Robo Lab

What issue did the project tackle?

Social robots are increasingly finding their way into various areas of everyday life and the professional world. Artificial intelligence means they can move autonomously, make contact with, and react to, people. In doing so, they simulate human emotions and feelings. This makes them of interest in applications such as retirement and care homes, hotels, shopping centres and education. Our case study (see Tanner, Burkhard & Schulze, 2019) indicates that their deployment in everyday life remains fairly limited. Following an initial «euphoric phase» with its focus on entertainment, which attracted public attention in e.g. department stores and reception areas, noticeable disillusionment has now set in. Companies with such robots are confronted with having to find specific applications that are also economically profitable. The FHNW Robo Lab strategic initiative (2018-2020) sought to identify tasks in which social robots can add value and the technical functionalities necessary to do so. The aim was to help industry partners identify potential applications for robots with social functions and find appropriate operating models.

What results and impacts did the project achieve?

Field and laboratory studies involving a variety of social robots such as Pepper, Nao, Einstein and P-Rob enabled the gathering of knowledge and experience in the following areas: the technological state of development and distribution of social robots, the programming methods used, the public’s acceptance, and ethical issues. Social robots are generally accepted by various groups of users, such as care home residents, hotel employees and guests, library visitors and students. Pilot studies in four areas were carried out with industry partners, the aim being to investigate in more detail the opportunities and risks of deploying robots, and their specifications in respect of their function and design:

Social robots in healthcare

In our estimation, the healthcare sector holds clear potential for the use of social robots, for example to support nursing staff. Our projects evaluated the technical feasibility and benefits of a number of deployment scenarios. Social robots can, for example, support staff in administrative tasks such as registering patients or assume assistive tasks such as activating patients. Effective deployment of social robots in a healthcare setting calls for the active involvement of patients and people in need of care.

Pepper in the healthcare sector.

Social robots in education/training and continuing education

As things stand, social robots deployed in a higher education setting are primarily used on technical/computer science courses, such as for learning programming languages. That said, they are equally suitable for interactions or as teaching assistants. Moreover, from an educational perspective, it appears that robots can themselves be made the object of the teaching for the purposes of reflection. For instance, in a number of courses at the FHNW’s schools, a discourse on digitality, e.g. on (information) ethics, was initiated through the presence of, and interaction with, robots. This is because interaction with technologies going forward calls for a critically reflective media education in the tertiary education sector.

Pepper in pilot trials at FHNW schools.

Social robots in industrial production

The current generation of social robots is still not advanced enough for typical production tasks. This is because, compared to industrial robots, they are not good at handling objects. Social robots such as Pepper and Nao use motion mainly for gesture, which places lower demands on accuracy than following exact motion trajectories in production. In future, however, we can expect to see hybrid solutions such as industrial robots supplemented by social skills that enhance their user-friendliness. In a bid to identify possible applications, a test rig was set up in the FHNW Robo Lab with input from partners with an industrial production background.

Social robots in public spaces

The public realm offers social robots more and more scope for their deployment, such as in hotels and shopping centres. For instance, we implemented robots to act as first point of contact in a company’s reception area and as an information guide in a library. Applications such as these call for a step-by-step development process involving several test cycles in order to give equal weight to the technical, organisational and social requirements. Also, for social robots to be accepted, the technical possibilities have to be communicated transparently and the robots must undergo continuous development.

>>Video of Pepper the library robot helping with book selection.

Pepper in the FHNW libraries at Muttenz and Brugg.

Social robots in virtual reality

Using the virtual realm, future deployment scenarios of social robots can be realised in advance and tested for usefulness, acceptance and technical necessity. The FHNW Robo Lab developed a virtual environment capable of accommodating multiple representations of the real world. In scenarios involving e.g. Pepper the robot as a guide in a virtual museum, we were able to show that the capabilities of the social robot can be tested in a «customer contact» setting and input gathered for its subsequent programming.

The pilot tests showed that sustained acceptance depends very much on how the specific application-related capabilities of such robots are designed. Due to the time and effort involved in the if-then-else programming of the current generation of social robots, their deployment is best restricted to scenarios where the same interaction is repeated continuously with large numbers of people. The greatest potential benefits resulted when robots were incorporated within teams of employees and their tasks and applications were developed jointly. Methods showing promise include user-centred development of application scenarios, scenario-based process and interaction design, and using VR to pre-evaluate processes.

Examples of virtual testing of social robots.

What distinguished the interdisciplinary collaboration in the project?

The development and introduction of successful deployment scenarios involving social robots depend on an interplay of factors that require interdisciplinary cooperation. These include, among other things, the programming of the interaction behaviour alongside speech input and output, the sensorimotor movement design of the robot, and the recognition and interpretation of the environment as well as of people and faces. Recognition of the usefulness of robots for both users and employees is also a prerequisite for the emergence of acceptance. In the strategic initiative, the interdisciplinary interweaving of specialist and experiential knowledge from computer science, engineering, education and applied psychology was key to understanding the human-robot interaction and central to the robots’ integrated design. The following factors were instrumental to the success of the interdisciplinary cooperation:

• Development of a mutual, shared understanding of goals through serious gaming vision workshops
• Intensive familiarisation with the disciplinary concepts and maxims of action on the basis of jointly conducted interdisciplinary case and laboratory studies
• Joint interdisciplinary conception, realisation, testing and implementation of pilot tests in the field
• Interdisciplinary staffing of teams at the management and operational levels
  

Cooperation with industry and research partners

Over the past three years, the FHNW Robo Lab strategic initiative has, we are pleased to note, involved numerous partners from science and industry. We are grateful to the following companies and institutions consisting of manufacturers, users, associations and research partners from Switzerland and abroad:
Avatarion, raumCode, F&P Robotics, JinnBot, Regionales Pflegezentrum Baden, SATW, Zentrum für Bilddiagnostik, SWITCH, ABB Technikerschule, Angestellte Schweiz, metrobasel, Universität Basel, Solothurner Handelskammer, Universität Bielefeld, SVEB, Inselspital (Bereich Innovation Pearl), Universität Basel Bereich Psychiatrie Chronobiology und Bereich Radiologie, Projekt H.E.A.R.T (BMBF) der Universität Marburg, Universitätsspital Zürich, Bereich Pflege, Rehab Basel, TU Aachen, Universität Bielefeld, EAO Olten, Die Post (Connecta Bern), Bundesamt für Sport BASPO, ZHAW Winterthur Institut für Pflegewissenschaft, Aargauische Industrie- und Handelskammer, Handelskammer beider Basel.

Building on the experience gained, the FHNW Robo Lab interdisciplinary expertise network continues to offer support in, e.g. programming social robots, clarifying legal and ethical aspects, and designing introduction processes for companies and organisations. Collaboration can take a variety of forms. It ranges from partnerships in publicly funded projects and bilateral cooperation in the context of applied research or services all the way to delivering input lectures and implementing student projects.