A Synergic Expert-Neural Network System to Identify Relocatable Targets using M

Period of Performance: 12/23/1993 - 12/23/1995

$466K

Phase 2 SBIR

Recipient Firm

LNK Corp., Inc.
6811 Kenilworth Avenue, Suite 306
Riverdale, MD 20737
Principal Investigator

Research Topics

Abstract

Intelligent and advanced ATR capabilities will be strongly desired in future for targeting applications during unmanned missions with missiles such as TOMAHAWK or manned missions with multi-mission aircraft such as F/A 18. Different sensors and feature data sources provided different types of information about the terrain features. Also, different computational paradigms offer different advantages which when combined together in an efficient way offer the strength of a unified framework. The primary goal of the proposed work is to develop a prototype hybrid multi-source integration system for ATR applications. The prototype makes use of a three-tier hybrid architecture developed in Phase I for identifying relocatable targets by integrating information from multiple sensors using a synergistic framework of neural networks, expert systems and fuzzy logic. The first level of this system addresses issues related to data representation and registration of multiple sources. The second level achieves featue extraction and partial recognition results using neural networks. The third level consists of a decision making expert system with fuzzy logic reasoning to reach a collective decision from multiple sources using an object-oriented representation of the target. The proposed Phase II effort directly supports the on-going programs associated with automated targeting. For example, for the TOMAHAWK program the prototype can help achieve better targeting accuracy through intelligent integration of multiple sensors and cartographic sources, and for the F/A-18 program this prototype development can help significiantly reduce the information assimilation load on the pilots. The Phase II work will serve as a foundation for a stand-alone hardware supported ATR system development in Phase III.