IEEE Consumer Electronics Magazine - January/February 2022 - 7

enough to be worn on or implanted within the
human body.1 These devices enable continuous
patient vital sign monitoring in addition to
remote control of medical devices such as insulin
pump, pacemaker, or continuous glucose
monitoring. Due to their nature (i.e., limited
power resources, memory usage, and computational
capabilities) and the environment, they
operate within, securing these devices along
with their network is crucial to the success and
wide adoption of this technology.
Authentication is a crucial process in safeguarding
BANs and their operations such as the
transmission of patients' health data or receiving
commands. Many lightweight authentication
schemes have been proposed over the years to
ensure the required security level for BAN sensors
while meeting their stringent constraints.2
Authentication enables BAN sensors or devices
to prove their identity for verification by completing
an authentication process or challenge. Without
any form of authentication, BAN sensors are
vulnerable to various security attacks that may
be carried out by a motivated adversary. Pirbhulal
et al.3 outline the security requirements for
BANs-based smart healthcare and highlight the
unique challenges that need to be overcome to
develop robust biometrics based authentication
schemes for BAN. Moreover, using fingerprints as
an authentication mechanism to unlock mobile
devices for payment authorization purpose was
investigated by Paul and Irvine.4 This study raised
some concerns about fingerprint-based authentication
and potentially other biometrics, such as
electrocardiogram (ECG), due to their associated
vulnerabilities. Issues discussed include how fingerprints
may be captured, how users cannot easily
avoid leaving their fingerprint behind and how
once an adversary compromises it, the revocation
is impossible. A review of the developed
authentication protocols for deployment within
implanted medical devices (IMDs) was presented
by Challa et al.5 This article outlines the security
challenges and functionality requirements for
authentication schemes targeting IMDs that can
be classified as either proximity-based, biometric-based,
or a combination of the two.
Fuzzy vaults are one such scheme that is being
used to facilitate a key agreement between communicating
BAN nodes to achieve a robust
January/February 2022
authentication. Fuzzy vault6 is a cryptographic
construction that allows a user to conceal a secret
value (e.g., message, password, or cryptographic
key) by locking it using elements obtained from a
publicly known universe. Any attempt to unlock
the vault and uncover the secret contained within
it requires significant overlap between the set of
elements used to lock the vault and the set being
used to attempt unlocking it. The order in which
these elements are used when unlocking the vault
has no impact on success due to the fuzzy vault's
possessing order invariance. These two properties
make unlocking a fuzzy vault possible without
a perfect or near-perfect recall of the elements
used to lock the secret. This may appear as a
security vulnerability, however, the fuzzy vault
scheme is more suitable for environments where
shared knowledge of the elements used to lock
the vault originates from a source that is incapable
of 100% successful capture. An appropriate
example of such a source is " fingerprint minutiae " 7
where the positioning of the finger, the moisture
present and any movement can have a
significant impact on the sensor's ability to
extract identifying markers. Even successive readings
of thesamefinger by thesamedevice can
vary to such an extent that they would appear to
be from separate users.
FUZZY VAULT OVERVIEW
The fuzzy vault scheme was first introduced
by Juels and Sudan6 to conceal a secret ðSÞ using
a set of elements A known at the time of vault
construction. Once locked, the vault can only be
unlocked using another set of elements B that
has significant overlap with the set A. The fuzzy
vault can be constructed in such a way to allow
unlocking through the use of either error correction
codes or polynomial interpolation. For the
sake of simplicity, we will be using polynomial
interpolation, specifically Lagrangian Interpolation.
The vault is constructed as follows. 1) Generate
an nth order univariate polynomial P over
the indeterminate x whose coefficients are an
encoding of the secret S. 2) Evaluate PðXÞ using
the elements of the set A creating a new set
R ¼fAi;PðAiÞg. 3) Randomly generate a set of
chaff points C which do not share the same x
with any point within R or any other chaff point,
7
IEEE Consumer Electronics Magazine - January/February 2022

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - January/February 2022
