Tracking Events across Time and Space

March 13, 2009 By: Cyrena Respini-Irwin

ObjectFX's spatiotemporal rules engine helps analysts determine which incidents are important.

Over the past few decades, the primary challenge facing information analysts has evolved dramatically: a data drought has given way to a deluge. The proliferation of the number and types of data sources available — including static and video image collectors, electronic emissions, and GPS-equipped humans — threatens to overwhelm government and commercial users alike.

"We're just getting this flood of data . . . how do you find the nuggets of good data in that ocean that's coming in?" asked Steve Panzer, vice president of ObjectFX's Government Division. The answer, according to ObjectFX, is Spatial Rules; the spatiotemporal rules engine uses correlations and user-defined rule layers to automatically detect incidents of interest. The company, which began by offering Web-based visualization tools, introduced this contribution to geospatial event processing (GEP) about three years ago, said Panzer.

Like its umbrella discipline, complex event processing, GEP is a system for sifting through various inputs and distinguishing meaningful happenings from background noise. By analyzing the spatial and temporal relationships among events, it's possible to reveal the importance of an occurrence that would seem unremarkable on its own.

A spatiotemporal rules engine filters a stream of sensor-based data, indentifying actionable intelligence. Image courtesy of ObjectFX.

This automation also enables superhuman levels of monitoring, which are crucial when lives are at stake. Panzer explained that there are, on average, 1,500 to 2,000 commercial aircraft over the United States. "Obviously, no one can pay attention to all of those flights," he said. But with rules for proximity in place, Spatial Rules can automatically raise the alert level if two jumbo jets come uncomfortably close to one another. Shifting the burden of incident filtering onto a rules engine allows personnel — whether they be air traffic controllers, intelligence analysts, or battlefield commanders — to focus their attention on problem areas.

Before being filtered by a spatiotemporal rules engine, sensor data can be overwhelming and difficult to sift through. In this example, every one of the yellow dots that crowd the display represents a boat or ship. Image courtesy of ObjectFX.

GEP Supports Security Initiatives

National security is another arena where sifting out relevant information is of life-and-death importance. Intelligence applications, including monitoring borders, ports, and war zones, pose an additional challenge to GEP users. As Panzer said, they're tracking "not only things that are providing a cooperative signal [such as commercial-aircraft transponder signals] . . . for the intelligence community, we're tracking a lot of non-cooperative signals — things that don't want to be tracked."

In addition to drawing on both cooperative and non-cooperative signals, Spatial Rules can also incorporate real-time and historical data. Analyzing patterns of data interaction in the past, said Panzer, can help analysts make better use of intelligence assets in the future. For example, if it's determined that certain signals preceded a foreign country's missile launch, those signals can serve as key indicators when they reappear. "It becomes sort of our early warning system," he explained.

This scenario supports Panzer's assertion that in event processing, "proximity in time is as important as proximity in space." Scrutinizing temporal connections can even help organizations like the DEA and Coast Guard foil drug or weapon smugglers. A suspicious low-flying aircraft from South America may be empty of cargo upon inspection, but with GEP, agents can determine that the aircraft crossed a particular stretch of ocean at the same time that a fishing boat was in the area, and search the vessel for transferred contraband.

The Future of GEP

Panzer anticipates lots of potential for commercial applications of GEP technology, such as incorporating proximity-triggered actions in mobile marketing and social networking tools. For example, retailers might send coupons or ads to a shopper's cell phone as they approach a particular mall. "They're taking into account not only your buying profile, but maybe things about your demographic . . . and also how close are you to a brick-and-mortar store."

Panzer also foresees an elimination of data silos, noting that many analysts in the intelligence community already work with multiple sources of data. He likened this data-fusion approach to the way that humans naturally live their lives, using information from all their senses in conjunction. "That hasn't been the case, for the most part, in intelligence processing," he observed.

As for ObjectFX offerings, Panzer stressed the importance of time-based correlation; he expects GEP products that are more temporally, as well as spatially, sophisticated. He also hinted that the next Spatial Rules release may include expanded "thresholding" features — meaning that the system issues alerts when the number of events in a particular context drops below, or rises above, expected levels.

Bookmark it: digg del.icio.us technorati yahoo facebook twitter

Add Comment