FAS | Intelligence | Programs | Process |||| Index | Search |


FAS Project on Intelligence Reform

Electronic Warfare Flagging Analysis Expert System (EWFAES)

The Electronic Warfare Flagging Analysis Expert System (EWFAES) was developed by the Air Force to reduce the workload on ELINT analysts when identifying unknown emitters. Some signals may still require manual analysis in instances where analysis by the expert system does not produce an ID. However, those numbers are a small fraction of those entering the system.

While many have heard flagging defined as identifying signals whose parameters are outside expected or programmed limits, the practical value of flagging is that it reduces to a manageable level the number of signals needing analysis. During a contingency, thousands of TACELINT messages flow in daily. It is not possible to analyze that amount of traffic and make reprogramming decisions. The flagging function identifies out-of-limit signals and, thus eliminates 98 to 99% of the ELINT volume that must be analyzed.

Battlefield management depends heavily on EW assets to provide current threat analysis for successful deployment of war-fighting resources. One area that supports this analysis is target specific operating characterization and equipment/unit correlation which is normally a "brute-force/labor intensive" effort. Dense signal environment dictates the need for automated/reduced manual collection, signal ID and sorting, and first-level signal processing. This area should also allow for further discrimination or "flagging" of out-of-limits signals, especially threat systems whose measurable values already lend themselves to characterization.

To reduce the volume of ELINT traffic that is forwarded for analysis, flagging employs a substantial amount of automation. Typically, up to 10,000 signals per day in various formats must be evaluated to determine how each of these signals would be processed by radar signal detection sets (RSDS) and jammers. To accomplish this sizable task, signals are processed in several steps, with expert system technology employed throughout the process to minimize human intervention. ELINT intercepts arrive from the field in several formats. The first task for an automated system is to parse these messages into a single, consistent format. Of course, message traffic has been known to contain format errors. When the parser detects a message with a format error, it signals for help from an ELINT analyst. The analyst corrects the format and sends the message back into the system.

Before these signals are processed by the flagging models of the RSDS and jammers, one must know with high confidence what type of system emitted the signal. Signal IDs are not always correct in ELINT reports and sometimes the signals are not identified at all. Correct identification of signals is the bread and butter of flagging. When threat emitters change their operating parameters, the signal may not be recognizable any more and is reported without identification. It is the task of flagging to use all available means, parametric, geographic, or other intelligence sources such as imagery or COMINT, to correctly identify the emitter of an unknown signal. These are the signals that will most likely cause problems for aviators.

After all the signals are correctly identified, each is processed through software simulations, or models, of various EW systems. Those signals that are not processed correctly by the models are flagged for further analysis by the reprogramming center. That is where the term flagging originates. The software models are designed by the flagging engineer to simulate the identification algorithms of the EW system. The signals are processed by the model using the mission data sets (MDS) that are applicable. The model response to the signal is evaluated, and if the response is correct, the model determines why it is incorrect and which parameter is outside limits. The output of the flagging model is reviewed by an ELINT analyst who ensures the signal is valid and also by the flagging engineer who looks for problems with the model itself.

The addition of flagging to the reprogramming process has eliminated the potential log jam of ELINT traffic that could delay new mission data from getting into aircraft.

Sources and Methods



FAS | Intelligence | Programs| Process |||| Index | Search |


http://www.fas.org/irp/program/process/ewfaes.htm
Created by John Pike
Maintained by Steven Aftergood

Updated Friday, May 05, 2000 8:10:39 AM