Cooperative Multiresolutional Information Fusion (CMIF) for Dim Target Detection and Tracking

MRLets Technologies, Inc. will develop a novel approach for detecting and tracking dim targets using information fusion of sensors from multiple platforms with the constraints of computation resource, data link bandwidth and network topology. This approach integrates the advances in distributed detection, sensor management, multiple target tracking and data communication by its feature technology _ multiresolutional technology. The center of this approach is Cooperative Multiresolutional Information Fusion (CMIF). Overcoming the difficulties faced by existing algorithms, CMIF extracts and integrates information directly at the measurement level, and provides coherence between detection, track initiation and track maintenance. CMIF employs both temporal and spatial domain multiresolutional techniques to extract coherent information of dim targets and control computational load. Local processing and cooperative processing extract measurement features and patterns and pack them effectively into Virtual Sequences (VSs). By communicating and fusing VSs among platforms, a major obstacle for measurement level information fusionis removed, while maintaining in the main advantage of measurement level fusion: better detection/tracking dim targets and reducing the effect of data eclipsing. A new data association algorithm - Multi-Pattern Data Association (MPDA) will be developed to take advantages of measurement level modeling. The goal of the work is computationally and communicationally tractable algorithms for multi-platform dim target detection and tracking in clutter.