December 4, 2012 | Written by: John Palfreyman
The proliferation of increasingly capable sensors presents a threat and an opportunity to Defence & Intelligence organizations. Commercial Big Data technology can offer significant improvements in dealing with the ever-increasing volume, velocity and variety of sensor data.
FROM: Let’s use Passive Sonar as an example of how current systems deal with the real time / low latency needs of military sensor data processing. The simplified system shown in the Figure 1 classifies & tracks vessels based on the detection and analysis of their passive noise signature. New track intercepts are compared with a library of known target signatures to effect classification.
Figure 1 – Simplified Passive Sonar Block Diagram
Current generation of systems uses Digital Signal Processing Hardware & Firmware to do this; this works fine, but it’s difficult, expensive and time consuming to change as new analysis techniques are needed to cope with threat evolution and mission change.
TO: Figure 2 shows the application of Big Data technologies to this challenge. The Signal Processing elements can be programmed in a Stream Computing array, such as InfoSphere Streams. This copes with the real time / low latency needs of analyzing and classifying the acoustic signatures.
We can also apply Infosphere BigInsights to look for patterns in the stored, historical data that may represent – for example – gradual threat signature changes over time.
Figure 2 – Application of Big Data Technologies
The combination of these two technologies gives the ability to make the analysis & classification processes ADAPTIVE; results of the slow time analysis of historical data can be used to condition the real time Streaming analysis without disruption to the processing flow. The same would apply to most sensor types including radar, electro-optic, electronic warfare etc.
The sensor systems are thus better able to cope with changing mission needs and the sophisticated threat, deriving timely actionable intelligence from the available sensors.