Low-cost mitigation of privacy loss due to radiometric identification

  • Authors:
  • Jason J. Haas;Yih-Chun Hu;Nicola Laurenti

  • Affiliations:
  • Sandia National Laboratories, Albuquerque, NM, USA;University of Illinois at Urbana-Champaign, Urbana, IL, USA;University of Padova, Padova, Italy

  • Venue:
  • VANET '11 Proceedings of the Eighth ACM international workshop on Vehicular inter-networking
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Recently, there has been much interest in using radiometric identification (also known as wireless fingerprinting) for the purposes of authentication. Previous work has shown that using radiometric identification can discriminate among devices with a high degree of accuracy when simultaneously using multiple radiometric characteristics. Additionally, researchers have noted the potential for wireless fingerprinting to be used for more devious purposes, specifically that of privacy invasion or compromise. In fact, any such radiometric characteristic that is useful for authentication is useful for privacy compromise. To date, there has not been any proposal of how to mitigate such privacy loss for many of these radiometric characteristics, and specifically no such proposal for how to mitigate such privacy loss in a low-cost manner. In this paper, we investigate some limits of an attacker's ability to compromise privacy, specifically an attacker that uses a transmitter's carrier frequency. We propose low-cost mechanisms for mitigating privacy loss for various radiometric characteristics. In our development and evaluation, we specifically consider a vehicular network (VANET) environment. We consider this environment in particular because VANETs will have the potential to leak significant, long-term information that could be used to compromise drivers' personal information such as home address, work address, and the locations of any businesses the driver frequents. While tracking a vehicle using visually observable information (e.g., license plates) to obtain personal information is possible, such means require line-of-sight, whereas radiometric identification would not. Finally, we evaluate one of our proposed mechanisms via simulation. Specifically, we evaluate our carrier frequency switching mechanism, comparing it to the theory we develop, and we show the precision with which vehicles will need to switch their physical layer identities given our parameterization for VANETs.