A Multi-Disciplinary University Research Initiative in Hard and Soft information fusion: Overview, research strategies and initial results

Abstract

The University at Buffalo (UB) Center for Multisource Information Fusion (CMIF) along with a team including the Pennsylvania State University (PSU), Iona College (Iona), and Tennessee State University (TSU) is conducting research to develop a generalized framework, mathematical techniques, and test and evaluation methods to address the ingestion and harmonized fusion of Hard and Soft information in a distributed Level 1 and Level 2 data fusion environment. The primary Research Thrusts addressed are framed around the major functional components of the JDL Fusion Process; these include: 1. Source Characterization of Soft Data input streams including; human observation-direct, indirect, open source inputs, linguistic framing, and text processing. 2. Common Referencing and Alignment of Hard and Soft Data, especially strategies and methods for meta-data generation for Hard-Soft data normalization. 3. Generalized Data Association Strategies and Algorithms for Hard and Soft Data. Robust Estimation Methods that exploit associated Hard and Soft Data. 5. Dynamic Network-based Effects on Hard-Soft Data Fusion Architectures and Methods. 6. Test and Evaluation Methodology Development to include Human-in-the-Loop. 7. Extensibility, Adaptability, and Robustness Assessment. 8. Fusion Process Framework. 9. Technology Concept of Employment. This program is a large, 5-year effort and considered distinctive in being a major academic thrust into the complexities of the hard and soft fusion problem. This paper summarizes the research strategy, the early technology decisions made, and the very early results of both design approaches and prototyping.