Application of Multi-level Fusion for Pattern of Life Analysis

Pattern of Life (POL) analysis constitutes a subset of Activity-based Intelligence (ABI) understanding those complex spatiotemporal contexts within which entities (e.g., cancer cells, people, etc.) move about and interact, normally-but not always-with a type of recognizable regularity. POL analysis methods are particularly important when attempting to detect and track complex behaviors in stochastic environments such as biological systems or urban terrains, where many interlocking entities coexist and share relationships (e.g., within metabolic pathways, air traffic, ground traffic, shipping environments, businesses, public transit systems, social organizations, etc.). We have developed a Pattern of Life Integrated System (POLIS), which provides a solution combining different but complementary techniques (mathematical and logical approaches) together to form an automated, scalable (i.e., cloud-capable) fusion-based estimation process that can exploit a variety of information sources, including contextual, hard sensor data and other types of soft or human reported data (past reports, existing data models, etc.) to provide POL analyses and alerts which enable efficiencies in analysis and effectiveness in decision support. This approach gives shape to an innovative, unique, and defendable framework for the evolution of an advanced software system applicable to layered POL analysis and enhanced decision-making. This paper provides an overview of the Pattern of Life Integrated System, inclusive of all fusion levels which collectively support the POL analysis. In addition to the system and methodological overviews, a case study is presented which demonstrates the importance and value of the multi-level fusion approach for analyst decision support.