Multi-Agent-Cooperation

The Multi-Agent-Cooperation research group investigates interaction processes between a human agent and one or multiple cooperative partners. These partners can represent other humans as well as technological devices such as robots, vehicles or aircraft. In the context of cooperation, it is important to balance partially competing individual and shared goals of the parties involved, which can be achieved by diverse mechanisms of communication. Metrics are to be developed to allow for the performance-oriented assessment of cooperation across multiple domains.

With regard to robots, effective strategies are required for both locomotive and evasive actions by designing legible and predictable movements. A robot’s motion can implicitly convey intentions to the human agent, who, visually perceiving and interpreting the respective information, reacts by himself performing certain movements within the shared space. The adequate design of movement strategies thus fosters the achievement of both individual and shared goals. In general, acceptance and trust of the human operator have to be considered in addition to the interpretation of human movements.

Concerning automated vehicles, similar mechanisms apply to the design of driving behavior. To ensure smooth procedures, driving strategies must be unambiguous and easily recognizable to other traffic participants such as human drivers, cyclists, and pedestrians. Accordingly, methods have to be defined for the design and evaluation of unequivocal and safe trajectories.

Also within vehicles, the progressive automation of driving functions renders new ways of cooperation between the two agents man and machine both possible and necessary. Depending on the context, central operating elements such as the steering wheel can directly be controlled by both cooperation partners. Consequently, it has to be obvious to both driver and automation who is in charge of lateral control at any time. To this end, the employment of haptic cues on the steering wheel is investigated as a means to convey information regarding for example the direction or certainty of a maneuver.

In civil aviation, pilots are superior to automated systems due to their teamwork skills and their adaptability to new and complex situations. It is therefore essential to structure communication and resulting coordination within the pilot group, which requires necessary training as well as objective and valid assessment criteria. As the relationships involved frequently cannot be modelled linearly, we are currently working on non-linear metrics. The latter allow to evaluate the adaptation of communicative and coordinative patterns and can serve as a basis for additional analyses.

Publications

Bengler, K., Zimmermann, M., Bortot, D., Kienle, M., & Damböck, D. (2012). Interaction Principles for Cooperative Human-Machine Systems. It - Information Technology, 54(4), 157–164. doi.org/10.1524/itit.2012.0680

Fuest, T. / Sorokin, L., Bellem, H., & Bengler, K. (2017, in press.). Taxonomy of traffic situations for the interaction between automated vehicles and human road users. Proceedings of the Conference on Applied Human Factors and Ergonomics (AHFE), July 17-21, Los Angeles, California, USA

Gontar, P., Homans, H., Rostalski, M., Behrend, J., Dehais, F., & Bengler, K. (2017, manuscript under review). Are Pilots Prepared for a Cyber-Attack? A Human Factors Approach to the Experimental Evaluation of Pilots’ Behavior. Journal of Air Transport Management.

Gontar, P., Schneider, S. A. E., Schmidt-Moll, C., Bollin, C., & Bengler, K. (2017, manuscript under review). Hate to Interrupt You, but… Analyzing Turn-Arounds from a Cockpit Perspective. Cognition, Technology & Work.

Gontar, P., Fischer, U., & Bengler, K. (2017, in press). Methods to Evaluate Crew Communication in a Training Environment: Speech Act Based Analyses vs. Cross Recurrence Analysis. Journal of Cognitive Engineering and Decision Making.

Gontar, P., & Mulligan, J. B. (2016). Cross Recurrence Analysis as a Measure of Pilots’ Coordination Strategy. In A. Droog, M. Schwarz, & R. Schmidt (Eds.), Proceedings of the 32nd Conference of the European Association for Aviation Psychology (pp. 524–544). Groningen, NL.

Reinhardt, J., Schmidtler, J., Körber, M., & Bengler, K. (2016). Follow Me! How Robots Should Guide Humans. Zeitschrift Für Arbeitswissenschaft, 70(4), 203–210. doi.org/10.1007/s41449-016-0039-2

Schmidtler, J., Knott, V., Hölzel, C., & Bengler, K. (2015). Human Centered Assistance Applications for the Working Environment of the Future. Occupational Ergonomics.