(a) How to efficiently process imperfect, heterogeneous sensor data over multiple signal

     modalities to detect weak signal signatures using wireless sensor networks?

(b) How can we locally model sensor data to reduce the communication burden of a

To address above (a), we developed a novel paradigm named for Evidence Filtering for multi-modality information processing in wireless sensor networks. Then, we extended this method using Distributed m-D State Space Models to enable local processing of sensor information inside nodes, to address the issue in above (b).