Spectrum models (1900.5.2)

A frequency assignment tells you a frequency, a place and a time. It doesn't tell you how much energy spills next door, or how much a receiver can take before it fails. IEEE 1900.5.2 spectrum consumption models capture both, for any transmitter, receiver or system, in a form machines can compare.

What a model says

A transmitter's model says how much energy it puts out, at which frequencies, in which directions, where and when. A receiver's model says how much interference it can take at each frequency, where and when. Put the two side by side and you can answer the question every spectrum manager asks: will these two work together?

Masks, in one picture

Power margin

Power margin is the number NEMESynth-EMSO's spectrum decisions turn on. It is the amount, in dB, by which the interference a receiver would get exceeds what its underlay mask allows. At 0 dB or below, the two systems are compatible. Above 0 dB, it says how much has to change: a margin of 8 dB is 8 dB of separation, filtering, power or antenna gain to find.

The standard computes it in two settings:

• From the masks alone, the margin is the attenuation the path between the two must supply through antennas, distance and terrain. It answers "how far apart must these be?" before anyone picks a site.
• From the masks scaled by the link budget, it says whether the two are compatible where and when they will actually operate.

The link budget that scales the masks

Both models carry a propagation map, so each pair has to choose one. Conservative, the default, takes the map that predicts the least loss, and so the most separation. Receiver dominant uses the receiver's map when the receiver marks it dominant. Optimistic uses whichever map is marked dominant and, when both are, the one that predicts the most loss, for the most reuse. Which approach an enterprise uses is a policy decision, and it changes the answer.

Two ways to compute it

Each underlay mask names the method that applies to it.

The receiver's vulnerability masks

A receiver model can carry several underlay masks, because a receiver tolerates some kinds of interference better than others. The margin uses the least restrictive mask whose conditions the interferer meets.

• Duty-cycle rated. Brief, narrow interference is tolerated at a higher level. Each rating gives a duty cycle, a maximum dwell time and a level adjustment. When several such interferers arrive, their duty cycles and dwell times add before a rating is chosen.
• Hop rated. A frequency-hopping receiver that can lose a stated fraction of its hops.
• Protocol or policy rated. A transmitter that follows a stated protocol, such as listen before talk, is compatible by its behavior. Such a mask can be marked independent, so its interferers add nothing to the aggregate.
• Intermodulation mask. The receiver's vulnerability to the products of two or more signals mixing, mostly from transmitters on the same platform or site.

Masks can carry a confidence rating too, so a set of them describes how the tolerable interference is distributed instead of drawing one line.

The interferer's power masks

On the transmitter side the spectrum mask gives the signal's power density across frequency, relative to the reference power, and the power map gives its antenna gain by direction. A spectrum mask can describe a continuous signal, pulses or frequency hops. For a hopping signal, the duty cycle inside the receiver's band comes from its dwell time, its revisit period and the share of its hop frequencies that land in that band.

Prediction and 1900.2

IEEE 1900.2 is the analysis the margin belongs to. Its measurement events can be measured or calculated, and the first of its harmful-interference categories, conceivable interference, is established by analysis alone. A power margin computed from two models is a calculated measurement event. Whether a positive margin makes an interference event, and how many make harmful interference, are criteria the analysis has to state and justify. 1900.2 also asks who owns a receiver's protection margin, the decibels it keeps against fading: the receiver, or the interferers that would like to use it.

Where and when

Models carry their own geography and schedule: a point, a circle around it, a polygon for an area of operations, or a path for a convoy or an aircraft. A schedule says when the model applies.

In NEMESynth-EMSO

SFAF says what was authorized. Spectrum models say what that authorization means for everyone else nearby. NEMESynth-EMSO computes margins today from SFAF records and link budgets, and will compute them from the models' masks as SFAFx carries them. 1900.5R uses the same terms for spectrum use at every level.

• Interference to noise. For every pair within reach, the interference at the receiver is the transmitter's EIRP, less ITM path loss over terrain, plus receive antenna gain, less a frequency-dependent rejection. Today the rejection comes from how much of the transmitter's bandwidth overlaps the receiver's, with a fixed rejection for an adjacent channel. An interference-to-noise ratio at or above the protection criterion, −6 dB by default, makes the pair a conflict.
• Aggregate interference. The optimizer adds up, in watts, the interference at each receiver from every transmitter it has chosen (the mean power), and adds a constraint whenever the sum crosses the criterion.
• Link margin on the clock. Each PACE step's link margin is checked against the margin it needs at every change in the scenario. A short margin is one of the conditions that moves a net to its alternate.