A sensor ontology for the domain of firefighting robots

Amy R Wagoner, Purdue University

Abstract

Fires create thousands of dollars in damage and thousands of deaths each year. Firefighters risk their lives everyday and are often killed in action. Firefighting robots may be able to reduce the loss of lives and damage due to fires. Robots are often used for redundant tasks that require the consistency and efficiency of a machine. They are especially optimal for tasks that require strength that exceeds that of a typical human being or for environments that are hazardous to people. Robots' metallic exteriors are far more durable and easier to replace than flesh and blood, thus they are ideal for fighting fire that may be unreachable or too dangerous for humaning beings. Firefighting robots are most often shaped like tanks and are equipped with fire extinguishers, sensors, and cameras. The robots are typically operated via remote control and lack autonomy. Because of the volatile nature of fires, it is difficult for software engineers to create algorithms to make firefighting robots more autonomous. Ontologies are commonly used for sharing domain information and structuring and analyzing data. This study proposes using an ontology that is designed specifically for a firefighting robot programmed to rescue a human in danger in order to make a decision making algorithm. The methodology uses ontological tools to build the ontology. A decision-making algorithm is created using the information that is stored in the ontology. The study is evaluated on the accuracy rate of making the correct decision. It is also evaluated on if the decision-making algorithm performs significantly better than decisions chosen at random..

Degree

M.S.

Advisors

Matson, Purdue University.

Subject Area

Robotics

COinS