Soft and hard fusion to improve prediction and retrodiction

Multitarget tracking and multisensor fusion have received much attention in the literature during the last four decades. Significant advances have been made in these fields during this period. With the existing computational resources, it is possible to track thousands of targets and fuse information from a multitude of heterogeneous sensors in real-time. However, the threats faced today are radically different from those during yesteryears and they continue to evolve, increasingly becoming difficult to track. In order to improve overall results, soft and hard fusion is essential in that highly quantifiable hard data and fuzzy soft data need to be fused. Soft and hard fusion can improve the results not only at the current time, but it can be used to improve prediction and retrodiction results as well. In this position paper, we explore the use of soft and hard fusion to improve tracking, prediction and forensic results. © 2011 SPIE.