You are here

Validation of Automatic Ground Moving Target Indicator Exploitation Algorithms

Description:

OBJECTIVE: Develop a cost effective method to provide independent validation of automated Ground Moving Target Indication (GMTI) exploitation algorithm performance. DESCRIPTION: Ground Moving Target Indicator (GMTI) data has proved to be a valuable tool for the warfighter. GMTI data provides an indication of ground target movement, as well as the direction of that movement and an estimate on target size. GMTI data can be collected over extremely large areas of interest, with high update rates, and provides all weather, day and night coverage. Exploitation data products include force protection, detection of traffic in restricted areas or near an installation of interest, link analysis, patterns of life discovery, and milling activity. The current volume of GMTI sensor data is too large for the analyst community to fully exploit, and data volumes will only grow as new sensors are deployed. Proper exploitation of GMTI data requires a higher degree of training than for exploitation of other common ISR INTs, therefore GMTI data exploitation is severely limited by the availability of trained personnel. Development of high quality automated GMTI data exploitation systems is essential in order to extract the full intelligence benefit from current and future GMTI systems. Many research groups are actively developing algorithms and tools for automatic exploitation of GMTI data in answer to this need. These tools are designed to provide a variety of data products, including target tracks and indications of specific target behaviors. Generally, GMTI exploitation algorithms have only limited testing at best, due to the prohibitive cost of running live sensor experiments, and the limited number of data sets for which ground truth has been collected. Analyst acceptance of exploitation algorithms is hampered by the lack of verified metrics quantifying the probability that exploitation data products are correct. GMTI data analysts will only accept automated exploitation systems if they feel confident that the systems are providing accurate intelligence. Without this confidence, tools are not adopted, and data exploitation suffers. This solicitation is for the development of a GMTI exploitation algorithm validation capability that will ensure that efforts to bring exploitation tools to the war fighter provide high quality, high confidence data products. The successful proposal will provide a statistically significant validation framework within reasonable cost constraints. This framework will be able to accommodate validation of a diverse set of exploitation algorithms goals, to include target tracking and detection of milling activity and other GMTI exploitation data products. Potential non-Government applications of this framework include validation of traffic modeling and site security techniques and algorithms. The Government will provide data as needed for the development; simulated unclassified data if required for Phase I and real GMTI data for Phase II and beyond. PHASE I: Develop a low-cost and statistically rigorous approach to GMTI exploitation algorithm validation. This approach should be applicable to general GMTI sensor systems and flexible enough to be useful for different exploitation goals. Describe how the methodology will be used to evaluate data exploitation algorithms created by third party developers, including interfaces and software dependencies. PHASE II: Based on the techniques demonstrated in phase I, develop a GMTI exploitation algorithm validation framework which provides statistically rigorous and unbiased analysis of candidate exploitation algorithms. Demonstrate proof-of-concept of this validation approach by applying it to government supplied data and algorithms in order to evaluate those algorithms for probability of correctly extracting information data products. Provide documentation of system ICDs and basic user's manual. PHASE III: Transition validation tool to government system(s), provide training in tool functionality to increase analyst acceptance of the validation methodology and tool adoption. Provide instantiations of the tool and test data sets to algorithm developers for early Q & A of automated exploitation tools. REFERENCES: 1. United States Air Force Chief Scientist (AF/ST) Report on Technology Horizons Volume 1, AF/ST-TR-10-01-PR, 15 May 2010. 2. Fundamentals of Radar Signal Processing, Mark A. Richards, McGraw-Hill.
US Flag An Official Website of the United States Government