Doctoral students at the University of Stuttgart are researching ways to get robots to handle things with a more gentle touch, which will allow them to operate in completely new fi elds. Among other things, they are taking inspiration from farms in New Zealand.
Robots can already do many things: they can paint cars, cut out panels, glue glass panels in place and react to specifi c commands. However, there are also many things that robots are still incapable of: they have trouble picking fruit, they cannot fi llet a side of pork and are unable to recognise human tissue. In a nutshell: when hard robots come into contact with soft materials, problems ensue. Oliver Röhrle and Alexander Verl are just two scientists among many who want to change that. Oliver Röhrle, Professor of Continuum Biomechanics and Mechanised Biology at the University of Stuttgart, heads up the International Graduate College of Soft Tissue Robotics.
The team uses computer simulations to fi nd out how robots could learn how to handle soft materials. The project receives funding from the German Research Foundation (DFG) and involves a collaboration with the University of Auckland in New Zealand, a long-term partner of the University of Stuttgart. This research was also at the centre of the New Zealand-Germany Science Circle, Stuttgart, an event held in March 2017 to celebrate the 40th anniversary of German-New Zealand Science Cooperation Agreement.
At least ten scientists in Germany and another ten in New Zealand are working on the relevant research at any given time. The college operates on an interdisciplinary basis. “We’re exploiting synergies that arise from the research specialities of both universities”, Röhrle explains. In addition to simulation technologies, these include biomedical sciences, robotic engineering and cyber-physical systems.
Simplifying Abattoir Work
One of the ideas for the project originated in New Zealand which is dominated by its agricultural industry. The country is home to a population of just four million people, yet some 30 million sheep are slaughtered there each year. There are plans to automate the meat processing industry. However, the plan is currently coming up against certain technical limitations, as identifying and sorting various cuts of meat such as hearts, liver and kidneys is no easy task for a robot. The situation is similar when it comes to harvesting apples or kiwis. To date, all attempts to automate this task have failed. Robots are apt to squash the fruit they pluck, which makes them worthless as an export product.
Robots at the Operating Table
The same problem is encountered in medical engineering. Whereas robots can carry out operations with a far greater degree of precision than humans and already perform millions of operations on humans, there are risks involved: any uncontrolled movement by the robot could injure the human patient. Yet, the future applications in the medical engineering are enormous. For example, it has long been possible for doctors to differentiate between a tumour and healthy tissue using computed tomography scanning (CAT), and operations are performed based on this information. “A robot may be able to do that better in the future”, says Professor Alexander Verl, co-spokesman of the graduate college and head of the Institute for Control Engineering of Machine Tools and Manufacturing Units at the University of Stuttgart. But for that, scientists would fi rst have to know much more about the interface between robots and soft tissue.
The researchers base their work on practical questions such as these. “For example”, says Röhrle, we consider how we can control the robot and how we can simulate the materials with which it works.” This fundamental knowledge will simplify future research. “Only in a later step will we address the question as to how the results could be transferred to the real world.”
The graduate college is also carrying out research into sensor technology, which refers not just to cameras or contact sensors, but to much more complex combinations. Some ideas are being taken from recent developments in exoskeleton technology, for example, i.e., robots that are worn on the body like a corset, which function as strength boosters. Minimum impulses from the human musculature control the actuators which amplify the movements of the wearer.
Comprehending Soft Tissues
The electrical signals of the muscle fi bres can be recorded as a so-called electromyogram. Another project is about developing a sensor system that can also analyse these impulses so that they can then be used as control signals. The fi ndings from this endeavour are helping to expand our basic knowledge about soft tissues. The results could be benefi cial wherever there is an interface between man and machine. Exoskeletons could revolutionise an industrial company’s production processes. For example, factory workers who need to lift heavy components can sometimes do so using their own muscle power, but sometimes need help. Measurable impulses in the worker’s muscles could be used to activate an electric lifting aid that would provide a boost when needed. This would enable older workers or those weakened by illness to remain in gainful employment.
The initial funding phase of the project is scheduled to last four years. The subsequent phase of the same length could deal with another aspect, namely the material from which the robot is constructed; by no means does this always have to be metal. Machines made of malleable plastic are also conceivable. Tiny channels fi lled with gas or liquids could run through this material much like the blood vessels and nerves in the human hand. Robots made of this material could then apply a gentle touch and maybe even develop a kind of tactile sense. Heimo Fischer