Birds of a Feather Flock Together: Fuzzy Extractor and Gait-Based Robust Group Secret Key Generation for Smart Wearables

The recent surge in the usage of smart wearables for health monitoring highlights securing the communication among a group of personal devices using group secret keys (GSK). Simultaneous GSK generation on multiple wearables is very challenging as finding a common feature among the devices that has good entropy is difficult. In this paper, we present two novel GSK protocols - FEAT-GSK and FEST-GSK, employing the unique gait characteristics of a person and fuzzy extractors. FEST-GSK eliminates the reconciliation and privacy amplification stages as it employs error correcting code and strong extractor. We implement our protocols on android devices and conduct various experiments. Our results demonstrate that the gait features extracted on user's devices show highest correlation (Pearson-correlation-coefficient >0.9), and guarantees matching group key generation e.g., 256-bit key in less than 4 s, whereas, the adversaries show as low as 20% key agreement with respect to the user.