The department has had a long history of conducting research to provide improvements in fire detection methods. The research has sought to identify improvements to provide more rapid detection while having fewer nuisance (i.e. false) alarms.
Much of the research has been directed at understanding the components of fire signatures as well as those associated with non-fire environments coupled with an assessment of the response of sensors to those conditions. Research activities have included development of a prototype for a multi-sensor detector that included traditional smoke sensors, heat sensors and gas sensors. In an initial project, analysis of experimental data from multiple sensors was done via artificial intelligence methods to distinguish between flaming fires, smoldering fires and nuisance alarms. The resulting algorithm provided faster response times than traditional smoke detectors with very few nuisance alarms. In a related project, multi-criteria algorithms with smoke and gas sensors were developed which yielded promising results for prompt detection with few nuisance alarms for applications on Navy ships. The department has also worked with a Maryland company to explore the ability of a smoke detector which they had recently to ignore nuisance alarms.
Department researchers have had several opportunities to explore the performance of contemporary detection technologies including air sampling, video detection and dual band projected beam detectors. The thrust of the work on video detection was to assess the performance of video detection in a tall space. By conducting experiments in an indoor arena, the response of video smoke detection was demonstrated. In addition, the department team developed an improved algorithm for identification of a rising smoke plume by the video detector.
Nuisance alarm rejection has been the topic of numerous research efforts. In one project, research conducted by a team of sophomores in the department developed an experimental protocol for the assessment of nuisance alarm rejection by dual wavelength, projected beam detectors.
The applicability of contemporary detection technologies in aircraft cargo compartments has been studied. The purpose of the project was to assess response characteristics of a variety of detection technologies, including air sampling detectors, dual wavelength spot detectors and gas sensors to a variety of fire sources. Experiments were conducted in three spaces, resulting in a demonstration of the scalability of the response of detectors to fire signatures.
With the multitude of sensors being distributed throughout contemporary buildings to monitor a variety of conditions, research has been conducted by the department to determine if information from an array of distributed sensors can be utilized to provide improved situation awareness. BIM software was used to represent the building and provide visualization of fire conditions produced in experiments.
Most of the projects involving advanced detection have been led by Dr. Jim Milke, though the project on distributed sensors was led by Dr. Andre Marshall.