In this work, we present a novel improvement to classical vehicle tracking algorithms by implementing a three-tier architecture consisting of a data-centric vehicle tracker paired with a hypothetical thinking layer that is controlled by an overarching goal layer – this models more effectively how a human thinks about and analyzes situations like vehicle tracking. The upper two layers are disassociated from the data itself and instead operate from the idea of qualia in event space. Our proof-of-concept results show how a classical vehicle tracker can be improved by fusing multiple input sources using coincident SAR and EO data paired with a thinking layer that is able to detect, hypothesize, and resolve conflicts.

Date Published: 14 May 2015
PDF: 11 pages
Proc. SPIE 9458, Cyber Sensing 2015, 945806 (14 May 2015); doi: 10.1117/12.2176582

Published in SPIE Proceedings Vol. 9458:
Cyber Sensing 2015
Igor V. Ternovskiy; Peter Chin, Editor(s)