SBIR-STTR Award

Under the advanced weapon station automation/integration topic, we propose an innovative research to integrate target extraction, automation and expert systems into one demonstratable weapon system for training intelligent gunners, and developing fire control decisions. Three phase i objectives are: (1) to demonstrate that this weaponn system can be developed using "off-the-shelf" hardware technologies and the existing sr&e atr algorithms; (2) to determine dominant factors affecting an totally automated segmentation process and (3) to conduct a series of demonstrations after the final system is developed and tested using real world image data and tactical information embedded as expert systems controlling the segmentation and target extraction processes. The unique properties of the weapon system are: (1) automation as the basis of atr action; (2) embedded expert systems as the decision maker to control the automation processes; and (3) the weapon system serving as the training round for developing the expert system. In this system concept, the human analyst is to control the automated atr processes, but at the same time he/she is to learn from the weapon how to develop rules for an automation file.

In the SBIR Phase I effort on Advanced Weapon Station development, we examined three related topics: (1) automation, (2) embedded expert systems, and (3) automated target recognition (ATR) system integration.The tasks were accomplished by using FLIR imagery from the PAIRSTECH program. On automation, we examined the factors influencing the hit-rates when classification is performed using Fourier descriptors. On embedded expert systems, we demonstrated how a classification process can be controlled using system parameters such as our Auto-start and Auto-stop. On integration, we examined the ATR components including sensors, trackers, segmentors, classifiers and interface modules, and discussed how an integrated ATR system can be achieved. Based on these Phase I results, we propose a Phase II effort centering on the delivery of a prototype weapon station system to ARDEC via these major tasks: (1) development of system modules: tracker, segmentor and classifier; (2) development of interface modules: tracker to segmentor, classifier to classifier and classifier to wire frame models: (3) acquiring image data and conducting demonstrations, and (4) delivery of a prototype system. In addition, we will conduct a number of field tests with the developed system at ARDEC and possible at the Aberdeen test site where the PAIRSTECH imagery was acquired.