Symbolic modelling of the dynamic environments of autonomous agents
Sammanfattning: To interact with a dynamic environment in a reliable and predictable manner, an autonomous agent must be able to continuously sense and “understand” the environment in which it is operating, while also meeting strict temporal constraints.In this thesis we present means to support this activity within a unified framework aimed to facilitate autonomous agent design and implementation. The central role in this approach is played by models at different levels of abstraction. Those models are continuously updated on the basis of available information about the dynamic environment. We emphasize the interface between the numeric and symbolic models, and present an approach for recognizing discrete events in a dynamic environment based on sequences of observations. Furthermore, we propose a logic to specify these characterization procedures.A prototype driver support system is used as a means for testing our framework on a real world application with considerable complexity. The characterization procedures are specified in the logic, and an implementation of the prototype is presented.
Denna avhandling är EVENTUELLT nedladdningsbar som PDF. Kolla denna länk för att se om den går att ladda ner.