Publication: hospOS: A Platform for Service Robot Orchestration in Hospitals – Saving Time of Nurses With Robot Management
Authors: S. Schmidt, T. Greiler, S. Fischer, D. Sommer, F. Wahl
Published in: ICT4AWE 2024, Springer CCIS 2762, pp. 202–221 (2026)
DOI: 10.1007/978-3-032-10685-8_12
In brief
The nursing shortage is no longer a forecast – it is everyday reality. Germany is expected to be short around 500,000 nurses by 2030, while the number of people needing care keeps rising. At the same time, nurses report that they lack time for what matters most: direct care at the bedside. Service robots are seen as part of the answer – but how much they actually contribute has mostly remained a claim.
This is where the paper comes in. It introduces hospOS, a platform that orchestrates robots from different manufacturers centrally, and – for the first time – provides solid figures on time savings from real hospital operations. Rather than arguing from pilot demonstrations or laboratory scenarios, the evaluation rests on systematic observation in two active hospitals. That is the difference between a plausible idea and a verifiable statement – and it is the basis on which investment decisions can responsibly be made.
Why nursing time is so scarce
To understand the value of relief, it helps to look at how a nursing shift is actually structured. A substantial part of working time on a ward goes not into direct care but into supporting activities: walking through the building, fetching and delivering supplies, meals and samples, answering the same questions over and over. These tasks are necessary, but they require no nursing qualification. Every minute spent on them is a minute missing at the bedside – exactly where nursing expertise is genuinely needed.
It is this gap that makes the topic so relevant. Reliably offloading non-nursing tasks does not conjure staff out of thin air, but it redistributes the working time that exists toward the work only people can do. In a situation where additional staff is barely available, that redistribution is often the only realistic lever.
Why an orchestration platform?
Today's service robots are mostly closed, single-purpose islands: one robot, one task, one proprietary controller. Covering several functions quickly leads to expensive special-purpose systems that barely integrate with hospital IT. Our approach flips this: instead of an all-in-one machine, we combine existing, affordable robots through a shared layer. hospOS automatically assigns each task to the most suitable available robot – modular, vendor-independent and embedded into existing infrastructure.
The advantage of this architecture is not only the purchase price. An orchestration layer makes the hospital independent of individual manufacturers and their product cycles. New devices can be added and older ones replaced without rebuilding the entire logic each time. Tasks are dispatched centrally rather than triggered manually for each robot, and the system stays comprehensible and manageable for the hospital's IT. A collection of individual machines thus becomes a coordinated whole that aligns with what the ward needs – not the other way around.
Three use cases, tested in two hospitals
We installed hospOS in two rural hospitals and evaluated it across three use cases: telemedicine, transport and orientation.
The benefit is clearest in transport. A preceding observational study found that 77.15 % of all recorded transports (N = 1,629) are done by hand – often short, repetitive trips that pull nurses away from actual care. Robots are especially worthwhile for transporting meals and drinks: from an economic perspective, this yields a potential saving of roughly EUR 10,000 per year and hospital – enough for a low-cost transport robot to pay for itself within a year. That most transports are still carried out manually is the decisive finding here: this is not about edge cases but about routine operations that can be automated systematically.
Demand in orientation is just as real: in a seven-day observation at the reception (N = 1,499 requests), most concerned visitors (51.3 %) and patients (38.5 %) – routine questions about directions and room numbers that a robot can handle reliably. Such requests follow clear patterns and can be answered in a standardised way without tying staff to the reception desk. For telemedicine, the paper outlines the savings from reduced travel time for on-call physicians; precise quantification is the subject of a follow-up study.
Field data instead of promises
The real contribution of this work lies in its methodology. Many claims about the value of service robotics rest on assumptions, vendor figures or one-off, unrepresentative demonstrations. The numbers behind this paper, by contrast, come from structured observation in the live operation of two hospitals – with traceable sample sizes and a clear separation between what is already proven and what remains to be investigated.
This restraint is deliberate. Where the evidence is still thin – as with telemedicine – it is named as an open question and assigned to a follow-up study, rather than filled in with optimistic estimates. It is precisely this clarity that makes the results usable for hospital management: they can decide on the basis of what was actually measured, instead of relying on promises.
What this means for hospitals
The message is not "the robot replaces the nurse" but rather: deliberately offloading repetitive, non-nursing tasks measurably wins back time for care. What matters is not the individual robot but the layer above it that integrates it sensibly into hospital operations.
This platform idea is where Athegus begins. With hospOS and the vendor-independent middleware Axiona, we turn individual robots into an orchestrated system – the software layer between the robots and the hospital world. The field data published here is the scientific foundation for it.