Secure multicarrier DS/SS with induced random flipping

For secure transmissions, we consider direct sequence/spread spectrum (DS/SS) systems where pseudo-random (PN) sequences are used for spreading. To generate PN sequences, linear feedback shift registers (LFSRs) are employed, which are also used to generate keystreams in stream ciphers. For steam ciphers that use LFSRs, there are known attacks including correlation attacks that can regenerate keystreams. Since those attacks can also be used to regenerate spreading sequences, in this paper, in order to mitigate this problem, we consider induced random (chip) flipping of spreading sequences. While perturbed spreading sequences make correlation attacks infeasible, a legitimate receiver also suffers from random flipping. To take advantage of known spreading sequences for good performances, the maximum likelihood (ML) detection can be used, which is, however, computationally prohibitive as the complexity grows exponentially with the processing gain. To avoid this difficulty, we derive an expectation-maximization (EM) algorithm to perform the ML detection with low computational complexity in this paper.