Institute of Computer Technology at Vienna University of Technology, Gußhausstraße 27-29, 1040 Vienna.

---

Abstract—An autonomous agent must know where it is, which objects are in the nearby surrounding area and where it can find already seen, far away objects. On the basis of the existing cognitive architecture Simulation of Mental Apparatus and Applications (SiMA) at the Institute of Computer Technology (ICT) we show, how cognitive architectures can get extended by location awareness and navigation. This paper deals with the problem of self-localization and path finding. Therefore, landmarks, locations and regions will be used. Actions lead the agent from one location to another. Locations are hierarchically grouped together to regions. The functionality was implemented exemplary in a multiagent simulation using simplified scenarios. The results confirm the prospects. Now we can start to introduce the functionality in our full featured main simulator.

Index Terms—Artificial general intelligence, artificial recognition system, cognitive architectures, cognitive automation, landmarks, localization, navigation, psychoanalytically inspired AI, software agents.

[PDF]

Cite: Christian Brandstaetter, Samer Schaat, Alexander Wendt, Martin Fittner, "How Agents Use Breadcrumbs to Find Their Way," Journal of Computers vol. 12, no. 1, pp. 89-96, 2017.