‘When people ask me what robotics is, they often think of surgical robots or little robots in the corridor,’ says Papenburg. ‘That’s true, but it’s also much more than that.’
Bernke Papenburg is the innovation and robotics manager at Rijnstate. In her work, where she and her team carry out dozens of projects relating to 3D, sensing, smart technology, robotisation, robotic surgery and AI, the focus is precisely on the interconnection between technology, data and care processes.
From sensors to action: the heart of robotics
According to Papenburg, robotics should be understood within the Sense–Compute–Act framework. This consists of: Sense, sensors that gather information from the environment; Compute, systems that process that information; and Act, actuators that carry out a physical action.
“Those sensors are actually a system’s artificial senses,” she explains. “And what you do with them happens in the computing stage. That’s where the power of AI lies.”
And what’s important is data. ‘Only structured, clear data can really drive this technology forward.’
AI: much broader than ChatGPT
There are also many misconceptions about AI. Many people immediately think of Large Language Models such as ChatGPT or Copilot. But according to Papenburg, that is only part of the story.
‘There are also neural networks, machine learning and predictive models,’ she says. ‘These recognise patterns in data and learn from them. You see this, for example, in diagnostic support or image analysis.’
AI and robotics therefore overlap significantly, particularly in the computing stage. ‘Together, they form the basis for truly intelligent systems.’
From generative AI to physical AI
Developments are happening at breakneck speed. Papenburg refers to a vision in which AI develops in stages:
- Generative AI (such as current language models)
- Agentic AI: AI systems that work together as independent ‘agents’
- Physical AI: the combination of AI and robotics
‘With agentic AI, systems communicate with one another and manage tasks independently,’ she says. ‘But physical AI is where it really gets interesting. That’s where AI and robotics truly converge.’
The combination
Robotics without AI often gets bogged down in fixed scripts. Papenburg gives a striking example: a robot at Disneyland that simply fell over when it ‘froze’.
‘Without AI, you have rule-based systems,’ she says. ‘If they don’t know what to do, they simply stop.’ AI adds flexibility. But without robotics, AI remains confined to digital environments. ‘It’s only when you combine them that you get truly intelligent applications.’
Applications in healthcare: from administration to the operating theatre
At Rijnstate, the focus is on various aspects of healthcare. This translates into several areas:
Administration and registration
‘There’s an enormous amount of work here,’ says Papenburg, ‘and very little job satisfaction.’ AI helps, for example, with summarising patient records, automatically preparing orders and generating letters. In addition, there are digital assistants that transcribe conversations into text and produce summaries. ‘This allows doctors to give more attention to their patients.’
Monitoring and patient care
One example is a smart mat placed under a patient that detects movement via air pressure. ‘The system can detect whether someone is getting out of bed, even before it happens.’ This provides insights such as: what is the patient’s level of activity, are they restless or, conversely, moving very little, and this gives you early warning signs of risk.
Medication and logistics
From pharmacy robots to smart medication trolleys: technology helps reduce errors and speed up processes. Think of smart support for dispensing medication and barcode scans for verification. ‘You don’t immediately do away with the four-eyes principle, but you are taking steps towards greater safety and efficiency.’
Diagnostics and treatment
AI is increasingly being used to support medical decision-making. Examples include: detecting fractures on X-rays, analysing tumour growth, interpreting ECGs and EEGs, and virtual surgical planning. Medical equipment such as CT and MRI scanners is also becoming smarter. ‘Scans can be performed 25 per cent faster and with better quality.’
Learning from what doesn’t work yet
Not everything is a success. Papenburg cites a UV disinfection robot that struggled with variable layouts. ‘Beds weren’t always in the same place. That’s when you realise that robotics alone isn’t enough.’ A social robot in the paediatric ward also regularly got stuck. ‘Just like that robot at Disneyland. That’s when you realise: they’re not smart enough yet.’ Nevertheless, experiments like these provide valuable insights. ‘It’s all about learning by doing.’
3D and digital twins: simulating the operation in advance
One of the most promising developments is the use of 3D models and digital twins. This involves creating a digital copy of a patient. ‘You can test in advance where to make the incision, how much tissue to remove and what the best outcome will be,’ says Papenburg.
This information can then be viewed in mixed reality, projected into the operating theatre and combined with robotic surgery. ‘That’s really where those worlds come together.’
The major challenge: implementation
Technology is advancing faster than hospitals can keep up with. The main challenges are the complexity of care processes, strict regulations (AI Act, MDR, GDPR) and issues surrounding responsibility and liability.
And above all: the role of people – there is no question of 100% autonomous machines. ‘We always work with a human in the loop.’ Papenburg does not believe in robots that replace people one-to-one. ‘The point is that people and technology reinforce each other. Robots, for example, take over repetitive tasks, physically demanding work and administrative duties. Meanwhile, people retain empathy, creativity and the ability to make complex decisions. Let the robot do the tedious jobs, so that healthcare professionals have time for their patients.’
Start with the ‘why’, not with technology
Papenburg explains: ‘We start every project with one question: what are we solving? Not: we have a great piece of technology. But: where is the problem?’ This is followed by an iterative process of testing, learning and adapting. ‘You don’t get it right first time.’
Looking ahead
A ten-year strategy? There isn’t one. ‘Developments are moving too fast. However, the direction is clear: greater integration of AI and robotics, a focus on labour-saving measures, and practising experimentation and learning. Don’t try to make it too perfect,’ says Papenburg, ‘just get started, keep it small and learn.’
‘Embrace the complexity,’ emphasises Papenburg. ‘Look beyond your own organisation, learn from one another and share knowledge.’ Because ultimately, technology alone won’t solve the problem. ‘We have to do it together.’
Wings Programme: Health Valley
This article is a summary of an online webinar delivered by Bernke Papenburg as part of the Wings programme.
Read more here about this programme for healthcare professionals and healthcare entrepreneurs.
