DARPA plans tools to lighten load of battlefield data

DARPA's Insight program will develop automated tools to help human analysts sift and identify valuable battlefied data from a variety of information sources and platforms.

Today's soldiers and the analysts who support them suffer from an embarrassment of riches. The constant flow of battlefield sensor data from space, aerial and ground-based sensor platforms provides an unsurpassed view of the battlefield.

But many of these platforms cannot easily share or mix information such as video and radar data, and human operators are often overwhelmed by the sheer volume of material.

The Defense Advanced Research Projects Agency’s Insight program plans to create an automated system that helps human analysts by blending together sensor feeds from a variety of platforms and sources. The agency’s announcement noted that the lack of an automatic exploitation and cross cueing capability for multi intelligence sources is a major shortfall for current intelligence, surveillance and reconnaissance systems.

The shortcomings of current ISR platforms and systems include a lack of automated tools to interpret, edit and weave data streams into a form useful to human analysts. According to the solicitation, vital information is often lost or overlooked due to the overwhelming flow of incoming data. A lack of integrated human-machine reasoning tools limits the ability of system to use operators’ knowledge and ability to understand complex data.

Insight plans to address these shortcomings by developing a next-generation ISR Exploitation and Resource Management system. The program will build on several areas. The first is model-based correlation using behavioral modeling and threat networks analysis tools to automatically combine data across sources and manage random and uncertain events.

DARPA plans to build resource management tools to identify data collection opportunities, permitting the efficient use and selection of a variety of sensors and platforms for each mission. A unified data management and processing environment will allow the rapid integration of new data sources, algorithms and analysis methods. Finally, the effort seeks to integrate human and machine processing in areas such as visualization, hypothesis manipulation and online learning.

The program’s E&RM system research has three parts: an analysis and resource management system, a unified all-source ISR human-machine interface, and an integration, processing and data archive environment. The analysis and resource management component will develop a range of information analysis and sensor management tools and their underlying algorithms, while the goal of the ISR human-machine interface is to develop a “modern human-computer interface that is an adaptable, customizable, context aware, unified interface capable of supporting dynamic, interactive missions.” Besides developing a variety of interface tools, this part of the program also seeks to conduct core research into enhancing human-computer interaction.

The last component, the integration, processing and data archive environment, will serve as a unified data source for all data processing. According to DARPA it will provide a distributed environment for visualization and computing resources, rapidly integrate new algorithms and manage existing ones.

To evaluate the results of the E&RM system, DARPA will develop the Insight Test Bed. The test bed will have three components, a physical test bed, a virtual environment to evaluate ISR exploitation systems with simulated sensor data, and a development incubator designed to lower the cost of entry and increase the number of program participants. The incubator also will serve as a repository for all of the program’s collected, simulated and processed data.

By combining Insight’s E&RM system and the test bed, DARPA hopes to “provide the unprecedented capability to rapidly and cost-effectively develop and demonstrate novel approaches to agile ISR system evolution.”

DARPA also plans to link Insight to other related sensor and data fusion programs, such as Wide Area Network Detection, Autonomous Real-time Ground Ubiquitous Surveillance, Video and Image Retrieval and Analysis Tool, Persistent Stare Exploitation and Analysis System and the Graph Understanding and Analysis for Rapid Detection – Deployed On the Ground (GUARD DOG).