Page 503 - Fundamentals of Radar Signal Processing
P. 503

ground  clutter  renders  the  interference  nonhomogeneous  in  Doppler  and

               possibly in range.


               6.5.3   Analysis of Cell-Averaging CFAR
               The intent of the adaptive calculation of the threshold is to provide a constant
               false alarm rate despite varying interference power levels. In this subsection,
               the detector performance is analyzed for the case of a square law detector to see
               if this goal has been achieved. The threshold computed according to Eq. (6.129)

               will  be  a  random  variable,  and  therefore  so  will  be  the  probability  of  false
               alarm. The detector will be considered to be CFAR if the expected value of P                 FA

               does not depend on the actual value of            .
                     Combining Eqs. (6.128) and (6.129) gives an expression for the estimated
               threshold







                                                                                                     (6.130)

               Define z  = (α/N)x ;  thus               Standard results from probability theory and
                                     i
                         i
               Eq. (6.124) give the PDF of z      i





                                                                                                     (6.131)

               The PDF of   is the Erlang density











                                                                                                     (6.132)


               The P  observed with the estimated threshold will be                           . This is now
                      FA
               also a random variable; its expected value is










                                                                                                     (6.133)


               Completing this standard integral and performing some algebraic manipulations
   498   499   500   501   502   503   504   505   506   507   508