Cognitive Perspective on QSR

Positioning. The article, published in the AAAI Spring Symposium, investigates the preconditions for combining qualitative reasoning and cognitive aspects on a navigation task.

Method. Task analysis and review of cognitive adequacy of QSR concepts.

Results. The article follows the original motivation of qualitative reasoning as the investigation of a “naïve physics”. This comprises to describe the world with human “everyday” concepts. Therefore a pure formal description is not sufficient – findings from Cognitive Science about human concepts. The article presents: (i) main findings from Cognitive Science about QSR and cognitive modeling, (ii) relates them to current research about QSR, (iii) suggest methodological adjustments that allow both for a more comprehensive evaluation of formalisms and for more realistic and diverse test-beds. Navigation deals as a test-bed for a cognitive QSR.

Ragni, M., & Kuhnmünch, G. (2009). A Cognitive Perspective on QSR: Navigation as an Example. In B. Nebel & S. Woelfl (Eds.), AAAI Spring Symposium: Benchmarking of Qualitative Spatial and Temporal Reasoning Systems(pp. 23–28). AAAI.

Qualitative Reasoning and Robotics

Positioning. The article, published in the Journal of Visual Languages and Computing, investigates a calculus with relative orientation, i.e., egocentric views in contrast to orientations that refer to a global reference system.

Method. Formal analysis; application to robotics.

Results. Reasoning about relative orientations poses additional difficulties compared to reasoning about orientations in an absolute reference frame. A complexity analysis of the ternary calculus reveals that it is in PSPACE; it utilizes finer distinctions than previously published calculi. Additionally, it permits differentiations which are useful in realistic application scenarios such as robot navigation that cannot be directly dealt with in coarser calculi.

 

Moratz, R., & Ragni, M. (2008). Qualitative spatial reasoning about relative point position. Journal of Visual Languages & Computing, 19(1), 75–98.

Spatial Cognition, Artificial Intelligence, and Robotics

Positioning. The article, published in the journal “Künstliche Intelligenz”, presents an overview of the research field Cognitive Science and its connection to Artificial Intelligence.

Ragni, M., Wiener, J. M., Hölscher, C., Brösamle, M., Büchner, S. J., Kalff, C.Strube, G. (2009). Human Spatial Cognition: Cognitive Science Joins AI and Robotics. KI, 23(2), 46-9.

Cognition and Artificial Intelligence

Positioning. The article, published in the textbook “Künstliche Intelligenz”, presents an overview of the research field Cognitive Science and its connection to Artificial Intelligence.

Strube, G., Ferstl, E., Konieczny, L., & Ragni, M. (2013b, July). Kognition. In G. Görz, J. Schneeberger, & U. Schmid (Eds.), Handbuch der Künstlichen Intelligenz. Oldenburg

Cognitive Robotics and Navigation

Positioning. The article, published at the International Conference on Cognitive Modeling 2012, presents a cognitive ACT-R agent controlling a simple robot by the cognitive architecture ACT-R.

Research Question. Can we replicate human behavior in a navigation task by implementing ACT-R on a Lego MindStorm Robot?

Method. Cognitive Modeling

Results. The cognitive robotic system (Lego MindStorm Robot) shows a similar behavior to humans while navigating in a labyrinth. Especially the construction of a pseudo mental map (based on chunks) shows that many limitations of humans depend on the restriction of the working memory.

Bennati, S., & Ragni, M. (2012). Cognitive Robotics: Analysis of Preconditions and Implementation of a Cognitive Robotic System for Navigation Tasks. In N. Rußwinkel, U. Drewitz, & H. van Rijn (Eds.), Proceedings of the 11th International Conference on  Cognitive Modeling (pp. 157–162). Berlin: Universitätsverlag der TU Berlin.