Page 193 - Fundamentals of Radar Signal Processing
P. 193

FIGURE 3.8   Datacube illustrating one CPI of data from a multichannel pulsed
               radar.



                     How large is a datacube? The number of samples in each dimension are
               determined  by  the  characteristics  of  the  desired  radar  measurements.  The
               number L of range samples is simply the length of the range swath divided by
               the range bin spacing, L  = R /ΔR . The swath length is determined by mission
                                                  w
                                                       s
               requirements,  while  the  range  bin  spacing  is  determined  primarily  by  the
               waveform bandwidth as seen in Eq. (3.1). Both may vary significantly for the

               same  radar  as  it  switches  between  various  operating  modes  with  different
               search ranges and range resolutions.
                     One  important  determinant  of  the  number  of  pulses M  in  a  CPI  is  the
               desired  Doppler  resolution  (equivalently,  velocity  resolution).  The  Doppler
               spectrum  is  the  DTFT  of  the  slow-time  data.  The  duration  of  the  slow-time
               signal is the CPI length of MT seconds. The Doppler resolution will therefore
                                                      5
               be on the order of ΔF  = 1/MT,  giving the required number of pulses as M =
                                          D
               1/ΔF T  = PRF/ΔF .  Thus, M  depends  on  the PRF  as  well  as  the  Doppler
                     D
                                       D
               resolution, and can vary widely. In pulse-Doppler processing for basic target
               detection  and  tracking, M  is  frequently  a  small  number  of  tens  of  pulses.
               However, in fine-resolution imaging it can be hundreds or even thousands of
               pulses.

                     For a multichannel receiver, the number N of channels is more difficult to
               characterize. A  phased  array  antenna  with  a  receiver  per  element  may  have
               hundreds or thousands of phase centers, each constituting a receiver channel. A
               subarrayed architecture may have many fewer, perhaps ranging from as little as
               three or four to a few tens. A monopulse antenna has three phase centers. The
               antenna type, size, and architecture all significantly influence N.
                     The  datacube  view  of  a  CPI  of  data  from  a  multichannel  pulsed  radar

               provides a good conceptual model for understanding most digital radar signal
               processing  operations.  Many  of  the  basic  radar  signal  processing  operations
               considered  in  the  remainder  of  this  text  correspond  to  processing  one-
   188   189   190   191   192   193   194   195   196   197   198