Systems, Man, and Cybernetics

 

Home

• [P863] Simulation of Learning and Planning by a Novel Architecture for Cognitive Technical Systems
  Author(s): Dennis Gamrad, Dirk Soeffker
  
  

  Abstract
  In this contribution, the concept and implementation of a novel architecture to simulate cognition is presented. The proposed architecture is based on a high-level Petri Net description and implemented with a suitable software (Renew). It contains an approximate model of the reality based on a principal layer of the representation of facts allowing cognitive functions and procedures. Here, this is used for planning purposes. The model describing the reality is based on the Situation-Operator-Modeling (SOM) approach and can be modified during simulation in order to reduce differences to the reality. As an example for illustration, the cognitive functions learning and planning are presented by means of the interaction of an agent in a grid-based environment.

  
   

Full Text (PDF)