eduzhai > Applied Sciences > Engineering >

Adversarial Filters for Secure Modulation Classification

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 17 pages

Abstract: Modulation Classification (MC) refers to the problem of classifying themodulation class of a wireless signal. In the wireless communications pipeline,MC is the first operation performed on the received signal and is critical forreliable decoding. This paper considers the problem of secure modulationclassification, where a transmitter (Alice) wants to maximize MC accuracy at alegitimate receiver (Bob) while minimizing MC accuracy at an eavesdropper(Eve).The contribution of this work is to design novel adversarial learningtechniques for secure MC. In particular, we present adversarial filtering basedalgorithms for secure MC, in which Alice uses a carefully designed adversarialfilter to mask the transmitted signal, that can maximize MC accuracy at Bobwhile minimizing MC accuracy at Eve. We present two filtering based algorithms,namely gradient ascent filter (GAF), and a fast gradient filter method (FGFM),with varying levels of complexity.Our proposed adversarial filtering based approaches significantly outperformadditive adversarial perturbations (used in the traditional ML community andother prior works on secure MC) and also have several other desirableproperties. In particular, GAF and FGFM algorithms are a) computationalefficient (allow fast decoding at Bob), b) power-efficient (do not requireexcessive transmit power at Alice); and c) SNR efficient (i.e., perform welleven at low SNR values at Bob).

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×