U.S. Department of Energy

Pacific Northwest National Laboratory

Hierarchical Signature Detection in High-throughput Environments

Principle Investigator: 

Hierarchical signature detection uses an integrated approach that leverages both disaggregation techniques and Bayesian evidence accrual methods to accelerate detection tasks in distributed systems defined by high-throughput, multi-INT data.

Approach

Hierarchical signature detection (HSD) accelerates detection tasks through a two-step process. First, HSD disaggregates primary detection tasks into intermediate tasks that can be distributed to the sources of data collection (e.g., various sensors and instrumentation). Second, we leverage the results of these intermediate tasks to create early situational awareness for the primary detection task. This early awareness allows for hierarchical systems to deploy contingency and mitigation plans earlier, and allows such systems to operate more flexibly and efficiently.

Impact

Hierarchical signature detection enables a wide range of users to push sophisticated analytics out to the edges of their networks and will allow new and existing sensors to be transformed from simple collection devices into intelligent distributed signature detection platforms. We believe the ability to support scalable analytics for detection and monitoring will broadly interest many efforts within the Laboratory and be of keen interest to many of PNNL's clients who reside across a variety of domains include power grid, cyber systems, high-energy physics, and human health.

Project Staff: 
Luke Gosink, Paul Bruillard, Jim Fast, Ken Jarman, Lynn Wood
| Pacific Northwest National Laboratory