Robert Bratuša and Marko Javornik (1998) . Prešeren awards for students.

Full text not available from this repository.

Abstract

One of the topical research problems is the reconstruction of the operator's skill of controlling the dynamic systems , such as aircraft or crane. One of the difficulties we face is that the skill is to a large extent subconsious and operator is often unable to express it in a functional form. Recently, the artifical intelligence for reconstruction of the subconsious skills. The process uses the operator's traces for the example-based learning. An appropriate design of the model was apllied. Experiments with reconstruction of the controlling skill were performed on dynamic systems using some of the machine learning programs. this has been tested on the pole and cart, and crane domains. The conceptual model uses the quantitative information, as well as the quantiative rules, operator's sub-goals and desired control trajectory. The work involves an extensive research of the combination of different strategies. The research work is concluded with an estimate of the robustness of clones anq qualitative strategies, by changing the starting position and parameters of the dynamic system. Additionally, a comparison to other methods has been made.

Actions (login required)