Page 192 - Fundamentals of Radar Signal Processing

and use only the data received after steady state is achieved. Conversely, if it is

desired to have a certain number of steady state PRIs for subsequent processing,
the number of pulses is often augmented by the necessary number of clutter fill
pulses.
It is tempting to conclude that range ambiguities could be resolved by
observing whether or not a target detection appears in all of the pulses. A target
detection missing from the first n pulses in a CPI suggests that the actual range is

the apparent range plus n times the unambiguous range. This idea will work if
detection algorithms are applied to the fast time data for each pulse separately
and the SNR is high enough that the probability of missed detections is small.
However, it is rare to use a CPI of data in this manner. More commonly, the
SNR of the single-pulse data is not adequate for reliable detection so that it is
not known whether the target is absent in the first n pulses. Instead, the slow
time data will be coherently or noncoherently integrated in order to obtain an

adequate SNR.
In addition to creating the possibility of range ambiguities, the use of
multiple pulses also aggravates the eclipsing phenomenon. A target at any
integer multiple of the range R ≈ cT/2, corresponding to time delays that are
ua
integer multiples of the PRI T, will produce an echo that arrives as the next
pulse is being transmitted. During this interval, the receiver will again be

isolated, so the target echo will be eclipsed. Targets at other time delays within
the interval (nT – τ, nT + τ) for any integer n will be partially eclipsed. Thus,
the pulse burst creates a series of blind zones in range or time delay. Targets in
these blind zones will be difficult or impossible to detect, even when they have
adequate SNR. Techniques to overcome this limitation are discussed in Chap. 5.

3.1.4 Multiple Channels: The Datacube
Some radars, but by no means all, have antennas that provide multiple

simultaneous outputs. The most obvious example is a system using a phased
array antenna with multiple subarrays, each having its own receiver, or even
with one receiver per array element in some cases. Each receiver will generate
a matrix of data like that of Fig. 3.2b for every pulse burst. The complete set of
data y[l, m, n] from all N channels is called a datacube and is illustrated in Fig.
3.8. The third dimension is often referred to as the receiver channel or phase

center dimension. Another type of system that generates a datacube uses a
monopulse antenna, common in some types of tracking radars. A monopulse
antenna has three output channels and so generates a datacube having N = 3
layers. Radar data is often explicitly organized in the processor memory in a
datacube format, i.e., as a three-dimensional structure of complex-valued data.