TAO_MRI

Mensch-Roboter-Interaktion (Human-Robot Interaction; in German)

The full title of the project is Mensch-Roboter-Interaktion (Human-Robot Interaction; in German), partly funded by the Technology Alliance of Upper Franconia (TAO).

Description

The goal of the TAO project Human-Robot Interaction is the research and development of innovative intuitive interaction concepts, which enable users to communicate with robots even without technical prior knowledge. The application scenarios are manifold and range from robotic helpers in private households to industrial robots in small and medium-sized enterprises. The robots should, for example be able to recognize objects independently and move them when the user requests it. We develop innovative intuitive interaction concepts, illustrate those concepts using prototypes and evaluate them with future users. In addition, we create design recommendations in order to support the situation-dependent development of forms of interaction. The user-centred approach enables intensive user participation in all project phases.

Approach

On the part of robotics, the methods used are based on the concepts of programming by demonstrating (PdV), which are already being investigated in other projects (e.g. INTROP) in Bayreuth. For example, the user intuitively takes the robot "by the hand" and guides it through the desired task. A declarative and an imperative approach are available for this purpose. So far, however, feedback from the robot to the user has not been considered – an aspect that we examine in this project. On the human-computer interaction side, we use the procedures of human-centred computing developed in Bamberg, which involve the end users in all phases of the design, implementation and evaluation of the new interaction concepts. The interface between user and robot will provide both autonomy and adaptivity of the robot to the users.

Activities

• Interview study and experimental study that provide an extensive insight in the requirements, expectations and use cases for Human-Robot Interaction
• Prototyping with concluding user studies
• Expansion stages of the prototype with interaction concept
• Evaluation and demonstration of the results
• Documentation and publication of the insights

Current State and Results

The project MRI has six project phases with individual research and prototyping cycles.

1. Requirements analysis
In a user-centered approach, we collect ideas and wishes regarding simple and intuitive programming of robots in interviews with individuals and in focus groups with future users. From this, we derive and prioritise requirements with regard to complexity and urgency.

2. Basic interaction concept
The interaction concept should cover the entire range from easy entry for beginners to great power for experts in robot programming. Based on the requirements, we develop an abstract model of user interaction, which offers an input and output comprehensible for end users through a suitable interface metaphor. Its usability is defined here according to the ISO 9241 standard as efficiency (accuracy and completeness of the achievement of objectives by the users), efficiency (resources which the users must use to achieve the objectives) and satisfaction (freedom of the users from inconveniences as well as the positive attitude of the users when using the software).

3. Prototype realisation and evaluation
We demonstrate a first integrated system for commanding a simple command, which enables parts of the interaction concept with the actual programming of a robot. We summarize the feedback from the user tests by a formative evaluation and, if necessary, convert it into requirements for the extension and revision of the interaction concept.

4. Expansion stages of the prototype with interaction concept

• Picking a complex task
  The users should be able to assign a complex task to the robot, which consists of several simple sequential operations that are executed independently by the robot.||
• Programming of case distinctions
  Users should be able to teach the robot different sensor-based reaction patterns for different situations.
• Programming repetitions
  Users should be able to instruct the robot to repeat one or more simple or complex operations.
• Modification of the program
  Users should be able to interactively change or correct an already created program.
• Calling subprograms
  Users should be able to reuse already created program parts (subprograms) in a complete program without having to recreate the subprograms.

5. Final evaluation
In the final summative evaluation, we evaluate the overall system with the users in terms of usability. We use quantitative methods such as user tests and eye tracking in the laboratory of the Chair of Human-Computer Interaction, as well as qualitative methods such as, in particular, heuristic evaluation in the field.

6.Demonstration and Documentation
We set up an overall demonstration in which the performance of all integrated forms of interaction can be clearly demonstrated in one or a few tasks.

The outcome of this cooperative project provides insights that allow conclusions about the needs and expectations of users in the household context. Thus, our research results emphasise that the use of robotical support in household scenarios always require the individual consideration of individual circumstances to create the best user experience. Future household robots need to be both pre-equipped with fast possibilities of recognising and interacting in their environment with users, and be easily personalisable to respond to user-specific needs.

Partners

Human-Computer Interaction Group, University of Bamberg
Chair for Applied Computer Science III (Robotics and Embedded Systems), University of Bayreuth

People

Prof. Dr. Tom Gross
Prof. Dr. Dominik Henrich
Michael Gradmann
Sascha Herr
Pia Marquart
Hesham Omran

Funding

Technology Alliance of Upper Franconia

Further Information

Publications

Herr, S., Gross, T., Gradmann, M. and Henrich, D. Exploring Interaction Modalities and Task Allocation for Household Robotic Arms. In Extended Abstracts of the Conference on Human Factors in Computing Systems - CHI 2016 (May 7-12, San Jose, CA). ACM, N.Y., 2016. pp. 2844-2850. (ISBN: 978-1-4503-4082-3/16/05).

Omran, H. and Gross, T. An Explorative Study on Requirements for Ambient Displays Presenting Mood Awareness. In Mensch & Computer - 16. Fachuebergreifende Konferenz fuer interaktive und kooperative Medien - M&C 2016 (Sept. 4-7, Aachen, Germany). Gesellschaft fuer Informatik e.V., Bonn, 2016. pp. 1-4. doi:10.18420/muc2016-mci-0227.

Herr, S. and Gross, T. Multimodale Mensch-Roboter-Interaktion im privaten Haushalt (Multimodal Human-Robot-Interaction in Private Households; in German). In Mensch & Computer - 15. Fachuebergreifende Konferenz fuer interaktive und kooperative Medien - M&C 2015 (Sept. 6-9, Stuttgart, Germany). Diefenbach, S., Henze, N. and Pielot, M., eds. Walter de Gruyter GmbH, Berlin/Boston, 2015. pp. 347-350. (ISBN: 987-3-11-044334-9).

Gross, T. Human-Centred Collaborative Computing (in German). In Proceedings of the Usability Day - Digitale Medien in Arbeits- und Lernumgebungen - uDay XIII 2015 (June 12, Dornbirn, Austria). Jost, P. and Kuenz, A., eds. Pabst Science Publishers, Lengerich, Germany, 2015. pp. 30-39. (ISBN: 978-3-95853-067-6).