Description:
OBJECTIVE: The objective of this research is to use spatial, temporal and graph analysis techniques to take very large data streams over wide areas and autonomously highlight areas of interest for a decision maker without a priori knowledge of the area and/or location of high value. DESCRIPTION: To protect U.S. national interests and achieve the objectives of the 2010 National Security Strategy, the Joint Force will need to recalibrate its capabilities and make selective additional investments [1]. The well-publicized push to the Pacific Rim drives a need to improved mission planning over much larger areas, including some for which conventional information gathering is challenging. Improved pattern recognition capabilities enabled by spatial, temporal and graphical analysis are one way to effectively improve mission planning and execution given limited resourced. The U.S. Marine Corp desires to improve battlespace awareness (to know the enemy and environment) through improved analysis, prediction and production [2]. Relevant raw data can usually be stored and subsequently visualized spatially, temporally or as a graph. The automated recognition of an area and/or location of high value is basically a pattern recognition problem that must be solved in the absence of a data that is naturally tied to a uniform grid. Data about remote areas can still include imagery, radar and open source which can include social and cultural information. The diversity of this data set makes the establishment of the uniform grid often required by pattern recognition algorithms challenging [3]. The extraction and preprocessing of features from raw data is also challenging due to the well-known influences of environmental, social and cultural influences on the importance of other observed features [4]. Cognitive processes and context need to be considered by location/area of interest pattern recognition algorithms. [5] The technical challenges of this topic are as follows: 1) Select and automate the population of a feature layer; 2) Preprocess data as required to approximate a uniform grid; 3) uncover linear/ nonlinear correlations between environmental, social and cultural variables and feature importance; 4) Mature models for why/when a location/area would be of interest; 5) Mature pattern recognition algorithms that approximate human recognition and classification skills utilizing cognitive insight; 5) provide a means to visualize current and predicted states (e.g. heat maps showing locations/areas of interest given a context). Define and apply metrics to measure modeling and prediction accuracy and potential for success of products produced. Creative solutions are desired. Data used should be relevant to potential use for product transition, such as a government agency, program of record or commercial market place. Use of open standards is encouraged to reduce costs and improve system interoperability. PHASE I: Identify a geographic region for study and data layer design concept. Demonstrate that diverse types of data can be ported to a spatial grid structure relevant to pattern recognition. Develop features models for areas of interest and then perform a proof of concept demonstration of a pattern recognition capability. The demonstration should be conducted using open source data. Document results from analysis and tests in a technical report or paper at a selected conference. The final Phase I brief should show plans for Phase I Option 1 and Phase II. PHASE II: Produce a prototype system that is capable of rapidly identifying locations/areas of interest based on a given context and visualizing that information as spatial heat maps with data traceability. The system should be able to automatically process data as sequential batch files or streaming data, accepting all standard raw data formats for images, maps, tracks, text and graphs. It is desired that context and pedigree of information be maintained for operator review. At this point the performer should focus on a proof-of-concept of capability of interest to transition program. PHASE III: Produce a system capable of deployment and operational evaluation. The system should consume available operational data sets and focus on areas that are of interest to specific transition programs or commercial applications. Machine based processing steps and metadata should be accessible by operator and presented in human understandable form. The software and hardware should be modified to operate in accordance with guidelines provided by transition sponsor.
