2013 Seventh International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing

A Network Mobility Solution Based on 6LoWPAN Hospital Wireless Sensor

Network (NEMO-HWSN)

Mohammadreza Sahebi Borhanuddin Bin Mohd Pooria Varahram Antonio J. Jara

Shahamabadi Ali Dept. of Computer & Computer Science Faculty,
Dept. of Computer & Dept. of Computer & Communication Systems University of Murcia
Communication Systems Communication Systems Engineering Murcia, Spain
Engineering Engineering Faculty of Engineering, UPM [email protected]
Faculty of Engineering, UPM Faculty of Engineering, UPM University
University University UPM Serdang, Malaysia
UPM Serdang, Malaysia UPM Serdang, Malaysia [email protected]
[email protected] [email protected]

Abstract—IPv6 Low-power Personal Area Networks peer-to-peer, WiFi hot-spots, vehicular networks, sensor
(6LoWPANs) have recently found renewed interest because of networks, and many more.
the emergence of Internet of Things (IoT).However, mobility Wireless Sensore Networks (WSNs) have a wide range
support in 6LoWPANs is still in its infancy for large-scale IP- of potential applications, from environmental monitoring,
based sensor technology in future IoT. The hospital wireless
home and industrial automation, military, to education,
network is one important 6LoWPAN application of the IoT,
where it keeps continuous monitoring of patients’ vital signs transport, healthcare and many more. It is developed over
while the patients are on the move. Proper mobility IEEE 802.15.4 which is a layer_2 standard defined for
management is needed to maintain connectivity between personal area network (WPAN). WSN is designed for
patient nodes and the hospital network to monitor their exact infrastructure-less type of networks which does not require
locations. It should also support fault tolerance and optimize an established network to be set up unlike a cellular based
energy consumption of the devices. In this paper, we survey network. WSN is also designed to connect to the Internet via
IPv6 mobility protocols and propose some solutions which a suitable node called the gateway [4]. However, IEEE
make it more suitable to a hospital architecture based on 802.15.4 is defined to be of limited capabilities by way of
6LoWPAN technology. Our initial numerical results show a
smaller frame sizes, low memory capacity and data rate. It
reduction of the handoff costs on the mobile router which
normally constitute a bottleneck in such a system. We also was primarily designed for short range communications with
discuss important metrics such as signaling overload, efficient power management that will eventually created a
bandwidth efficiency and power consumption and how they low-power personal area networks (LoWPANs) to support a
can be optimized through the mobility management. large number of nodes with energy saving capability [5].
The Internet Engineering Task Force (IETF) defines IPv6
KEYWORDS-6LoWPAN, NEMO Mobility, Healthcare,
Wireless Sensor Networks, Hand-off. Low-power Personal Area Networks (6LoWPAN) which is
an IPv6-based LoWPAN on the basis of IEEE 802.15.4 for
I. INTRODUCTION communications with the Internet. With its vast address
Over the past two decades, communication networks space, 6LoWPAN allows global connectivity between a
have experienced tremendous growth and expansion all over large number of IPv6 intelligent devices over large areas.
the world. The explosive growth of many types of mobile The protocol also enables the nodes to be self-organized i.e.
devices such as smart phones, variations of tablet computers, can do self-detection, self-healing, and self-configuring,
and laptops, has fueled the demand for more bandwidth with without human intervention [4].
varying Quality of Service (QoS), with pervasive For the success of IoT in general, and for healthcare in
connectivity and at affordable costs [1]. These mobile particular, mobility support is essential [6]. Mobility is
devices are generally very powerful in themselves with ever required to maintain the fault tolerance of the network and
more innovative user interfaces, better information security full access to information regardless of their locations. In
and privacy, capability for higher end-to-end data transfer healthcare, some of the main applications for 6LoWPAN are
rate, streaming or interactive communications, and many for real-time monitoring of vital signs such as ECG
other features [2]. Mobile wireless network generally (electrocardiogram), heart rate, SPo2, blood pressure, weight
encompasses wireless sensor networks, ad-hoc and mesh and breathing rate of patient under treatment in the hospital.
networks and infrastructure based cellular networks. These Moreover, it is important that these monitoring could be
groups of networks can service a wide array of application done while the patients move around within the hospital [7].
areas such as the ubiquitous broadband access [3], mobile In addition, a mobility protocol needs to be reliable to
reduce end-to-end delays, packet losses, and network

978-0-7695-4974-3/13 $26.00 © 2013 IEEE 433

DOI 10.1109/IMIS.2013.157
failures under any conditions because of the criticalness of Information System (HIS) and users (physicians, surgeons
healthcare provisioning. Therefore, the aims of a portable and nurses). As shown in Fig. 1, each part of the hospital
monitoring system are: (i) to control and monitor the such as operating theatre, observation rooms and wards are
patients in any location, (ii) to store the information as the organised as a PAN is under network coverage to keep the
Knowledge Based System (KBS) in order to study and connectivity within nodes and the Internet. Each PAN with
survey symptoms and predict illness [8]. all the nodes belong to the same domain deployed with a
The design features of 6LoWPAN node like packet size 6LoWPAN Border Router (BR) to connect the Internet, HIS,
restrictions, energy and power restrictions and delays in the and other PANs via the network backbone [4].
reception of messages, have constrained host-based mobility
protocols such as MIPv6, HMIPv6, FMIPv6. The Mobile
Node (MN) which a patient would carry, is involved in most
of the mobility management signals which is power hungry
[9]. In this respect, Proxy MIPv6 (PMIPv6) is more
appropriate to support 6LoWPAN mobility than the host-
based solutions, but it has two short-comings: it cannot
support multi-hop and it requires the 64 bit network prefix to
be assigned to each MN [10].
Mobility solutions can give different kinds of efficiency
and performance depending on its applications. Thus,
hospital wireless sensor network based on 6LoWPAN
(HWSN6) has been defined for hospitals, with mobile
nodes, border routers and gateways. Although this mobility
solution has been designed for hospital applications, but the
energy constraint of mobile patient nodes which comprises a Fig. 1. Hospital network architecture
set of sensor nodes has not been considered in [4]. It also did
not consider the energy consumption of Mobile Routers A. Gateway and HIS Node
(MR) which will constrain the Personal Area Network
This gateway manages its domain, establishes
(PAN) lifetime.
connections between networks, and interconnects with each
From this brief discourse it is anticipated that 6LoWPAN
other through wireless or wired links. HIS is a system based
will become more popular in the near future. This is
on Open Services Gateway Initiative (OSGi) technology for
primarily because it has a vast address space that is well
the management of all the other systems from the hospital.
suited to individually address all objects that are connected
HIS saves the important monitoring information of all nodes
to the Internet. Nevertheless, power consumption is a serious
and provides information and services to the other systems
issue in 6LoWPAN, therefore mechanisms need to be sought
belonging to the hospital such as follows: management of
in order to optimize this resource. One example of a busy
alarms from the Monere systems, Electronic Health Record
device is the MR; it is a very complex device that manages
(EHR), health status, localization service, and directory
and performs significant mobility functions [1].
service [4].
In this paper, we propose a new mobility protocol for
mobile patient nodes, it comprises of a set of sensor nodes B. Monere System
that forms a single unit called mobile patient node in
Monere system is a BR that covers each part (domain)
HWSN6. In this scheme, the MR that acts as a coordinator
of a hospital and is a Mobile Data Collector (MDC) from the
manages the mobility scenario and PAN functions. This
patient sensors, similar to a sink node in each PAN. It is
mobility solution decreases the number of messages and
equipped with several interfaces that establish the
overheading, and prolongs the lifetime of a patient PAN via
connection with other networks like Bluetooth, cellular and
MR.
Ethernet and controls home automation (ZigBee, X10 and
This paper is organized as follows: A discussion on
EIB) and standards such as CANBus, Ethernet and Serial
system architecture is given in section II. Then the mobility
[11]. The area under the interconnected BRs is referred to as
scenario is reviewed in section 0. In section IV, our mobility
a PAN or domain. 6LoWPAN BR plays two roles: it is
mechanism scheme is evaluated. Finally, comparison three
identified as Home Agent (HA) responsible for buffering
models and conclusion are discussed in section V and VI
and forwarding packets to the MN, and as Foreign Agent
respectively.
(FA) coordinates visited network [4].
II. SYSTEM ARCHITECTURE
C. Patient Node
The hospital system architecture is made up of patients A mobile patient node is an active device which moves
(mobile node, a set of sensors), Monere system (local between multiple PANs. A set of sensors as one unit are
gateway or border router), Internet gateway, Hospital fixed on the patient's body (6LoWPAN MN) to measure and

434
collect health data continuously such as heart rate, SpO2,
peripheral and core body temperature, glucose etc. [12].
From Fig. 2, two types of devices as sensor nodes are
defined in IEEE 802.15.4: there are Full-Function Device
(FFD) and Reduced-Function Device (RFD) respectively.
FFDs are designed to support all network functionalities and
participate in peer-to-peer topologies with multi-hop
communications. The functionalities of RFD devices are
limited to measure physical parameters and to process non-
complex tasks in star topologies (they do not support multi-
hop communications). Normally each PAN coordinator
controls the PAN, which has the function of setting up and
maintaining the PAN (hence, only a FFD device can assume
the role of PAN coordinator) [10].

Fig. 3. Mobility scenario

2. Movement detection time: when an MN moves, it detects

that its link quality has degraded beyond a certain
threshold, it means the existing router is no longer
reachable, or a new access router is available [13].
Fig. 2. Patient node model 3. Entering the visited network: upon the mobile patient
node entering the threshold or new network area (PAN),
III. MOBILITY SCENARIO then it receives a Beacon message (message1) which is
The WSNs mobility protocols proposed a large scope of broadcasted periodically by 6LoWPAN BR as coordinator
applicability with the conjunction of the variety of case (Monere system). Hence it detects the movement and
scenarios make it difficult to generate a standard mobility. sends Association Request (message 2).
To overcome this challenge, a specific scheme in mobility 4. Confirmation of MN in visited network: the following
management for hospital WSNs has been proposed. The messages as authentication messages are necessary to
requirements of this scheme are continuous monitoring, low validate that the new MN is from the hospital area:
latency, no loose of packets and low signaling. Binding Request (message 3), Binding Challenge
Fig. 3 shows a mobility scenario of a patient that moves (message 4), Challenge Request (message 5), Challenge
between the home network and visited networks and then Reply (message 6), Challenge Forward (message 7),
returns to base/home network. This kind of scenario is Binding Confirm (message 8), Location Update (message
common at hospitals when the patients walk or move to 9), and Association Reply (message 10) message.
other rooms to do medical tests. Phase 1 shows an initial 5. Interchange of data frames in the visited network: The
state of the patient node in home network that exchanges the messages from 11 to 14 show how a data frame and its
vital signs via the Monere system to maintain continuous Ack are exchanged.
monitoring. In phase 2 and 3, it moves to visited network 6. Movement between visited networks: When a patient
and runs mobility scenario, and finally it returns to the home leaves the visited networks, FA informs the HA via Node
network in phase 4. Left and Ack messages (messages 15-16).
Fig. 4 shows a mobility diagram with the messages 7. Returning to the base network: Finally the patient node
exchanged in each step of mobility scenario that discussed comes back to the base network, then it informs HA of its
as follows: new location by sending a Re-association Request
1. Exchange of messages in home network: the general message (messages 17-20).
frames (data, requests, responses and ACK frames)
interchange between sensors and BR such as SPo2 level
per each 5 seconds.

435
 ( b ) FFD devices without the MR

( c ) FFD devices with the MR

Fig. 5. Three connection models between mobile sensors and border router

Fig. 4. Messages exchanging in mobility scenario.

Fig. 5(a) presents the first model in which the MR as
sink node controls, maintains and collects data from body
IV. NEMO-HWSN MOBILITY MECHANISM
sensors and transmits to BR in based network or visited
SCHEME
network, and finally executes the mobility process. Although
As mentioned in section C, the mobile patient node with this model suffers less handoff cost due to use of MR to
its attached sensors is considered as a network or PAN that support mobility, but the MR presents a bottleneck to the
moves between different PANs. This mobility scenario is PAN (should collect all data from sensors), which is a
very similar to NEMO, because the PAN or a group of serious problem in 6LoWPAN. As a result, the MR works as
mobile sensor nodes moves together and also they need a coordinator to handle the mobility scenario and collect data
strong power device as MR to coordinate and collect the as sink node.
PAN data. Hence the partial of mobility cost have close Fig. 5(b) shows the second model in which all body
relation to PAN architecture such as type and number of sensors are FFD devices without any coordinator that attend
sensors, message overhead, and MR as coordinator which to the mobility process. Hence all FFDs repeat and execute
management the mobility in mobile PAN. the mobility scenario such as coordinator node (in previous
model) and send their mobility messages to BR directly. The
handoff process will be increased compared with the first
model and it is clear that the second model costs higher in
terms of energy requirement and data exchange during
mobility [12].
Our mobility management solution is designed to solve
the challenges of previous mobility models. We present the
new scheme with a low overhead and light traffic on MR to
optimize the bandwidth consumption and PAN lifetime.Fig.
5(c) illustrates the proposed architecture which comprises of
FFD as sensor nodes with an MR as the coordinator. In this
model, the MR as coordinator runs the mobility process
based on mobility diagram (Fig. 8) to exchange the
( a ) RFD devices with the MR
handover messages in movement situation; but FFDs data is

436
directly transmitted to BR. Hence the duty of sensing data
transmission is eliminated from MR which leads to longer
lifetime of MR during the tunneling process and sensor
nodes can be located behind the MR without mobility
message support. Finally, the MR registers FFDs in the BR
as a FA in order to create a connection with a new FA and
transmit their data frames into hospital networks. By way of
this technique, we provide the best handoff cost and
mobility scenario for MR. Hence any increase in the
number of FFD will not increase the cost of handoff during
mobility.
Fig. 6, Fig. 7, and Fig. 8 show a comparison of the
mobility diagram in terms of messages exchanged in three
models. The thin and dotted lines show when the MR as the
coordinator runs the mobility scenario which exchanges one
message to follow the mobility process. The bold lines
present the case when all FFDs are attending the mobility
process which exchange the more mobility messages to
complete the mobility scenario. In other words, when the
mobility process repeats for all FFDs, the diagram lines
become bold. As shown in Fig. 8, the signaling cost of our
proposed scheme lies between the two previous models.

Fig. 7. Mobility messages exchanged of Fig. 5(b) with a set of sensors

Fig. 6. Mobility messages exchanged of Fig. 5(a) with one MR

Fig. 8. Mobility messages exchanged of Fig. 5(C) with one optimized MR

437
 V. COMPARISON THREE MODELS [2] S. Dey and N. Shilpa, “Issues in IPv4 to IPv6 Migration,” International
Journal of Computer Applications in Engineering Sciences, vol. I, no. I, pp.
Hence in the NEMO-HWSN scheme, the MR power 9–13, 2011.
consumption and mobility overhead are optimized by way of [3] C. Shelby, Zach and Bormann, 6LoWPAN: the wireless embedded
reduction in the MR traffic and the number of mobility internet. John Wiley & Sons, Ltd, 2009.
messages. As has been pointed out, the mobility cost is [4] A. J. Jara, M. A. Zamora, and A. F. G. Skarmeta, “An Initial Approach
related to handoff process time and tunneling. To calculate to Support Mobility in Hospital Wireless Sensor Networks based on
6LoWPAN ( HWSN6 ),” Journal of Wireless Mobile Networks, Ubiquitous
the handoff process time, we used the process time of each Computing, and Dependable Applications, vol. 1, no. 2/3, pp. 107–122,
message that exchange between MN and HA in during the 2010.
mobility scenario [14]. The handoff times of the NEMO- [5] Jonathan W. Huai and D. Culler, “Extending IP to Personal Area
HWSN against that of model A and B of the previous Networks,” Internet Computing, IEEE, vol. 12, no. 4, pp. 37–45, 2008.
schemes. The handoff process time in the NEMO-HWSN is [6] A. J. Jabir, S. K. Subramaniam, Z. Z. Ahmad, and N. A. W. a Hamid,
smaller in comparison to the second model (Fig. 5(b)). As “A cluster-based proxy mobile IPv6 for IP-WSNs,” EURASIP Journal on
Wireless Communications and Networking, vol. 2012, no. 1, p. 173, 2012.
mentioned before, the proposed scheme (Fig. 5(c))
[7] X. Wang, S. Zhong, and R. Zhou, “A mobility support scheme for
minimizes the signaling overload of the MR; although its 6LoWPAN,” Computer Communications, vol. 35, no. 3, pp. 392–404, Feb.
handoff process time is not as good as the first model but it 2012.
is close enough (Fig. 5(a)), because the number of mobility [8] A. J. Jara, F. J. Blaya, M. A. Zamora, and A. F. G. Skarmeta, “An
messages are increased a little to complete the mobility Ontology and Rule Based Intelligent Information System to Detect and
process. Predict Myocardial Diseases,” in Proceedings of the 9th International
Conference on Information Technology and Applications in Biomedicine,
Finally, This type of scenario is suitable for network 2009, pp. 5–7.
mobility such as NEMO, with 6LoWPAN nodes which [9] J. H. Kim, C. S. Hong, and T. Shon, “A Lightweight NEMO Protocol
suffers from energy challenges such as energy constraint, to Support 6LoWPAN,” ERTI Journal, vol. 30, no. 5, pp. 685–695, 2008.
limited battery or accessing to energy resources. [10] L. M. L. Oliveira, A. F. De Sousa, and J. J. P. C. Rodrigues, “Routing
and mobility approaches in IPv6 over LoWPAN mesh networks,”
VI. Conclusion International Journal of Communication Systems, no. February, pp. 1445–
1466, 2011.
This paper described a mobility solution for mobile
[11] M. A. Zamora-izquierdo and A. F. Gómez-skarmeta, “Integral and
patient node in hospital settings to maintain the continuous Networked Home Automation Solution towards Indoor Ambient
connectivity between the patient nodes and hospital area Intelligence,” IEEE Pervasive Computing, vol. 9, pp. 66–77, 2010.
network. This solution considers the hospital architecture in [12] A. J. Jara, M. a. Zamora, and A. F. G. Skarmeta, “HWSN6: Hospital
order to define a solution that reduces the amount of Wireless Sensor Networks Based on 6LoWPAN Technology: Mobility and
messages exchanged between the mobile patient node and Fault Tolerance Management,” in International Conference on
Computational Science and Engineering, 2009, pp. 879–884.
6LoWPAN hospital network, and the traffic on MR. This
[13] G. Bag, M. T. Raza, H. Mukhtar, A. H. Akbar, S. M. S. Shams, K.-H.
means that the bandwidth consumption is decreased and also Kim, S. Yoo, and D. Kim, “Energy-aware and bandwidth-efficient mobility
the lifetime of the MR is optimized. The patient node should architecture for 6LoWPAN,” in MILCOM 2008 - 2008 IEEE Military
not run a costly configuration for new topology when an MR Communications Conference, 2008, pp. 1–7.
dies early due to congestion. Finally, it is shown that this [14] A. J. Jara, R. M. Silva, J. S. Silva, M. a. Zamora, and A. F. G.
scheme provides the same handoff cost and light traffic on Skarmeta, “Mobile IP-Based Protocol for Wireless Personal Area Networks
in Critical Environments,” Wireless Personal Communications, vol. 61, no.
MR and BR regardless of the number of sensors attached to 4, pp. 711–737, Nov. 2011.
a patient node. Hence, the NEMO-HWSN mobility protocol
for hospital architecture should be more feasible in a WSN
based on 6LoWPAN.
ACKNOWLEDGEMENT
The work described in this paper is funded by the
Ministry of Science, Technology and Innovation (MOSTI)
of The GOVERNMENT OF MALAYSIA (No. 01-01-04-
SF1218) and Universiti Putra Malaysia. The authors would
like to thank all parties which have contributed towards the
success of this project.

REFERENCE
[1] Mohammadreza Sahebi Shahamabadi, P. Varahram, and M. Noura,
“On Power Consumption in IPv6 Over Low Power Wireless Personal Area
Network ( 6LoWPAN ),” in The First International Conference on Green
Computing, Technology and Innovation, 2013, pp. 21–28.

438