Learning from interaction with the environment using a situation-operator calculus with application to mobile robots

This contribution demonstrates the application of a situation-operator modeling technique to learning from interaction with the environment, planning to reach a given goal and executing the corresponding plan. A mobile robot is chosen to demonstrate the concept in a real world environment The proposed architecture of an autonomous learning system consists of different modules, e.g. modules for data compression, short term memory and goal translation. The learning, planning and execution modules as well as the database build the representational level of the proposed approach. The core of the approach is the assumption that changes of the world are understood as a sequence of scenes and actions. These parts are modeled using a special situation and operator calculus.