PCSS

Physical-Cyber-Social Systems

There is increasing amount of data on the web which is a collection of human knowledge and experience. Machines, when enabled access to this huge knowledge on the web, can not only empower them with an ability to interpret the observations they make, but also learn concepts from the observations. This ability of machines to interpret observations and deriving concepts form them can lead to powerful and smart machines which enhance the Human Experience.

We present a physical-cyber-social system, a Semantic Web research effort that exploits background knowledge and abductive reasoning realized in OWL which is basically deductive in nature. We present use of semantic knowledge base and abductive reasoning for processing realtime sensor data to derive human intelligible abstractions. This system shows identification of a fire from mobile platform mounted sensors sending realtime data streams.

Projects

Semantics in Rescue Environment

Semantics Empowered resCUe enviRonmEnt (SECURE) is a project in which we are using a cyber-physical-system like a robot, with various types of sensors like temperature, CO2, CO, Infrared to identify various types of fires using ontology of perception and background knowledge of the domain. In this project, we study the use of background knowledge by rescue robots, thereby enabling rescue robots to be knowledgeable leading to better situation awareness, quick control of emergency situation and increased lives being saved.

Perceiving the Environment

The goal is to study the use of background knowledge to create abstractions of sensor data. These abstractions are human intelligible and makes the first responders better understand the situation. These abstractions are derived from raw sensor observations using the perception process.

Use of Disaster Related Knowledge Bases

The goal here is to enable the robot to access various ontologies related to disaster like in [1]. We would like to explore the way in which we introduce concepts into the ontology. Since we are dealing with constraints with respect to the resource, we need to retain minimal concepts in the ontology. The size of the ontology decides the efficiency of the operation of the robot and there by impacts the rescue operations.

References

[1] Heather S. Packer, Nicholas Gibbins, Nicholas R. Jennings: An On-Line Algorithm for Semantic Forgetting. IJCAI 2011: 2704-2709.