Conclusions

Abstract

The fundamental hypothesis that “there is a need for intelligent physical compliance was formulated in the introduction of this work, in order to account for the shortcomings of today’s service robots. The hypothesis builds on the notion that particularly non-traditional robot tasks outside factory buildings demand compliant interaction capabilities in combination with advanced cognitive reasoning skills. However, this constitutes a need which is not even covered for application in laboratory environments. To overcome this issue, intelligent decision making strategies for compliant manipulation tasks have to be developed. That is, a robot has to process diverse task knowledge to plan effect-oriented actions, parameterize and execute these actions accordingly, and eventually interpret the outcome of the actions on a high-level of abstraction to estimate the task performance qualitatively. In order to facilitate this reasoning, suitable representations that incorporate the properties of physical compliant interaction have to be provided.