Hypotheses Generation for Process Recognition in a Domain Specified by Temporal Logic
| Type of publication: | Inproceedings |
| Citation: | Colonius:2012uh |
| Publication status: | Accepted |
| Booktitle: | ARTIFICIAL INTELLIGENCE AND LOGISTICS |
| Series: | Report Series of the Transregional Collaborative Research Center SFB/TR 8 Spatial Cognition |
| Number: | 3 |
| Year: | 2012 |
| Month: | August |
| Pages: | 49-54 |
| Location: | University Bremen |
| Abstract: | The degree of automation in the logistic domain is in- creasing, which amplifies the need of autonomous robotic systems. An autonomous system requires situational awareness as well as an understanding of processes in its environment to act goal orientated. Process Recognition methods show valid solutions for this problem. Considering the challenge of recognizing processes from a humans point of view, explainability of found solutions and simple process definitions for processes to search for becomes essential. This leads to the use of qualitative reasoning techniques. Especially logistic do- mains like warehouses are highly dynamic and mobile robots situated in that domain often do not have complete sensor coverage of the whole environment. Logic methods like deduction for model check- ing can thus be insufficient due to incomplete or wrong observations, breaking the reasoning chain. Topic of this paper is the enhance- ment of the deductive reasoning for process recognition with hy- potheses based on background knowledge and reasoning techniques to overcome these limitations. Therefore, a qualitative measurement for certainty will be motivated and benefits of hypotheses integration shown. |
| Keywords: | |
| Authors | |
| Editors | |
|
Attachments
|
|
|
|
|
Notes
|
|
|
|
|
|
Topics
|
|
|
|
|
