1st International ICST Workshop on Secure Wireless Networks

Research Article

Achievable secrecy rates for wiretap OFDM with QAM constellations

  • @INPROCEEDINGS{10.4108/icst.valuetools.2011.245791,
        author={Francesco Renna and Nicola Laurenti and H. Vincent Poor},
        title={Achievable secrecy rates for wiretap OFDM with QAM constellations},
        proceedings={1st International ICST Workshop on Secure Wireless Networks},
        publisher={ACM},
        proceedings_a={SECURENETS},
        year={2012},
        month={6},
        keywords={OFDM; wiretap channel; secrecy capacity},
        doi={10.4108/icst.valuetools.2011.245791}
    }
    
  • Francesco Renna
    Nicola Laurenti
    H. Vincent Poor
    Year: 2012
    Achievable secrecy rates for wiretap OFDM with QAM constellations
    SECURENETS
    ICST
    DOI: 10.4108/icst.valuetools.2011.245791
Francesco Renna1, Nicola Laurenti1,*, H. Vincent Poor2
  • 1: Department of Information Engineering - University of Padova
  • 2: Department of Electrical Engineering - Princeton University
*Contact email: nil@dei.unipd.it

Abstract

Orthogonal frequency division multiplexing (OFDM) systems have enjoyed widespread adoption as the physical layer standard for high data rate wired and wireless networks, due to their ability to efficiently cope with slowly varying dispersive channels. Therefore in searching for feasible implementations of physical layer security techniques, it is appropriate to analyze how existing information theoretic results can be applied to the OFDM structure This paper considers the information theoretic secrecy rates that are achievable in a wiretap OFDM channel when transmitting QAM constellation symbols. The loss with respect to the secrecy rates obtained with Gaussian distributed inputs is evaluated for both finite constellation cardinalities and in the asymptotic approximation of arbitrarily high cardinality. Moving from the insight gained with this analysis, we propose bit-loading strategies to efficiently allocate the appropriate number of bits in each subchannel, by considering the twofold objective of minimizing the loss with respect to the Gaussian input secrecy capacity and minimizing the total bit load.