Future Lab – The Automated Laboratory of the Future
Quo Vadis? What are the next developments of analytical laboratories?
- Fig. 1: Robot based system for fully automated sample preparation and determination of metalls.
- Fig. 2: Sample vials in racks in standardized MTP format.
- Fig. 3: Camera based tube detection.
- Fig. 4: Mobile robot for sample transport in complex automation systems.
- Fig. 5: Dual-Arm-Robot using a standard laboratory pipette (left) and delivering samples to a GC autosampler (right).
- Prof. Dr.-Ing. habil. Kerstin Thurow is Director of the Center for Life Science Automation at the University of Rostock (D). She graduated in 1995 at the Ludwig- Maximilians-University Munich and received her habilitation in 1999 in the field of measurement and control engineering at the University of Rostock. Since 1999 she has been working as a professor in the field of automation technology / Life Science Automation. The research areas include the fields of robotics, mobile robotics as well as data handling and management. Professor Thurow has published more than 200 publications. She is a founding member of the Academy of Sciences in Hamburg and a member of the German Academy of Engineering Sciences (acatech).
- Priv-Doz. Dr.-Ing. habil. Heidi Fleischer is head of the research area “Life Science Automation - Processes” at the Center for Life Science Automation of the University of Rostock. She studied Information Technology / Computer Engineering at the University of Rostock and received her Ph.D. here in 2011. Fleischer earned her habilitation in 2016 and received the Venia Legendi in the field of “Measurement and Automation Technology”. Her field of research includes the automation of sample preparation processes for analytical investigations.
Classic analytical laboratories today still have a comparatively low degree of automation. However, increasing numbers of samples and increasing cost pressure also make automation indispensable in this area. The complexity of analytical processes with regard to the process steps and the required lab ware requires new concepts and approaches.
Automated processes have long been established in the field of bio screening and the pharmaceutical industry. The situation is quite different in the field of classical analytical laboratories in environmental analysis, quality control or the food industry. Apart from the highly automated analyzers for analytical measurements, extensive manual activities still dominate here. This is a result of the often more complex process sequences, which, in contrast to biological samples, involve extensive sample preparation prior the analytical measurement (e.g. digestions, dissolution of solids, use of aggressive or volatile media, working at higher temperatures and pressures, complex separation of matrix and analyte, etc.).
On the other hand, there exist no standard for the sample vessels in this area. In contrast to the microtiter plate (MTP) as standard in the biological-pharmaceutical sector, a large number of vessels are used in analytical measuring technology, which differ in particular with regard to the volumes, vessel shapes and materials used. But even in the area of classical analytical measurements exists an increasing pressure for higher automation. Reasons for this are the increasing number of samples due to increasing regulations, increasing cost pressure, but also the growing shortage of skilled personnel. While automation in industrial production is usually dedicated to high sample amounts and uniform processes, analytical solutions are more likely to be flexible solutions that can be easily adapted to changing requirements. This makes automation solutions of interest to small and medium-sized companies, which are often confronted with changing problems and fewer sample numbers. For this reason, the field of automation of analytical methods has a great potential for development beyond the classical biological-pharmaceutical processes and will develop strongly in the coming years.
New Concepts for Tube Handling
While in the biological-pharmaceutical field with the microtiter plate a standard is available, which allows a simple programming, this is not the case in analytical measuring technology and the associated processes.
The large number of different vials and tubes used in analytical measurement technology requires new strategies for their handling. Here are different ways conceivable. On the one hand, there is the possibility of a consistent single vessel handling. This allows maximum flexibility since individual samples can be processed with different processes. At the same time, this concept places the highest demands on the programming, control, and scheduling of the systems.
On the other hand, it is also possible to combine samples into groups, which are arranged in the MTP footprint. This allows the use of classical automation systems such as pipettors, etc., which are aligned to the MTP format. Furthermore, higher throughputs can be achieved, since expensive transport steps of individual samples can be omitted (Fig. 2).
For the delivery of vials and tubes to the automation systems and to the automated sub-stations, suitable procedures must be developed. In general, systems can be used that supply the samples via conveyor belts. For identification of the samples, labels with barcodes or RFID (radio-frequency identification) can be used, whereby in the first case orientation-independent identification methods are required in order to minimize the error rates of the automation systems. Other challenges include the error-free identification of the vessels used, as well as the reliable determination of the solutions contained and the detection of phase boundaries (Fig. 3). While this is already possible with colored liquids, colorless liquids are still a challenge if, due to the process specificity, e.g. Capacitive measurements inside the sample tubes cannot be used. The degree of difficulty can increase, if e.g. pellets must be detected in tubes and separated in a targeted manner. In all cases, camera-based methods can be used. Here, the development of suitable image processing algorithms for the detection of volumes and phases as well as the feedback of these measurement data into the automation system (for example for the determination of immersion depths of dosing cannulas) are essential core tasks of the upcoming developments .
Use of Dual-Arm Robots
Automation of processes has often been limited in the past by requiring accurate protocols in highly regulated areas. Classical automation requires often a change in the standardized processes, e.g. by changing the dosing process from a manual piston-stroke pipette (e.g., Eppendorf) to an automated liquid handler. The comparability of the procedures is then no longer given. In addition, extensive re-validation of the now automated procedures is required.
With the introduction of dual-arm robots, these problems can be eliminated and automation solutions offered for such areas. Due to the two arms and the high number of degrees of freedom, these systems are able to carry out laboratory processes in analogy to humans with the identical laboratory equipment and process steps. This results in a 1:1 automation, a completely identical representation of the manual processes in the automated system. Due to the human-like movements even the integration of laboratory devices without interfaces, such as ultrasonic baths activated by switches and knobs are possible. In addition, the automatic handling of measuring systems such as gas and liquid chromatographs and mass spectrometers has hitherto been a problem since the systems are usually not designed for robot-based operation. Here, too, the use of dual-arm robots is a possible alternative  (Fig. 4).
Numerous manual processes, such as pipetting requires cooperation between both arms. This coordination of the two robotic arms is one of the existing challenges in transferring such processes to the robot. Furthermore, due to the limited working range of the dual-arm robots, optimal, collision-free path’s must be planned. This is usually done through teach-in procedures. However, recent developments also rely on computer-aided simulation methods. Another problem to be solved to allow a really 24/7 operation of the robots are intelligent methods of sample and lab ware delivery to the system.
Full Automation vs. Intelligent partial Automation
Full automation is generally the goal of all automation efforts, if this is economically viable. In the field of analytical measurement technology, there is the additional problem that sub-processes cannot be integrated into complex systems without considerable effort due to environmental conditions and safety requirements. For example, e.g. microwave digestions due to the production of toxic gases require appropriate safety precautions such as working under special hoods. In order to achieve full automation even in complex processes of this kind, the use of mobile robots is an obvious choice. These can take over the transport of samples and lab ware between the partial automation systems. Current and future research areas include questions of navigation and position determination of the used mobile robots as well as methods of collision detection and avoidance. Here are different methods such as laser scanners, UV or IR detectors can be used. In particular, strategies for accurately grasping and placing samples by the mobile systems are important to ensure safe, contamination-free transportation. This requires extensive descriptions of the kinematics of the robot arms as a basis for their programming. A further development consists in the expansion of the scope of activity of mobile robots from pure transport functions to manipulation possibilities of the samples (integrated robotics) .
Data Handling and Management of Automation Systems
The main goal of automation is to increase throughput in analytical processes. To prevent bottlenecks arising in the evaluation of the collected data, suitable automated data evaluation methods are required. In order to achieve the highest possible platform independence, web-based solutions are the means of choice here. The evaluation of the measurement results must also allow a combination of different measurement data of individual samples . For the analytical measurement of unknown samples, this also includes the extraction of relevant information from the different individual measurements and their combination for the unique identification of compounds.
Another problem is the management of the automation systems. If a complete automated system is not possible, but different sub-automation systems (possibly including mobile robots as transport systems) must be coordinated, the use of hierarchically organized workflow management systems is required. This requires new workflow control solutions that allow the integration of the wide range of heterogeneous automation IT systems with the mobile robot control system. For workflow planning, a data infrastructure is provided that supports process modeling and execution in terms of master and process data management and data access. By using a graphical planning editor, flexible workflow modeling is made possible from an ergonomic point of view. Optimized strategies for interpreting workflow models need to be developed for workflow control. Furthermore, the implementation of suitable scheduling methods for the workflow execution is required in order to enable an optimally use of the existing automation systems with a high degree of flexibility .
The analytical laboratory of the future will be characterized by a much higher degree of automation. In addition to fully automated systems that can take over complete process sequences, distributed solutions with automation islands will also be used if this makes sense due to economic or safety considerations. Robots, even in mobile form, are increasingly becoming supporters of laboratory personnel in the transport and manipulation of samples and lab ware. This enables higher throughputs with high flexibility at the same time. Automation systems will also become of interest to small and medium-sized companies.
Prof. Dr.-Ing. habil. Kerstin Thurow1, Priv.-Doz. Dr.-Ing. habil. Heidi Fleischer2
1 Center for Life Science Automation, University Rostock. Rostock, Germany
2 Institute of Automation, University Rostock, Rostock, Germany
Prof. Dr.-Ing. habil. Kerstin Thurow is Director of the Center for Life Science Automation at the University of Rostock (D). She graduated in 1995 at the Ludwig-Maximilians-University Munich and received her habilitation in 1999 in the field of measurement and control engineering at the University of Rostock. Since 1999 she has been working as a professor in the field of automation technology / Life Science Automation. The research areas include the fields of robotics, mobile robotics as well as data handling and management. Professor Thurow has published more than 200 publications. She is a founding member of the Academy of Sciences in Hamburg and a member of the German Academy of Engineering Sciences (acatech).
Priv-Doz. Dr.-Ing. habil. Heidi Fleischer is head of the research area “Life Science Automation - Processes” at the Center for Life Science Automation of the University of Rostock. She studied Information Technology / Computer Engineering at the University of Rostock and received her Ph.D. here in 2011. Fleischer earned her habilitation in 2016 and received the Venia Legendi in the field of “Measurement and Automation Technology”. Her field of research includes the automation of sample preparation processes for analytical investigations.
 Thurow, K., Junginger, S. und Roddelkopf, T.: Biospektrum 23(5), 531-534 (2017)
 Fleischer, H.; Drews, R., Janson, J., Chinna Patlolla, B. R., Chu, X., Klos, M. und Thurow, K: Journal of Laboratory Automation 21(5), 671-681 (2016)
 Thurow, K. und Liu, H. in: Mobile Robotics: Principles, Techniques and Applications. Nova Science Publisher (2015), 1-23
 Adam, M., Fleischer, H. und Thurow, K.: SLAS Technology 22(2), 186-194 (2017)
 Gu, X., Neubert, S., Stoll, N. und Thurow, K.: IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, 156-161 (2016)