Before the optimal actions of a robot in a system can be determined, the system itself has to be understood. The subparts of the system can interact with each other and a model is used to formalize the values with mathematical equations. A typical example is a bathtube process, in which water floods into a tube.[1] Such a system can be visualized with a computer graphic and can be described with the QSIM tool.
Another slightly more powerful modeling technique is a randomized model.[2] Here is the idea to create a event transition graph which describes the working of the system with a probabilistic approach. This allows to describe a system even not all details are understood.
Qualitative reasoning[]
- ontology "qualitative physics"[3]
- grounding costmap[4]
- "natural language instructions" interactive "Qualitative physics"
- "hierarchical task network" "qualitative physics"
- qualitative kinematics
References[]
- ↑ Hocaoğlu, Mehmet Fatih. "Qualitative Reasoning for Quantitative Simulation." Modelling and Simulation in Engineering 2018 (2018).
- ↑ Eveillard, Damien, et al. "Probabilistic modeling of microbial networks for integrating partial quantitative knowledge within the nitrogen cycle." Frontiers in microbiology 9 (2018): 3298.
- ↑ OUR-K framework, The Kautham Project
- ↑ “Learning Qualitative Spatial Relations for Robotic Navigation”, page 3 shows an example for a cost map