A new team at the Fraunhofer IPA is developing digital laboratory systems to support industrial partners in planning and managing laboratory processes. This benefits companies in terms of quality assurance, economic efficiency and, notably, ergonomic improvements to the work place.
The Fraunhofer Institute for Manufacturing Engineering and Automation IPA in Stuttgart is responding to the digitisation trend in laboratories in its own way: on March 1st 2018, the IPA founded its new "Digital Lab Services" work group. With four full-time employees and seven students, the team led by Matthias Freundel was launched with the aim of advancing the establishment of digital labs, especially in companies.
"We've had increasing numbers of enquiries from customers who are developing digitisation strategies in the context of Industry 4.0. Manufacturers of pharmaceuticals and in-vitro diagnostics already have a high degree of automation of up to 80 percent – the goal now is to digitise the remaining 20 percent,” says Freundel. With his team of computer scientists and bioinformaticians, Freundel develops solutions for the software-based support of work that is still done individually and by hand. This basically concerns all work that does not make sense to automate. "If, for example, a pipetting process is to be repeated a hundred or a thousand times, a machine can do this better than a human, but there are also pipetting jobs that require a lot of flexibility and have to be done manually. However, the digital recording and evaluation of the pipetting steps could be used to optimise the pipetting process, which would also contribute to quality assurance.”
In practice, in one of IPA’s pilot projects, this means that the user’s actions are digitally recorded at a laboratory workstation equipped with sensors and a camera. For example, what angle are the pipette and sample tube being held at? How is the required equipment arranged in the work area, is this facilitating or hindering user actions? The results of such analyses are expected to help improve work processes and the working environment. They can also help ergonomically if it turns out, for example, that the user holds the pipette in a way that increases the risk of pain. "Ultimately, we always look at how we can assist the user in laboratory work without using robotics," says Freundel.
The IPA team develops its own systems for object and gesture recognition. "We tend to use cameras for high-resolution images and depth sensors for distance and shape recognition. The data collected is forwarded to a higher-level system including databases and then evaluated," explains Freundel. He acknowledges that at the moment it is difficult to identify some of the hand positions and materials. Intensive development work is still needed so that processes can be presented without data gaps. Enormous amounts of data accumulate, depending on which activities are analysed. Therefore, "big data" is also an issue that has to be addressed. Suitable software and hardware solutions need to be found so that data can be managed in a meaningful way and interpreted easily and quickly.
Freundel and his team also anticipate advantages with regard to physical laboratory planning. "One of our goals is to use augmented reality to place devices in a room, so we can analyse how much space laboratory staff have to move around the device and how the device fits into the room layout. Consequently, these tools will also help to plan new buildings better, and we will continue developing them this year."
As good as all this sounds in theory, there are still some hurdles as far as implementation of the tools are concerned. Freundel is well aware of this: "Many technologies sound good, but like augmented reality, they still have their limitations. For example, the field of view that can be viewed with data goggles is still relatively small, which causes problems as far as applications close to the eye and on the body are concerned. Other goggles with more computing power might be the solution." Digital speech recognition, for example, used for recording information, also gives rise to particular difficulties in laboratory settings. Not only does it have to master the technical language, but it also has to cope with a lot of background noises. Another important issue is sensor placement: which sensors should be placed where? Should they be attached to the employee or is it better to put them somewhere in the room? The team has set out to clarify the many small details so that they can offer industry partners smoothly functioning systems or jointly develop solutions adapted to individual requirements.
However, the researchers are not just looking at the laboratory itself: digital laboratory services also focus on storage. The IPA has started a research project to investigate how often which items are removed from the storage room and whether it would be a good idea to store them according to frequency of use. "It might be useful, for example, to store pipette tips and gloves together," says Freundel. This then leads into how the laboratory needs to be designed so that users have quick access to the things that they need most frequently.
Freundel and his team also address privacy and data security. "As data should not enable automatic tracing of users, we're aiming for generalised data capture. And it always depends on how digital services are used. For example, digital protocols are regarded as positive. If an employee spends an average of 20 percent of his or her time taking notes, he or she would have significantly more time to focus on the actual work if notes were digitally recorded," says Freundel. Another important point is data transfer. Will it still be possible to read the data in 20 years’ time, will the software still be available? And what about the encryption of data in the cloud? "All these issues will keep us busy and they are ones that we very much want to contribute to," says Freundel.
Freundel is certain of one thing: laboratory technician as a profession will not die out despite automation and digitisation. This is due to the variety of tasks and the many irregularities associated with experimental work - humans are better suited to specific tasks than machines. "In Germany, however, professionals are relatively expensive and well-trained staff are difficult to find. So we need to make better use of our skilled labour force, and digital laboratory services provide crucial help," says Freundel.