Performance of convolutional codes with diversity under worst case partial-band noise jamming

The performance of convolutional codes with diversity are analyzed under worst-case partial-band jamming. Various receiver structures and decoders that do not use any form of side information are considered and their performances are compared using Chernoff bounds. The performance of systems considered in the literature that use perfect side information are given for comparison purposes. It is shown that good robust soft metrics perform very close to square-law-combining metrics with perfect side information. The self-normalized metric performed especially well, with performance almost identical to that of the square-law-combining metric with perfect side information.<>