Fusing Orthogonal Technologies
The final report for this workshop is available at:
ADSA05 focused on fusing orthogonal technologies for detecting explosives for aviation applications. This workshop was the fifth in a series dealing with algorithm development for security applications. The pattern with the series has been to deal broadly with a new topic in the odd-numbered workshops in order to scope out the topic. The even-number workshops are then used to address a focused application related to the topic. The present workshop followed this pattern by broadly addressing the topic of fusion, which is also known as sensor fusion or data fusion.
The topic of fusion was chosen for the workshop in order to support the Department of Homeland Security’s (DHS) objective of improving the performance of existing technologies, where performance is defined as increased probability of detection, decreased probability of false alarms, lower threat mass and increased number of types of explosives. There is evidence that existing technologies will eventually be unable to satisfy DHS’s requirements for improved performance unless they are fused with other technologies.
Much of the discussion at the workshop dealt with defining the terms fusing, orthogonal and technology. Multiple definitions exist for each of the terms. Additional discussions dealt with improved performance of new and emerging technologies when deployed as stand-alone systems and when fused. The discussion of improved performance was primarily addressed how to increase the probability of detection and reduce the probability of false alarms. Additional discussions dealt with operational issues such of cost of ownership, concept of operations and risk-based screening.
Finding: DHS is not well educated in fusion and its terminology.
Recommendation: DHS should define terms used by fusion experts in R&D and other fields.
Finding: DHS has experience with fusing systems and some of these have failed.
Recommendation: Need to learn why these systems failed.
Recommendation: Need to focus on a particular problem and try to solve it to set precedence.
Recommendation: Need to establish performance metrics to be able to judge effectiveness of individual sensor systems and compare improvements due to fusing two or more systems.
Recommendation: Address how technologies are designed and chosen so that, when fused, the resulting fused system has better performance than existing technologies.
Finding: Risk-based fusion was discussed, but was not part of the workshop.
Recommendation: This topic needs to be a focus and discussed.