Page 363 - Fundamentals of Radar Signal Processing

(5.63)

Dividing Eq. (5.62) by Eq. (5.63) gives the output SIR; further dividing that
ratio by the input SIR gives the improvement factor for a two-pulse canceller
operating against clutter only (no noise)

(5.64)

Since ρ in the matrix formulation is the same as ρ [1], this is the same
c
c
expression obtained using autocorrelation methods in Eq. (5.57). The same
result could also have been obtained in the vector analysis by simply using the
T
target model vector averaged over Doppler, t = [1 0] .
Additional MTI metrics can be defined. Improvement factor I is an average
of the improvement in signal-to-clutter ratio over one Doppler period. At some
Doppler shifts, the target is above the clutter energy, while at others it is below
the clutter and therefore not detectable. I does not indicate over what percentage

of the Doppler spectrum a target can be detected. The concept of MTI visibility
factor or target visibility V has been proposed to quantify this effect
(Kretschmer, 1986). V is the percentage of the Doppler spectrum over which the
improvement factor for a target at a specific frequency is greater than or equal to
the average improvement factor I. A related metric is the usable Doppler space
fraction (UDSF), which in turn is determined by the minimum detectable

velocity (MDV) or minimum detectable Doppler (MDD). These metrics are
common in space-time adaptive processing, so their discussion is deferred to
Chap. 9.

5.2.6 Limitations to MTI Performance
The basic idea of MTI processing is that repeated measurements of stationary
clutter yield the same echo amplitude and phase; thus successive measurements,
when subtracted from one another, should cancel. Any effect internal or external

to the radar that causes the received echo from a stationary target to vary will
cause imperfect cancellation, limiting the improvement factor.
Perhaps the simplest example is transmitter amplitude instability. Consider
a two-pulse canceller and suppose that the amplitude of each pulse may differ in
amplitude from the nominal amplitude by up to ±5 percent (equivalent to 20
log (1.05/1) = 0.42 dB). The signal resulting from subtracting two echoes from
10
a perfectly stationary target can have an amplitude that is as large as 10 percent