One long term goal in current robotics research is the development of robot systems, which have approximately the same cognitive, communicational, and handling abilities like humans. To achieve this goal, a variety of problems have to be solved by the researchers, for instance in the domain of communication. Here, it is necessary to bridge between different types of communication. On the one side there is the human, who typically prefers a symbolic communication using symbols like gestures or words. On the other side there is the robot system, which typically prefers a subsymbolic communication with subsymbolic information like joint positions, forces, or motor currents.
In general, there is no direct mapping from symbolic to subsymbolic information. Therefore, a transformation is needed, which is able to combine the two types of representations. Therefore, one objective of this work is the analysis of the requirements for such a transformation layer and the proposal of concepts for combining symbolic and subsymbolic representations. Based on this concepts, we describe how to ground symbols to a robot system. At last, we analyze how a robot system can utilize the proposed concepts in order to provide an intuitive, symbolic user interface.
Within this work, we propose the usage of physics, respectively physical effects and quantities, in order to bridge between symbolic and subsymbolic information. These effects and quantities are used for the description of executable actions, respectively for symbolic parameters of executable actions. For the description of executable actions we introduce the concept of Principal Physical Effects. This concept describes the relations between the transformation layer and the subsymbolic action representation suitable for a robot system. More specific, this concept is used to describe the relations between a physical process, a physical law, and a subsymbolic action representation based on sensor-based robot motions. To describe the relations between a symbolic representation and the transformation layer, we enhance the concept of Principal Physical Effects to Verbalized Physical Effects. This allows a mapping of the symbolic description of a verbalized instruction (e.g. shove, rotate), the symbolic parameter of a verbalized instruction (for instance nominal or prepositional phrases) to a corresponding Principal Physical Effect. To use symbolic parameters combined with an action representation based on Principal Physical Effects, these symbols have to be described in terms of physics. Therefore, we propose the usage of a physical dictionary, which stores the symbol information in context of physical properties.
Besides the grounding of symbol information, we analyze how a robot system can utilize the proposed symbol description in order to provide an intuitive, symbolic user interface. Therefore, we describe additional components, which illustrate the transformation of a symbolic instruction towards a parameterized, subsymbolic robot motion. At last, we evaluate the concepts using a prototype system.
|School:||Universitaet Bayreuth (Germany)|
|Source:||DAI-C 81/1(E), Dissertation Abstracts International|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be