![](bg3.png)
device, but provide no experimental results. A
wearable health-monitoring device using a Personal
Area Network (PAN) or Body Area Network (BAN)
can be integrated into a user’s clothing (Park and
Jayaraman, 2003), like Foster-Miller’s health
monitoring garment for soldiers. Along these lines,
Paradiso (Paradiso, 2003) describes preliminary work
on the WEALTHY system, a garment with embedded
ECG sensors for continuous monitoring of the heart.
Jovanov et al present in (Jovanov et al., 2005) a
wireless BAN with motion sensors for computer-
assisted physical rehabilitation and ambulatory
monitoring. In (Kemp et al., 2008), Kamp et al
develop a wearable system for manned bomb disposal
missions. Mihovska and Prasad (Mihovska and
Prasad, 2007) have developed an adaptive security
architecture for personal networks with an
asymmetric key agreement scheme on three levels by
using contextual information, such as the location of
the user and the capability of the devices. This
architecture is based on an elliptic curve
cryptosystem. It has, however, one shortcoming. It is
susceptible to impersonation via key compromise.
A global notice about these approaches shows that
traditional communication protocols are used to
transmit the collected data from the human body to an
external system (e.g., cellphone, laptop).
Unfortunately, this does not guarantee a real-time
transmission of this information since an important
variable delay can occur, especially when some
sensors transmit large units of data such as images.
Moreover, due to the use of radio communication, the
confidentiality of the transmitted data is not
intrinsically guaranteed, which may lead to privacy
violation. In several applications, including
healthcare, even the identity of the wearer should be
hidden.
3 PROPOSED WISSN
ARCHITECTURE
In this paper, we address two crucial issues regarding
wearable sensor systems:
•Improving Real-time Responsiveness: This is
achieved by building special communication frame
structures based on the non-uniform multiplexing of
the data generated by different types of sensors
•Combining Sensor Authentication and user
anonymity through the use of lightweight
cryptographic protocols: In order to adapt to the
severe resource limitations characterizing WISSNs,
we use an elliptic curve implementation of the
proposed security functions
In spite of its apparent simplicity, WISSNs exhibit
several complex features and therefore require
sophisticated engineering approaches in order to be
set up. In the following, we list the most relevant
factors that may shape the communication models
used in smart sensor networks.
1. Multi-functional framework: A sensor node
may be able to carry out multiple functions that can
be set on/off depending on the situation. Obviously,
the communication requirements may differ greatly
from one functionality to another according to the
data sent across the WISSN. For instance, when the
network is deployed in a mining structure, a first
category of sensor may be used to monitor the amount
of several toxic gases in the atmosphere. A second
type of sensor can serve to estimate the opacity of the
encountered obstacles. IRM sensors can be used in
such a context in order to predict, and possibly
prevent, disasters. Since the volume of data generated
by the latter category is by far greater than that
generated by the former, much more bandwidth must
be reserved to transmit image data.
2. Independent monitoring capability: Due to the
non-uniform nature of the monitored events
(irrespective of the application), some sensors may
exhaust their energy more rapidly than others. This
may result in the presence of uncovered regions
where the nodes in charge of gathering data related to
the environment are out of power. Since such a
situation significantly affects the efficiency of the
WISSN, solutions should be proposed to avoid it. One
alternative is to tune the quality of the data gathered
by a sensor node according to its residual energy
resources. This would extend considerably the
lifetime of this node at the cost of losing some refined
data, which is definitely better than totally losing the
functionalities provided by the node. As a result the
communication resources required to transmit the
data may vary from one sensor to another.
3. Exportable configuration: Configurations can
be exported from one sensor to another in order to
turn on/off several functionalities. Even though this
feature allows energy to be saved (by triggering
power-consuming time only when necessary), it
creates a significant security hole since node
imposture can be easily carried out. Hence,
authentication mechanisms should be set up to
prevent non-authorized nodes from manipulating the
WISSN. Two important issues must be taken into
consideration: First, the security algorithms must be
based on non-complex algorithms and use small
cryptographic credentials (to adapt to limited CPU
time and memory resources) and; Second For a wide
range of applications, the anonymity of the person
holding the wearable or implantable smart sensor
system should be preserved. Since this conflicts with
authentication, specific security infrastructures will
SECURE WEARABLE AND IMPLANTABLE BODY SENSOR NETWORKS IN HAZARDOUS ENVIRONMENTS
87