clients / case studies / universities

UA Electrical and Computer Engineering Department

Goal

To deliver to the Government a top-quality product in response to a research grant.

Solution

Ephibian helped the University of Arizona Electrical and Computer Engineering Department design and deliver the results of cutting-edge research. Doing so facilitated the University’s receipt of several additional phases of research grant funding.

Results

Introduction

The U.S. Army Battle Command Battle Lab wanted to create data visualization and "big data" analysis tools to enable military intelligence analysts to quickly digest, identify and display trends within large, complex and inconsistent data sets. By identifying trends in enemy courses of action, analysts could then search for signals that predicted future attacks.

What did we do?

  • Ephibian and the University of Arizona worked together to design and develop sophisticated search algorithms and intuitive query modeling tools, two and three-dimensional data visualization capabilities to rapidly inform the users, and collaboration mechanisms so the analysts could share the gathered knowledge. This program is called the Asymmetric Threat Response and Analysis Program (ATRAP).
  • ATRAP allows military intelligence analysts to identify enemy courses of action and their key indicators within massive data sets, which they can then use to more accurately predict the actions of these unconventional adversaries, including terrorists and insurgents.
  • While the University faculty and staff focused on research, Ephibian turned their research results into scalable and efficient software solutions.
  • Ephibian helped the University retain project knowledge and direction across multiple years and an ever-changing graduate student base.

How did we do it?

  • By developing the ATRAP application to contain a wide variety of data ingestion tools
  • By implementing flexible data organization and visualization tools including entity and link creation, time analytics, geospatial mapping, 2D and 3D visualization and more
  • By creating tools for the analysts to easily query the complex datasets, and score the likelihood of those models being correct (as well as scoring each substep within the model). The system also auto-derives other likely models
  • Ephibian applied UA theoretical research in the area of social network behavior to identify entities (e.g., people, locations, equipment) within the data that are crucial players and “central” to the human Intelligence networks, versus those which lie on the periphery
  • We created collaboration tools to help analysts share and integrate their collected knowledge and findings. This allows the users to subdivide the problem, yet integrate their overall knowledge and understanding
  • Ultimately we developed the capability of users to construct spatial and temporal models of assumed courses of action, verify the validity of these models by auto-scoring them against the mountains of available data, in order to unearth likely enemy actions. Then the system allows the user to apply these models to review incoming data streams to auto-alert when precursor signals of enemy actions emerge which indicate (predict) a future event.

What was the result?

The successful development and implementation of the ATRAP tool helps drive limited human collection teams to confirm or deny enemy courses of action. The program was given authority to operate on DoD networks and is available for use in the field by military intelligence analysts.

The University has achieved a successful program, that has resulted in several rounds of research funding and holds promise to deliver even more, as the versatile tool suite can be tailored for used in other knowledge domains.