ISR & Airborne Intelligence
Persistent, wide-area airborne intelligence, from tasking through dissemination, managed as a complete program.
Glossary
Processing, Exploitation & Dissemination.
Processing, exploitation and dissemination is the work that turns raw sensor collection into something a commander can act on: converting the signal into usable imagery or data, analyzing what it shows, and delivering the result to the people who need it, in time for them to do something about it. It is the largest recurring cost in most airborne intelligence programs and the one most often left out of the budget.
The part of ISR that is not the aircraft
A sensor produces a signal. PED is everything between that signal and a decision.
Processing is the mechanical conversion — georeferencing imagery, decompressing and indexing video, correcting radar returns, applying calibration. Largely automated, and the part people assume is all of it.
Exploitation is the analysis. A trained analyst watches the feed or works the imagery and determines what is actually happening: that the vehicle at the crossing is the same one seen two nights ago, that the pattern around a compound has changed, that a track is fishing rather than smuggling. This is human work, and it is the expensive part.
Dissemination is delivery. Getting the assessment to the person who can act, in a format they can use, in the language they work in, before the information expires.
Why it is the cost that surprises people
An aircraft is a capital decision with a visible price. PED is a headcount decision that recurs forever, and it scales with the sensor rather than with the airframe.
A single electro-optical turret watching one point produces a feed one analyst can monitor. A wide-area sensor covering a city produces imagery that a dozen analysts could work for a week and still not exhaust. The aircraft cost the same in both cases. The PED cell did not.
The arithmetic that catches programs is shift coverage. Twelve hours of daily collection does not mean twelve analyst-hours. It means a watch rotation, a supervisor, someone doing post-mission exploitation after the aircraft lands, and enough depth that leave and attrition do not stop the mission. A program flying two orbits usually needs a PED cell in the high teens, not four people.
The failure mode
The common outcome is a program that collects far more than it exploits. The aircraft flies, the archive fills, and a small cell works the top of the queue while everything underneath it ages out. Coverage statistics look excellent. Nothing reaches a commander in time.
This failure is quiet, because the measurable outputs — sorties flown, hours collected, terabytes archived — all look healthy. The number that matters is different: what proportion of collection reached someone who could act on it while it was still actionable.
How to size it before you buy
Work backwards from the decision, not forward from the sensor.
Establish who receives the product and what they will do with it. Establish how quickly it stops being useful — minutes for an interdiction cue, days for a pattern assessment. That latency requirement sets whether exploitation has to happen onboard, at a ground station in real time, or in post-mission analysis.
Then size the cell to the shift pattern that latency demands, and only then choose the sensor. A program that picks the sensor first has already fixed its analyst bill without knowing it.
Where it can be reduced, and where it cannot
Automation genuinely helps at the processing layer and at triage — change detection, motion indexing, and flagging the segments worth a human’s attention. Those tools cut the volume an analyst has to look at.
What they do not do is replace judgment about intent. Deciding whether an observed pattern means what it appears to mean is analytical work, and the programs that assumed software would absorb it have generally had to hire the analysts anyway, later and in a hurry.
How Alpha Federal treats it
We size the PED cell in the same conversation as the airframe and the sensor, and we say what it will cost to run rather than quoting the aircraft and leaving the rest for later. Where a partner nation intends to own the capability, the analyst pipeline is part of the capability transfer from the start, because an institution that can fly but cannot exploit has bought a surveillance aircraft and not an intelligence capability.
Common Questions
What does PED stand for in intelligence?
Processing, exploitation and dissemination. It is the work between a sensor producing a signal and a commander acting on it: converting the collection into usable form, analyzing what it shows, and delivering the result to whoever can act while it is still current.
Why is PED the largest cost in an ISR program?
Because it is recurring headcount rather than a one-off purchase, and it scales with the sensor rather than the aircraft. A wide-area sensor produces imagery a dozen analysts could work for a week; a single turret produces a feed one analyst can monitor. The airframe cost the same in both cases.
How many analysts does an ISR program need?
It is set by the shift pattern the latency requirement demands, not by the aircraft count. A program flying two orbits with a real-time requirement typically needs a cell in the high teens once watch rotation, supervision, post-mission exploitation and cover for leave and attrition are included.
Can automation replace PED analysts?
It genuinely helps at the processing and triage layers, with change detection, motion indexing and flagging the segments worth a human's attention. It does not replace judgment about intent. Programs that assumed software would absorb that have generally had to hire the analysts later and in a hurry.
How do you size a PED cell before choosing a sensor?
Work backwards from the decision. Establish who receives the product, what they will do with it, and how quickly it stops being useful. That latency sets whether exploitation happens onboard, in real time at a ground station, or post-mission, and only then does the sensor get chosen.