the operating room there has been using a Senhance Surgical Robotic System ever since. Now, researchers aim to determine whether and how an assistive robot like Helga can support the team on the ward.

“We need to consider how humans and machines can work together on the wards, if only because of the shortage of skilled personnel,” says Gödeke. The goal: to relieve staff of tasks that can be “automated as standard.” This will free up more time for other important medical tasks, for direct patient contact, and for discussions with parents. The Hauner Association is supporting the project with a generous donation.

Together with the Fraunhofer Institute for Manufacturing Engineering and Automation (IPA) in Stuttgart, Gödeke has analyzed where and how digital services can support doctors and nursing staff. He says that highly standardized processes and activities that require listening and observation are primarily suitable. “When it comes to palpation or similar tasks, it becomes more difficult.” The first test case at the Hauner Children’s Hospital: monitoring children with mild concussions.

Concussion is one of the most common reasons for inpatient hospitalization among children in a pediatric surgery clinic. These children are then kept on the ward for 24 to 48 hours for monitoring. Nurses regularly check pupil response, heart rate, verbal and motor responses, and general well-being—initially every hour, then at slightly longer intervals. Each check takes five to ten minutes and involves the use of standardized assessment tools such as the Glasgow Coma Scale. To date, centralized bedside monitoring systems have been unable to adequately perform these tasks.

For a year, Dr. Birgit Graf and her team at Fraunhofer IPA have been fine-tuning the robot so that it can take on these tasks. For more than 25 years, the computer scientist has been working on the practical development of robotic assistance systems in the healthcare sector and in elder care. These systems are used in logistics and service, cleaning, housekeeping, documentation, and lifting tasks.

Helga’s “body” is a standard commercial serving robot, similar to those used in restaurants for delivering or clearing plates. The unique features lie in her “head”—a powerful computer with an integrated touchscreen and various sensors, which was developed entirely from scratch for this project. “Developing the methods for measuring pupil dilation and heart rate was new and particularly challenging,” says Graf. “We had to develop the sensor concept for the pupil test from scratch. We also had to train the robot so that it functions reliably in practice, even with patients’ widely varying anatomies and lighting conditions.” For example, during the pupil test, Helga must accurately detect the pupil and any changes in it. The darker the iris, the more difficult this is.

Sarah is ready for the pupil test. She looks at the screen, where two cars are driving side by side. “Now position your head so that both cars are driving in the middle of their lanes,” says Helga. A light flashes briefly. Helga uses a camera to record how the pupils change. She also checks—just as in a normal neurological examination—the children’s verbal and motor responses: She asks the children what animal they see on the screen. She also has them click on crying or laughing emojis on the tablet to show how they’re feeling and to test their movements.

For now, Helga is still a prototype. The first step is to determine whether she can reliably, safely, and in a child-friendly manner perform tasks such as monitoring patients with concussions. A doctoral candidate is monitoring approximately 120 patients as part of her dissertation at the Department of Pediatric Surgery in collaboration with the Institute for Nursing Science at LMU Medical Center. Among other things, vital signs are measured and compared using both human staff and the robot to identify areas where Helga still needs refinement. In addition, the study team is surveying children, parents, and medical staff to learn how they experience the robot’s use.

Even though the project is still in its early stages, there are numerous ideas for further applications: from the initial medical history intake upon admission to guided tours of the ward. “This is the future,” says Ward Manager Corina Tuch. She emphasizes the importance of carefully assessing which tasks a robot can take on in the ward’s daily routine—and how the team is receiving Helga. Her colleagues chose her name. And Helga even has a permanent spot on the ward: in the break room at the ward’s central station.

Concussion is one of the most common reasons for inpatient hospitalization among children in a pediatric surgery clinic. These children are then kept on the ward for 24 to 48 hours for monitoring. Nurses regularly check pupil response, heart rate, verbal and motor responses, and general well-being—initially every hour, then at slightly longer intervals. Each check takes five to ten minutes and involves the use of standardized assessment tools such as the Glasgow Coma Scale. To date, centralized bedside monitoring systems have been unable to adequately perform these tasks.

For a year, Dr. Birgit Graf and her team at Fraunhofer IPA have been fine-tuning the robot so that it can take on these tasks. For more than 25 years, the computer scientist has been working on the practical development of robotic assistance systems in the healthcare sector and in elder care. These systems are used in logistics and service, for cleaning, housekeeping, documentation, or lifting tasks.

Helga’s “body” is a standard commercial serving robot, similar to those used in restaurants for delivering or clearing plates. The unique features lie in her “head”—a powerful computer with an integrated touchscreen and various sensors, which was developed entirely from scratch for this project. “Developing the methods for measuring pupil dilation and heart rate was new and particularly challenging,” says Graf. “We had to develop the sensor concept for the pupil test from scratch. We also had to train the robot so that it functions reliably in practice, even with patients’ widely varying anatomies and lighting conditions.” For example, during the pupil test, Helga must accurately detect the pupil and any changes in it. The darker the iris, the more difficult this is.