Algorithms For Wireless Body Area Networks: A Survey

Full text

ii

iii

iv

v

vi

vii

viii

Table of Contents
1 Introduction 1
1.1 Introduction............................... 2
1.2 Motivation................................ 3
1.3 ProblemStatement........................... 4
1.4 Novelty ................................. 5
1.5 ThesisOutline.............................. 6
2 Algorithms for Wireless Body Area Sensor Networks: A Survey 7
2.1 RelatedWork .............................. 8
2.2 Algorithms for Minimum Energy Utilization in WBASNs . . . . . . 9
2.2.1 Protocol Selection for Minimum Energy Utilization . . . . . 9
2.2.2 Battery-dynamics Driven TDMA MAC Protocol to Restore
EnergyResources........................ 10
2.3 Algorithms for Delay Minimization . . . . . . . . . . . . . . . . . . 11
2.3.1 Minimizing Delay using Genetic Algorithm . . . . . . . . . . 11
2.3.2 Minimizing Delay using Energy Efficient Configuration Man-
agement for Multi-Hop WBASNs . . . . . . . . . . . . . . . 13
2.4 Algorithms for Route selection in WBASNs . . . . . . . . . . . . . 15
2.4.1 Environment-Adaptive Routing . . . . . . . . . . . . . . . . 16
2.4.2 PRPLC routing with LLF Capturing Multi-scale Connec-
tionLocalities.......................... 17
2.4.3 TARA: Thermal Aware Routing Algorithm . . . . . . . . . . 19
2.5 Algorithms for Computational Complexities . . . . . . . . . . . . . 21
2.5.1 Reducing Computational Complexity using Fast Synchro-
nization ............................. 21
2.5.2 Heuristic Path and Observer Based Algorithm for Reducing
Computational Complexity . . . . . . . . . . . . . . . . . . . 22
2.6 Algorithms for Collision Avoidance . . . . . . . . . . . . . . . . . . 24
2.6.1 Collision Avoidance with On Body Packet Routing . . . . . 24
2.6.2 Applying Tree Algorithm for Collision Avoidance . . . . . . 27
ix

2.6.3 Collision Avoidance by Broadcasting Scheme Messages . . . 28
2.7 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 ATTEMPT: Advanced Threshold-based Thermal-aware Energy-
efficient Multi-hop ProTocol for Heterogenous Wireless Body Area
Networks 35
3.1 Background ............................... 36
3.2 SystemModel.............................. 37
3.2.1 Initialization Phase . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.2 Routing Phase . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.3 Scheduling Phase . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.4 Data Transmission Phase . . . . . . . . . . . . . . . . . . . . 45
3.2.5 Mobility Support in ATTEMPT . . . . . . . . . . . . . . . . 46
3.2.6 Invitation Phase . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4 Conclusion 55
4.1 Conclusion................................ 56
References 56
x

List of Figures
2.1 MAC Protocols For WBASN . . . . . . . . . . . . . . . . . . . . . 10
2.2 Routing for WBASN . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Body Movements in WBASNs . . . . . . . . . . . . . . . . . . . . . 14
2.4 Environment Adaptive Routing . . . . . . . . . . . . . . . . . . . . 17
2.5 Thermal Aware Routing . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Heterogenous On body Sensors Network . . . . . . . . . . . . . . . 24
2.7 WBAN Medical Applications . . . . . . . . . . . . . . . . . . . . . 27
2.8 Hierarchy of WBASN Algorithms . . . . . . . . . . . . . . . . . . . 29
2.9 Energy Consumption (LOS) . . . . . . . . . . . . . . . . . . . . . . 32
2.10 Energy Consumption (NLOS) . . . . . . . . . . . . . . . . . . . . . 32
2.11Delay(LOS)............................... 33
2.12Delay(NLOS).............................. 33
2.13 Delay vs Average Power (LOS) . . . . . . . . . . . . . . . . . . . . 34
2.14 Delay vs Average Power (NLOS) . . . . . . . . . . . . . . . . . . . 34
3.1 Health-care Application of WBASNs . . . . . . . . . . . . . . . . . 38
3.2 Sequence of Phases in Each Round . . . . . . . . . . . . . . . . . . 39
3.3 Link Hot-spot Detection . . . . . . . . . . . . . . . . . . . . . . . . 43
3.4 Energy Management for Single-hop and Multi-hop Communication 45
3.5 Flow-chart of ATTEMPT . . . . . . . . . . . . . . . . . . . . . . . 46
3.6 Sequence of Phases in Each Round . . . . . . . . . . . . . . . . . . 47
3.7 Link Establishment and Link Breakage due to Mobility Of Human
body................................... 48
3.8 Packet Delivery Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.9 Number of Nodes Alive Over Time . . . . . . . . . . . . . . . . . . 53
3.10 Number of Dead Nodes Over Time . . . . . . . . . . . . . . . . . . 54
3.11 Total Energy of the Network . . . . . . . . . . . . . . . . . . . . . . 54
xi

List of Tables
2.1 WBASN Algorithms Energy Table . . . . . . . . . . . . . . . . . . 29
3.1 WBASNs Sensors Nodes Data Rates . . . . . . . . . . . . . . . . . 37
3.2 Comparison of ATTEMPT with Existing Algorithms . . . . . . . . 41
3.3 Overview of WBASNs Routing Protocol . . . . . . . . . . . . . . . 47
3.4 Comparison of Different WBASN Wearable Sensors . . . . . . . . . 50
3.5 Simulation Environment . . . . . . . . . . . . . . . . . . . . . . . . 51
xii

Chapter 1
Introduction
1

1.1 Introduction
Wireless communication technologies are playing its important role in different as-
pects of life such as entertainment, games, emergency services, sports and medical
health care. With advancement in wireless technologies and storage devices has
resulted in increase of patient monitoring devices. With these devices, the wire-
less communication technologies are used for data streaming. As a consequence
these technologies consume power for data transmission and reception. To reduce
the power consumption less power signal processing techniques should be intro-
duced. Wireless Body Area Sensor Networks (WBASNs) are used for medical and
nonmedical applications. The wireless sensor nodes used in WBASNs are tiny,
light-weight and limited power sources. These sensor nodes have different levels
of energy and generate different size of data while the Wireless sensor Networks
(WSNs) nodes almost have same level of energy and data rate. Thus, employing
routing algorithm of WSN can not support WBASNs sensor nodes. The selection
of WBASNs routing algorithms should support the heterogeneous sensors network.
To monitor the patient suffering from a chronical disease, different sensor nodes
are deployed on human body, these sensor nodes construct WBASNs. Sensor
nodes are connected to a sink node, sink node has better resources as compared
to all deployed nodes in term of energy and data rate. Sink node on receiving
data from sensors process data and this processed data send to out body server.
Patient is monitored in WBASNs, in case of critical scenarios immediate access
is required to save human life. For data packet delivery routes are constructed
according to cost metric. WBASNs use two types of routing protocols, reactive
routing and proactive routing, the former category of routing protocols maintains
routing table on periodic updates, and later category of routing discover routes
2

when request is arrived. WBASNs also support heterogenous traffic like low to
high traffic rates, and high to lower traffic rates.
1.2 Motivation
In [1], authors increase transmission range of sensor nodes and use single-hop
communication between sensor nodes and sink node, to overcome the problem of
topological partitioning due to constant human body movement and ultra short
Radio Frequency (RF) transmission range.
Sue et al. [2] used multi-hop communication to transfer data between sink and
root nods. In direct communication, increase in temperature of sensor nodes may
affect human tissues. The storage delay (due to topological disconnections) and
congestion delay increase delay in multi-hop communication. Delay is not sup-
portive for emergency services. Thus, Multi-hop communication cannot support
emergency services.
The authors in [3], define a prototype and a detailed topology for WBASN. They
increase the communication range of sensor nodes, and overcome the problem of
partitioning due to mobility of human body postural. They use proactive routing
approach to update sensor nodes by sending Hello messages, and use a higher layer
polling strategy to avoid collision. However, no priority is assigned for emergency
data delivery, and proactive routing is not suitable for WBASN, due to high energy
transmission cost.
We identified the issues of WBASNs for medical and health care services. This
study also discusses how different algorithms use different techniques to minimize
energy consumption, delay, path selection cost, computational complexity. To find
best path, nodes in WBASNs have to spend some time and consume some energy.
3

So, in this work, we calculate cost for time spent and energy consumed. MATLAB
simulations of algorithms are performed for the delay and energy consumption of
disconnected nodes, due to body movements. A detail hierarchy of algorithms is
discussed to better understand the issues.
1.3 Problem Statement
Routing protocol should be fast, scalable during human body movements. On
MAC layer their is a trade-off between reliability, latency and energy consump-
tion. The selected MAC Protocols should be low duty cycle and simplified. From
above all discussion, we see that energy is always a trade off on every layer. Thus,
for long network life algorithms selected for WBASNs should consume less energy.
Because the battery of implanted sensor nodes can not be changed on regular
basis. The implanted sensor nodes can not be changed on regular basis. However,
there is need to optimize energy of the implanted nodes. There is a trade-off be-
tween transmission distance, data rate and power consumption on physical layer.
Decreasing range and suitable data rate selection is important for energy opti-
mization. Modulation and demodulation schemes selection are also important to
save energy. The WBASN algorithms have the following problems:
1)- The implanted sensor nodes should no be changed on regular basis. So, the
selected WBASNs algorithm should be energy efficient.
2)- Single-hop communication between sensor root nodes and sink for normal data
delivery can increase the transmission cost.
3)- Using multi-hop communication for critical data delivery increases delay, since
each intermediate node receive, process and then transmit, this reception and
transmission increases delay in data delivery, which is not supportive for emergency
4

services.
4)- The priority is defined for emergency services.
5)- The selected protocol for WBASNS should be thermal-aware.
1.4 Novelty
We propose and validate a routing protocol both for heterogeneous and homo-
geneous networks. We named this routing protocol Advanced Threshold-based
Thermal-aware Energy-efficient Multi-hop ProTocol (ATTEMPT). The nodes are
placed around the sink in descending order of their data rate. As low data rate
sensor nodes can not forward data generated by higher data rate sensor node due
to insufficient buffer size. In the proposed protocol, priority is assigned to the
emergency services. To overcome delay, critical data is sent directly to the sink
node and normal data is sent through multi-hop communication.
We enhance ATTEMPT to support continuous human mobility. We present a
mechanism for placing heterogenous sensor nodes on human body. We placed
high data rate node on less mobile places on human body for Mobile-ATTEMPT
(M-ATTEMPT). Mobility of human body causes disconnection between previously
established links. It takes time to establish new connection to forward data and
causes delay. As delay is not supportive in real-time applications. To beat delay
and overcome problem of disconnection. We used energy management in our
proposed routing protocol. Using this energy management sensor nodes increase
their transmission range and directly communicate with sink node for critical data
delivery. For normal data delivery multi-hop communication is used. MATLAB
simulations of proposed routing algorithm are performed for lifetime and reliability
in comparison with multi-hop communication. The results show that the proposed
5

routing algorithm has less energy consumption and more reliable as compared to
multi-hop communication.
1.5 Thesis Outline
In next chapter, we present a survey on WBASNs algorithm for energy minimiza-
tion, fast path selection, minimizing delay, reducing the processing complexity
and dealing with collision avoidance. We propose a routing algorithm for het-
erogeneous WBASNs in chapter 4. Finally the conclusion is drawn in chapter
5.
6

Chapter 2
Algorithms for Wireless Body
Area Sensor Networks: A Survey
7

2.1 Related Work
An energy configuration management for multi-hops [2], authors propose Fast
Path Selection Scheme (FPSS) algorithm which make sure low delay and optimize
power consumption. Whereas, priority is assigned to emergency data, and normal
data on basis of power. However, use of Hello messages after a regular interval
result loss of energy.
In [4], authors discuss routing layer there is a trade-off between fast routing and
energy utilization. Routing protocol should be fast, scalable during human body
movements. On MAC layer their is a trade-off between reliability, latency and
energy consumption. The selected MAC Protocols should be low duty cycle and
simplified. From above all discussion, we see that energy is always a trade off on
every layer. Thus, for long network life algorithms selected for WBASNs should
consume less energy. Because the battery of implanted sensor nodes can not be
changed on regular basis.
In [5], authors use fast synchronization algorithm for GMSK and estimate the
Carrier Frequency Offset (CFO) and Phase Frequency Offset (PFO) generated
due to the unhinged oscillator. Coherent detection is used instead of non-coherent
detection due to its power efficiency which is appropriate for WBASN. However,
feed forward structure is implemented in hardware, with no proper discussion on
noise added by Gaussian low pass filter.
In [6], genetic algorithm is applied to find route which is the most steadfast and
having minimum interruption. Where weight is assigned to cost, delay and reli-
ability to find fitness. However, priority is not discussed for sending emergency
data.
8

For full utilization of unused time of the processor in [12], authors use track and
update scheme for path information and to maximize the use of the processors
time. Because different processor are being used in sensor nodes. All sensor nodes
have different execution time for different assigned tasks. However, emergency is
not discussed.
Emergency data has a dedicated channel to sink node. The lower priority data
is lagged until successful delivery of emergency data at sink. To overcome the
problem of collision, authors in [8] use tree base structure for normal and emer-
gency data. A dedicated path is used for emergency services. However, dedicated
channel for emergency is over all waste of resources.
In next section, we discuss algorithms for minimizing utilization of energy in
WBASNs.
2.2 Algorithms for Minimum Energy Utilization
in WBASNs
Different algorithms are used to minimize energy consumption and to achieve
efficient utilization of accessible resources. In this section, we discuss how different
algorithms use different techniques to optimize utilization of energy.
2.2.1 Protocol Selection for Minimum Energy Utilization
The growth and progress in the field of medical, laid some use of wireless sensors
implanted in human body and to monitor the patient’s chronicle disease. These
sensors are light weight. The used technologies are blue tooth and Zigbee. These
communication technologies are short range, inadequate data rate and high energy
consumption. Ultra-Wide-band (UWB) has less energy consumption and adequate
9

data rate which is suitable for communication between the sink node and out body
server.
Selection of communication protocol should be collision free for sending emergency
data. Two schemes are defined to avoid collision; contention based and scheduled
based. MAC protocols use contention based scheme where collision is probable,
when two sensors try to approach the same channel concurrently. Sensors wait for
predefined time and try again until they get an approach to send data. To make
it collision free sensors nodes consume more energy. However, in schedule based
scheme Time Division Multiple Access (TDMA) is used. In this scheme, sensor
nodes are assigned time slots to avoid collision.
Fig. 2.1 Hierarchy shows that the contention base channel approach is more energy
consuming than the scheduled based channel access approach due to overhearing.
MAC
Protocols
Contetion
Based
Schedule
Based
Energy
Consuming
CCA
Contention
Overhearing
Problem
No
Contention
Greater Control
Overhead
High
Synchronization
High Bandwidth
Consumption
Low Bandwidth
Consumption
Figure 2.1: MAC Protocols For WBASN
2.2.2 Battery-dynamics Driven TDMA MAC Protocol to
Restore Energy Resources
TDMA based scheduling techniques are adopted to save energy. However, all
devices must be synchronized. In [12], battery-dynamics driven TDMA MAC
protocol is proposed. Authors proposed the idle time of battery, to recover capacity
effect as a result device remain alive for a long time. In this scheme, sensor nodes
10

hold data un-till sufficient packets are available, and channel is in a good state to
send data.
2.3 Algorithms for Delay Minimization
In literature, multi-point or point-to-point architectures are used to overcome the
problem of delay due to the partitioning of human body movements. Another
type of delay is congestion delay which is always less as compared to end-to-end
packet delay caused by topological disconnections. For real-time applications such
as emergency and health care the presence of end-to-end packet delay is not sup-
portive. It is only supportive in non-real-time applications such as entertainments
and games. Our aim is to discuss the algorithms which are specifically discussing
the critical issue of delay for health care services. In this section, we discuss how
delay is minimized using different algorithms.
2.3.1 Minimizing Delay using Genetic Algorithm
Emergency conditions needs to be reliable and with minimum delay. As delay is
QoS parameters so, there is great require for quality of service in emergency. In
different human posture the different paths are selected to send data from source
to the destination. In [6], Genetic Algorithm is used for WBASN to select the
fast and reliable route. Weight is assigned to cost, reliability and delay for path
selection. This algorithm randomly selected prime population. For computing
fitness, this algorithm performs crossover and mutation on prime and median
population. In this way, most reliable path is selected for the destination with
minimum delay. Path with low delay consumes less energy. The pseudo code is
discussed to understand the working of genetic algorithm for path calculation on
the QoS as shown in Algorithm 1.
11

Algorithm 1 Genetic algorithm
1: Input (Initial Population)
2: for all Prime Population do
3: for all (i=1) do
4: Estimation based selection
5: Select P1 = P/*P denotes initial population, P1 denotes prime popula-
tion*/
6: Crossover
7: for all j=1 do
8: Select two i and j values form P
9: Select another two k and l values form i and j
10: Save k and l in P2 /*P2 denotes median population*/
11: end for
12: Mutation
13: for all j=1 do
14: Select a random point m0=mfrom P2
15: mutate m at specified rate
16: if (m0>)then
17: update m and m0
18: end if
19: update P=P1 + P2
20: Fitness=(Wcost ×Cost(s)) + (Wtime ×time(s) + (Wrel ×reliablity(s))
21: Now data on route which has high fitness value
22: end for
23: end for
24: end for
In Fig. 2.2. two paths are available one is from A to P and other is from A to
C. The reliability and delay from A to P are 0.9 and 6ns, respectively. Whereas,
reliability and delay from A to C are 0.3 and 5ns, respectively. The selected path
is from A to P, due to, better QoS.
Fitness is computed from [6] by using the following equation in which, weight of
cost, Wcost, weight of reliability, Wrel, weight of time, Wtime for calculated path
as:
F itness = (Wcost ×Cost(s)) + (Wtime ×time(s)+
(Wrel ×reliability(s))
(2.1)
12

Path2
Path1
A
B P
S
C
D
A
B P
S
C
D
Selected Path
(a)
(b)
Reliability=0.9
Delay= 6ns
Delay=5ns
Reliability=0.3
Figure 2.2: Routing for WBASN
2.3.2 Minimizing Delay using Energy Efficient Configura-
tion Management for Multi-Hop WBASNs
WBASNs posses limited resources, hence path selection and load balancing are
issues to optimize utilization of available resources. In [2], FPSS is use to minimize
energy and delay. For this purpose author propose energy management configu-
ration for multi-hop in WBASNs. FPSS makes sure optimization in delay and
resources for a fastest path selection. This algorithm defines two type of nodes,
parent nodes and child nodes. The child node selects their parent node for data
dissemination. The selection of parent nodes is based on deviation of remaining
power value in case of normal data packet [N]. For urgent data packet [U] the cri-
13

teria for parent selection is based on remaining power value of the parent nodes.
The child node selects parent node from the available parent node.
In [2], selection is made on QoS and power of nodes Minimizing:qPs
i=1(P(N i)−m)2
s
where P(Ni) is estimated remaining power, mis average remaining power, and s
is total number of nodes with constraints Pi,j Xr,p
i,j ≤δr,p, τ ≤100ms mentioned
in [3]. δr,p value ensures selection of the number of hop count, and τis delay for
Fig. 2.3 (a) depicts the M-hop routing in which node A sending its data through
node H. As the human body changes its position, in Fig. 2.3 (b) the node A
sending its data directly to sink node. The FPSS is design to calculate fast route
against human body movements as shown in Fig. 2.3 (c). The fast path results
minimum delay and low energy consumption which is suitable for WBASNs.
AD
CBx
(a) Phase-I
A
D
CB
(b) Phase-II
y
A
D
CB
(c) Phase-III
zLegend
Sink Node
Root Node
Figure 2.3: Body Movements in WBASNs
If an event occurs on an intermediate node during an establish link. Intermediate
node will receive this message, modify it and send it to specific destination as
shown in Algorithm 2.
14

Algorithm 2 Fast Path Selection Scheme
1: Preconditions: Pi,j Rr,p
i,j
2: Preconditions: C(Nc) = 0
3: Sp←Power of all nodes
4: while Sp>0do
5: ndis ←Disconnected node
6: if ndis = 1 then
7: if nihas a list of PLthen
8: Not connect to parent
9: else
10: Request of connection
11: Nc←List of child nodes
12: if nc< µ&Pi,j Xr,p
i,j < Sr,p then
13: Rp←Register connection to parent node
14: else
15: Rch ←Register child node
16: if Pch /∈Rch then
17: if node having low priority packet is disconnected then
18: Pni ∈P(Ns) select this node
19: else
20: Keep existing Nc
21: Modify routing table
22: Calculate Cni
23: end if
24: end if
25: end if
26: end if
27: end if
28: end while
2.4 Algorithms for Route selection in WBASNs
WBASN requires efficient route selection, some limitations restrict efficient route
selection these are: First, the bandwidth is limited therefore the control overhead
generated by sensor nodes should be less. Secondly, all the sensor nodes are
heterogenous in term of data rate and energy consumption their placing on body is
an issue. The sensor used in WBASNs are heterogenous in different characteristics.
When high data rate transmitted through low data sensor device it decreases life
time of sensor nodes.
15

2.4.1 Environment-Adaptive Routing
The Environment-Adaptive Routing (EAR) algorithm in [13] performs the follow-
ing three steps: Routing Table Constructor, Fault Detector, and Path Selector.
Fully Functional Devices (FFDs) broadcast its control messages, as the Reduced
Functional Devices (RFDs) receive these control messages. These RFDs execute
their constructor module to construct routing tables. RFDs use Path Selector
module to find path for sending their data to FFDs. If a link breakage or con-
gestion occurs on a route, RFDs use Fault Detector to find another best available
path to send their data to FFD, and RFD disseminate a control message to inform
neighbors about the stale route. So that, other nodes not use this route in future
to send their data. The EAR calculate different communication cost for heteroge-
nous WBASN devices. For deploying devices on human body, this algorithm also
defines level for arranging devices on basis of data rate. The EAR routing is very
suitable for WBASNs for health care, because the sensors such as EEG, EMG,
blood pressure have different data rates. Devices Utility function U(v) is use to
calculate the cost at any node in [13] as:
U(v) = α.d2
e+fu+Co(2.2)
Where αis device level, drepresents distance between node uand node v, and eis
residual energy of node u.fuis communication cost of node u, and Cois control
overhead generated by node u. And communication cost fuis receiving cost rc
and transmitting cost tcas:
fu=rc+tc(2.3)
16

In Fig. 2. 4 Environment-Adaptive routing is shown for heterogeneous WBASNs.
In which devices are arranged according to their data rates, the high data rate are
closer to FFDs (Coordinator).
Coordinator
Medical
Device
Medical
Device
Medical
Device
Medical
Device
Medical
Device
Medical
Device
Medical
Device
Medical
Device
CE
Device
CE
Device
CE
Device
CE
Device
Level 1: CE Devices
Data Rate: 10 Kbytes
Level 2: Medical
Devices
Data Rate: 1 Kbytes
Data Rate: 50 bytes
Level 3: Medical Devices
Figure 2.4: Environment Adaptive Routing
2.4.2 PRPLC routing with LLF Capturing Multi-scale Con-
nection Localities
Data transaction among sensor nodes for monitoring patient should have less delay,
due to topological partitioning of frequent human movement. For end-to-end de-
lay and packet loss due to low transmission range Delay Tolerant Network (DTN)
routing is discussed in [3]. They use a prototype and construct network topology
for WBASN, use distance vector and Probabilistic Routing with Postural Link
Cost (PRPLC) approach. In PRPLC, when a node-iwant to forward a packet to
sink node-d. The pseudo code shown in Algorithm. 3 describes multi-hop commu-
nication for high traffic for a common destination. This algorithm calculates its
route on the basis of HCQ and LLF. DTN routing and on body flood mechanism
serve with delay low bounds which is appropriate for real-time applications.
17

Algorithm 3 Probabilistic Routing With Postural Link Cost Algorithm
1: while True do
2: for all (For all nodes, Ni,j where i, jN and i6=j)do
3: ωt
i,j ←Historical Connectivity Quality between node iand node jfor
time-slot t
4: Twindow ←Number of slots
5: Lr
i,j ←links between node iand node jduring a time-slot r
6: ωt
i,j =Pt
r=t−Twindow
Lr
i,j
Twindow
7: Li,j ←link between node iand node jfor time-slot t
8: if (Li,j == 1 ) then
9: Pt
i,j ←Link Likelihood factor
10: Pi,j =Pt−1
i,j + (1 −Pt−1
i,j ).ωt
i,j
11: Check connectivity by Hello-message
12: for all (For all neighbor nodes, ni,j where i, jN and i6=j)do
13: nl←Neighbor List
14: if (Pt
i,d > P t
i,nl)then
15: Send Pt
i,d to node j
16: end if
17: end for
18: else
19: Pi,j =Pt−1
i,j .ωt
i,j
20: Continue buffering at node i
21: end if
22: end for
23: end while
24: if Packet is Received then
25: Modifying message
26: if Destination == 1 then
27: Forward Packet to Destination
28: else
29: Discard it
30: end if
31: end if
And not have any direct link with node-d, the node-ichecks the Historical Link
quality (HCQ) of its neighbor with node-d. And forward packet to neighbor node
has high HCQ with node-d. Link Likelihood Factor (LLf) calculated in [3] as:
Pi,j =Pt−1
i,j + (1 −Pt−1
i,j ).ωt
i,j If link is connected (2.4)
18

Pi,j =Pt−1
i,j .ωt
i,j If link is disconnected (2.5)
2.4.3 TARA: Thermal Aware Routing Algorithm
Biomedical sensor rooted inside human body also has harmful effects on tissues
caused by rising temperature. Thus, this issue must be taken into consideration
while implanting sensor nodes on human body. Thermal-aware routing protocol in
[14], that route the information away from high temperature nodes. This Routing
protocol calculate temperature rise of its next neighbors toward the sink. The
node with high temperature area above threshold are called Hot-spot. Packets
around the Hot-spot area by a withdraw strategy, as depicted in Fig. 2. 5.
(a) Area Hotspot
(Phase-I)
1
2
3
4
5 Sink Node 1
2
3
4
5 Sink Node 1
2
3
4
5 Sink Node
(b) Area Hotspot
(Phase-II)
(c) Area Hotspot
(Phase-III)
Figure 2.5: Thermal Aware Routing
WBASN applications for medical is concerned with body server receiving infor-
mation from sink node. For emergency situation, some priority base data has
been sent, for this function the delay is minimum. To attain minimum delay in
a given bandwidth. Channel is assigned for a priority base data and remaining
process should be lagged until the successful delivery of emergency data. The
computational overhead caused the delay problem. So, in next section we dis-
cuss computational complexity. We step by step describe and enhance the pseudo
code of TARA, to understand effect of rising temperate on tissues shown in Algo-
rithm. 4. TARA defines thermal effect for WBASNs however, the Hot-spot area
19

links are break and packets are withdrawn causes delay not supported in real-time
applications.
Algorithm 4 Thermal Aware Routing Algorithm
1: Initialization
2: while True do
3: for all forward Information packets do
4: Update Routing table (Information of node, Sink Location, Number of
hope count)
5: for all (Node Niwhere iN)do
6: if (Sensed Information ==1 ) then
7: Count Packet
8: CT H ←Count Packet Threshold
9: Tm+1
(i,j)= [1 −δtb
ρCp−4δtK
ρCpδ2]Tm(i, j)
10: +δt
CpSAR +δtb
CpρTb+δ
ρCpPc
11: if (Count Packet < CT H &T emp < T empT H )then
12: Count = Count Packet ++
13: Forward Packet
14: else
15: Mark node ias Hot spot
16: end if
17: else
18: Packet is Received
19: if Buffered Packet <TimeT H then
20: Check Temperature of next node
21: if (Temperature< T empT H )then
22: Forward Packet
23: Modifying message
24: else
25: Wait for Hot Spot area to become normal
26: end if
27: Forward it
28: else
29: Discard it
30: end if
31: end if
32: end for
33: end for
34: end while
20

2.5 Algorithms for Computational Complexities
Computational complexity is a problem in WBASN. In this section we discuss
how different algorithms tackle problem of computational complexity.
2.5.1 Reducing Computational Complexity using Fast Syn-
chronization
Limited resources like power and bandwidth are major major constraints. Transceiver
consumes less than 1mW power for given data rates. Gaussian Minimum Shift
Keying (GMSK) is modulation selection for WBASN. It is very fast, and power
efficient. IEEE 802.15.6 adopts GMSK due to its properties. These properties
are constant envelope and spectral efficiency. In literature different techniques
are used to avoid computational complexity and optimize energy utilization. Fast
synchronization for GMSK [5], adopts feed forward arrangement for efficiency and
burst communication. Timing Offset (TO) and Phase Frequency Offset (PFO)
produce due to unstable oscillator. Improved TO estimation algorithm is used to
minimize computational complexity, to decreases SNR in WBASN.
Use of coherent detection instead of non-coherent detection improves energy op-
timization in GMSK. Where as permeable is introduce in the receiver for packet
detection. Also feed forward timing and frequency recovery algorithm are use
which overall increase the performance, thus decreases the computational com-
plexity. We describe the pseudo code of fast synchronization algorithm and define
how PFO, TO and STO is estimated for reducing complexity of GMSK shown in
Algorithm 5.
21

Algorithm 5 Fast Synchronization Algorithm For GMSK
1: Take input as s(t) = ejψ(t;a)
2: ψ(t;a) = πPkakq(t−kT )←information phase
3: g(t)←Gaussian low pass filter
4: g(t) = 1
2TnQh2πB
√ln 2 t−3T
2i−Qh2πB
√ln 2 t−T
2io
5: nT ≤t≤(n+ 1)T
6: s(t) = eψ(t;a)=P1
i=0 Pn
k=n−2αi,khi(t−k T )
7: if (t [0,3T]) then
8: h0=P(t−T)p(t−2T)
9: else
10: for (t[0, T ]) do
11: h1(t) = P(t−T)p(t+T)
12: end for
13: Symbol Timing Offset estimation
14: x(t)←correlation function
15: x(t) = r(t)⊗(h(−t))
16: Carrier frequency Offset estimation
17: y(iT s)←CFO-compensated signal
18: y(iT s) = r(iT s)s∗(iT s −ˆτ)
19: Carrier phase Offset estimation
20: ˆ
θ= arctan PLθ−1
i=1 =[yc(iT s)]
PLθ−1
i=1 <[yc(iT s)]
21: y(iT s) = e−j[2π(kT +nTs)ˆv]y(iT s)
22: end if
2.5.2 Heuristic Path and Observer Based Algorithm for
Reducing Computational Complexity
WBASN consists of various types of biomedical sensor and actuators. All nodes
are having different processors with different processing capability. The processing
might be complex as some processors takes short time for processing and some
takes long time due to computational complexity. In WBASN there are needed to
reduce this computational complexities therefore, the processors perform better.
Algorithm. 6 defines how the path is computed to minimize the unused time of
processors.
22

Algorithm 6 Heuristic Path and Observer based Low energy Scheduling Algo-
rithm
1: dsink ←deadline of sink node
2: rroot ←Root node with arrival time
3: Wi←worst case execution time of node
4: liCommunication duration
5: Compute SK=dsink−rr oot−(Pi∈pathWi+Pi∈pathli)
Pi∈pathWi
6: P←subset define all paths with reduced complexity
7: while p > 0do
8: Sm←path with minimum scaling factor
9: m=MIN(Sm)
10: Vk←unallocated node
11: for all Vk∈mdo
12: if Vk=task then
13: Rk←start constraint node k
14: Dk←End constraint of knode
15: end if
16: if (1 + Sm).Wk< Dk−Rkthen
17: Sτ K ←1 + Sm
18: Wk←Worst execution time of node k
19: else
20: Sτ K ←Dk−Rk
Wk
21: end if
22: Dk←Rk+Sτ K .Wk
23: for all j∈P&τkdo
24: Update jusing
25: S0
j=Sj+(Sj)−(Sτk−1).Wk
Pi∈pathj
26: Remove jfrom P
27: if Si= 0 then
28: Dk←Rk+lk
29: else
30: Dk←Rk+lk
31: end if
32: end for
33: end for
34: end while
In [7], two types of scheduling algorithms are use. Static scheduling algorithm and
dynamic scheduling algorithm. By using these algorithms, slack (unused time after
task is completed) is distributed among the tasks fairly. As a result computational
complexity is decreased. We enhance the pseudo code of the HPOSS algorithm.
When tasks are assigned to sensors for availability of resources, then there is a
23

need to run static scheduling. As actual and worst case execution times of sensors
are different from each other then there is a need to repeatedly perform static
scheduling to provide low computational complexity. Thus, overall efficiency of
the system is increased. Fig. 2. 6 defines mechanism for placing heterogeneous
nodes on human body, suitable for WBASNs.
Sink Node
Pulse
Oximetry
Temperature
Sensor
Blood
Pressure
Sensor
ECG Sensor
EEG Sensor
EMG Sensor
Low Data
Rate
Low Data
Rate
High Data
Rate
Very Low
Data Rate
Very High
Data Rate
High Data
Rate
Figure 2.6: Heterogenous On body Sensors Network
2.6 Algorithms for Collision Avoidance
In this section we discuss how different algorithms use different techniques to
transfer data without any collision. Collision is avoided by different protocols.
Schedule based or contention based.
2.6.1 Collision Avoidance with On Body Packet Routing
WBASN sensors transmit their data to sink node which further switches data.
however, when sensor nodes are sending data to sink node there is a possibility of
collision. Collision avoidance techniques are implemented in WBASN to optimize
the resources.For sending data to sink node free from any collision and overcome
problem of partitioning [1], the authors specifies a body structure to mounting
24

sensor nodes on human body Structure. For avoiding collision is using higher
layer polling strategy is managed by sink node.
The sink node polls other sensor nodes in a round robin fashion. Data and Hello
messages are sent to transmission range of 0.3mto 0.6m. So, collision is avoided.
Nodes are connected to a sink which, collect the data, after a specific time slot
by using full power of the battery. In this way all node send data directly to the
sink and no need of the intermediate nodes for routing and path selection. The
On Body Store and Flood Routing (OBSFR) algorithm [18], has a better delay
performance by increasing the transmission range, and overcome the problem of
partitioning, using proactive routing approach.
We enhance pseudo code of OBSFR only define transmitting strategy of nodes,
but if a message is received to node how it forwarded to the sink node and the
table is then maintained in Algorithm. 7. OBSFR has high energy transmission
cost, uses single-hop communication and use of hello messages is loss of energy,
not appropriate for WBASNs.
25

Algorithm 7 On Body Store And Forward Routing Algorithm
1: while True do
2: for all forward all packets do
3: for all ni;j∈n;j6=ido
4: Li,j ←links between node iand node j
5: if (Li,j = 1 & iis neighbor ) then
6: PL←Parent List
7: nj∈PLof node i
8: Broadcast packet
9: else
10: Continue buffering at node i
11: end if
12: end for
13: end for
14: end while
15: if Packet is Received then
16: if Destination = 1 then
17: Received Packet
18: else
19: Modifying message
20: end if
21: if neighbor= 1 then
22: Forward it
23: else
24: Discard it
25: end if
26: end if
From the above discussion we propose equations to calculate the time and energy
cost paid for a route by a sensor node to transfer its data to sink node and after
further processing send it to out body server. CEis cost of energy paid by route
and CTis cost of time for a route.
CT=T otal time to col lect Data
Number Of Nodes +F orwardingT ime (2.6)
CT=Tt
Nt
+Ft(2.7)
26

CE=
n
X
i=1
(NE)i+
n
X
i=1
(FE)i(2.8)
Fig. 2. 7 depicts how the data is collected using high level polling strategy after
a regular interval.
Figure 2.7: WBAN Medical Applications
2.6.2 Applying Tree Algorithm for Collision Avoidance
MAC has an important role in the successful operations of WBASN. Especially, in
case of emergency in medical and non-medical. MAC should assign fast access to
the sensor nodes information dissemination to the sink node. In tree algorithm [8],
the authors discuss priority for emergency data. They assign priority varies from
0 to 7. The highest priority has more probability to access the channel to send as
compared to lower priority data. The tree algorithm scenario averts collision and
overcome the problem of wasted slots.
27

2.6.3 Collision Avoidance by Broadcasting Scheme Mes-
sages
Wireless Autonomous Spanning Tree Protocol (WASP) proposed in [15], which
uses a spanning tree for medium access control and data transmission. Parent node
inform its connected child node by sending their data in a time slot by a special
message called WASP-scheme. Every WASP-scheme is unique for every node and
is maintain by the parent node, and another advantages of the WASP-scheme
is that it broadcast recently connected wireless links. These WASP-schemes are
used to avoid the collision and control traffic, and request more resources for its
children. Algorithm 8 defines pseudo code and working of WASP.
Algorithm 8 Wireless Autonomous Spanning Tree Protocol
1: Take address of sending node
2: nc∈Nc←list of child
3: for all ncallocated a slot do
4: if Slot is empty then
5: Send WASP-scheme data
6: end if
7: for all silent period>0do
8: nc←child of nc
9: Send WASP scheme data
10: for all ncrecieves data from ncdo
11: Forward it to sink
12: Cτ←Contention Slot
13: for all Cτ≥0do
14: niscan wireless medium for certain time
15: if (QoS)≥requirement, ni∈PLthen
16: nisend JOINT-REQUEST
17: Register parent node
18: ni∈Rp
19: Register child node
20: Rpupdate WASP-scheme
21: else
22: Piggy-back acknowledgements
23: end if
24: end for
25: end for
26: end for
27: end for
28

Table 2.1: WBASN Algorithms Energy Table
Algorithm IEEE Stan-
dard
Data Rate Energy
Utiliza-
tion
Remarks
Tree 802.15.6 100 Mbps Low Excellent
GMSK 802.15.6 151.8 kbps Low Very good
OBSFR 802.15.6 19.2 kbps High good
FPSS 802.15.4 10 kbps Low good
Hierarchy in Fig. 2. 8, all WBASNs algorithms are discussed with respect to
different issues.
OBSFR Algorithm
Genetic Algorithm Low Energy
Scheduling
Algorithm
FPSS Algorithm Algorithm
For GMSK
WBAN Algorithms
Crossover
Mutation
Problem Specific
Optimization
QoS Path
Selection
Energy
Delay
Reliability
Fast
Overcome
Partioning
Problem Specific
Proactive
Routing
Collision
Avoidance
Fast
Power
Efficient
Problem Specific
Low
Complexity
Low SNR
Fast
Multi-Hop
Problem Specific
Priority
Delay
Energy
Efficient
Problem Specific
Low
Complexity
Low
Energy
Heterogeneous
Sensors
Low
Computational
Time
HPOSS
Algorithm
Static Scheduling
Dynamic Scheduling
Figure 2.8: Hierarchy of WBASN Algorithms
2.7 Simulation Results
We perform simulations to find out the effects of dynamic human body on dis-
connection between sink node and out body server. Because the sink node is
collecting data from all sensor nodes and aggregated data further send to access
points. As sink node is working as central entity. If the sink node has long battery
29

life, network remains alive for long time. Protocols selected for sink node should
be less energy consuming. For this purpose we placed sink on different stance
position and in different ranges which varies from 0 −4mand 4 −10mon Line
of sight (LOS) and NON-Line of Sight (NLOS) to check the connectivity between
sink node and out body server. The interference, absorption, effect of cloth and
effect of environment is also be taken into consideration. We simulate different
parameters variety distance and packet length placing nodes on LOS and NLOS,
to measure energy losses and delay. These path loses are due to mobile human
body. Fig. 2. 9 depicts that as the number of channels increase the channel
energy about the energy level increases on LOS. As channels increase it occupy
more bandwidth due to high data rate. As WBASNs are heterogeneous networks
in term of data rate. Fig. 2. 10 portrays with increase in number of channels
energy consumption about the mean energy level also increases on NLOS. This
is due to human body continually changing positions causes the link break and
utilizes more energy to establish new connection to transmit data.
Excess delay is defined as all possible values of delay in the instantaneous impulse
response after subtracting off the delay of the fastest arriving signal component.
The Root-Mean-Square (RMS) delay spread is probably the most important single
measure for the delay time extent of a multi-path radio channel. Fig. 2. 11 depicts
at LOS the delay increase with channel number is due to disconnections of establish
routes, the routes break due to the signal absorption, short range radio links. The
new connection takes time to establish. In case of NLOS as the channel number
increases delay also increased. Because at NLOS repeated disconnection increase
delay, the connections are breaking due to body movements and absorption shown
in Fig. 2. 12.
30

The communication between node and sink is either single-hop or M-hop. When
nodes are placed in LOS the communication is single-hop. But the communication
is M-hop when communicating nodes one is on chest and other is on back of person.
The single-hop communication is inefficient one sender and receiver are far away.
Fig. 2. 13 portrays average power decay profile and delay for LOS. The delay and
power decay increase when communication between nodes on chest and back of
person which is NLOS as shown in Fig. 2. 14. The routing should be fast against
human continuous movement to overcome the problem of partitioning. In next
sections we discuss the algorithms for WBASN routing.
31

0 10 20 30 40 50
−10
−8
−6
−4
−2
0
2
4
6
Channel number
Channel Power (dB)
Channel energy
Energy mean
Energy mean + Energy stddev)
Energy mean − Energy stddev
Figure 2.9: Energy Consumption (LOS)
0 10 20 30 40 50
−8
−6
−4
−2
0
2
4
6
Channel number
Channel Power (dB)
Channel energy
Energy mean
Energy mean + Energy stddev)
Energy mean − Energy stddev
Figure 2.10: Energy Consumption (NLOS)
32

0 10 20 30 40 50
2
4
6
8
10
12
14
16
18
20
Channel number
Delay (ns)
Excess delay
mean excess delay
RMS delay
mean RMS delay
Figure 2.11: Delay (LOS)
0 10 20 30 40 50
2
3
4
5
6
7
8
9
10
11
Channel number
Delay (ns)
Excess delay
mean excess delay
RMS delay
mean RMS delay
Figure 2.12: Delay (NLOS)
33

0 20 40 60 80 100
−60
−50
−40
−30
−20
−10
0Average Power Decay Profile
Delay (nsec)
Average power (dB)
Figure 2.13: Delay vs Average Power (LOS)
0 20 40 60 80 100 120
−60
−50
−40
−30
−20
−10
0Average Power Decay Profile
Delay (nsec)
Average power (dB)
Figure 2.14: Delay vs Average Power (NLOS)
34

Chapter 3
ATTEMPT: Advanced
Threshold-based Thermal-aware
Energy-efficient Multi-hop
ProTocol for Heterogenous
Wireless Body Area Networks
35

3.1 Background
The rising temperature of implanted sensor nodes due to communication radiations
and circuitry power consumption can affect the human body. In [14], authors use
thermal-aware routing to minimize the effect of rising temperature of implanted
sensor nodes. Quwaider et al. [3] use single-hop communication to transfer data
between root nodes and sink node. They also define a prototype for placing sensor
nodes on human body.
Environment Adaptive Routing (EAR) algorithm [13] define different communi-
cation cost for heterogeneous WBASNs devices. However, the single-hop com-
munication and proactive routing is not suitable for WBASNs. Multi-hop com-
munication is suitable for normal packet delivery and single-hop is only used for
emergency services due to high transmission cost. Use of Hello messages after a
regular interval results high energy consumption.
In [15], Wireless Autonomous Spanning Tree Protocol (WASP)is defined to achieve
low delay and network reliability for WBASNs. WASP-scheme message is dis-
seminate to update parent nodes with information of child nodes. However, in
WASP-scheme power balancing issue is not tackled.
Annur et al. in [8] apply tree algorithm with prioritization for WBASNs. A
channel is dedicated for emergency data delivery and normal data is lagged until
the successful delivery of critical data. However, the dedicated channel results loss
of available resources.
36

3.2 System Model
In order to introduce our model, we suppose that the sink is placed in the center
of the human body. Since WBASNs are heterogeneous networks, and placement
of nodes on human body is an issue. We resolved this issue by placing nodes
in descending order of their data rate with respect to sink, as depicted in Fig.
3. 1. Thus, the nodes with high data rate send data directly to the sink node,
and can easily forward the received data from low data rate sensors. Implanted
sensor nodes on human body with their data rate are mentioned Table. 3. 1.
Problems analyzed in previous section are set in following manner: 1) single-hop
communication is used for emergency services and on-demand data, 2) for normal
data delivery multi-hop communication is used, 3) to prolong life-time of network
by selecting the path with less hop-counts.
Table 3.1: WBASNs Sensors Nodes Data Rates
Sensors Data Rate
EMG Sensor Very High
Image/ Video
Sensor
Very High
Accelerometer High
Blood Glucose High
ECG Sensor High
EMG Sensor High
Blood Pres-
sure
Low
Tempraure
Sensor
Very Low
CO2Gas Sen-
sor
Very Low
Fig. 3. 2 depicts the phases of proposed routing protocol with above mentioned
features. There are four phases in our in proposed routing protocol. These are ini-
tialization phase, routing phase, scheduling phase and steady state or data trans-
mission phase. Initialization phase of the proposed routing algorithm is discussed
37

Figure 3.1: Health-care Application of WBASNs
in next subsection.
3.2.1 Initialization Phase
In initialization phase, all nodes broadcast Hello messages. These Hello message
contains neighbors information and distance of sink nodes in form of hope-counts.
In this way, all nodes are updated with their neighbors, sink node position and
available routes to the sink node. Route computation for data delivery to sink
node of the proposed routing algorithm is discussed in next subsection.
38

Initialization Phase Routing Phase Scheduling Phase Data Transmission Phase
Sink Information Neighbors Information Multi -hop Routes Single -hop Routes On- demand TDMA Slots Allocation
Figure 3.2: Sequence of Phases in Each Round
3.2.2 Routing Phase
In this phase, routes with fewer hopes to sink node are selected from available
routes. We suppose nodes have information of all nodes and position of the sink
nodes. So, selected routes are steadfast and consume less energy. Emergency
services are also defined in proposed routing protocol. In critical scenarios, all
processes are lagged untill successful reception of critical data at the sink node. In
case of emergency, all the implanted nodes on the body can communicate directly
to the base station. Moreover, all sensor nodes can communicate directly to the
sink node when demand is arrived from sink node. In direct communication,
delay is much less as compared to multi-hop communication. Because in multi-
hop communication, each intermediate node receives, processes and then sends
data to next node. The reception, processing and then transmitting the received
data on each intermediate node which causes delay. And some time congestion
also increased this delay. In critical scenario, delay is not acceptable. This delay
is minimized by sending data through single-hop communication. We calculate
39

energy consumed in single-hop communication ESHOP as:
ES−HOP =Etransmit (3.1)
And transmission energy Etransmit is calculated as:
Etransmit =Eelec +Eamp (3.2)
where, Eelec is the energy consumed for processing data and Eamp is energy con-
sumed by transmit amplifier. We suppose a linear network in which all nodes are
implanted at equal distance from each other. To transmit bbits up to nhops, the
transmission energy is given as:
Etransmit =n(bEelec +bEamp)d2(3.3)
here d2is the energy loss due to the transmission.
Etransmit =nb(Eelec +Eamp)d2(3.4)
Comparison of the attributes of proposed routing algorithm with existing routing
algorithms are given in Table 3. 2. Energy preservation is a prime consideration
in WBASNs, as the deployed sensor nodes have limited energy sources. So, de-
ployed nodes need reasonable use of battery for extended life-time of the network.
Implanted sensor nodes on human body have some heating effects.
In our single-hop/multi-hop traffic control algorithm 1, if a node sensed emergency
or on-demand data, then it uses single-hop communication. In single-hop commu-
nication, the sensor node uses full power of battery to sends its data. On the other
hand, for normal data is received, the multi-hop communication is used to send
data to sink node. Thus, collectively less power is consumed without effecting
40

Table 3.2: Comparison of ATTEMPT with Existing Algorithms
Algorithms Network Type Communication Thermal-
aware
Energy-
efficient
Emergency Mobility
Support
FPSS[2] Homogeneous Multi-hop Yes Yes Yes No
TARA[14] Homogeneous Multi-hop Yes No No No
OBSFR[1] Homogeneous Single-hop No No No No
EAR[13] Heterogeneous Multi-hop No Yes No No
WASP[15] Homogeneous Multi-hop No No No No
Tree [8] Homogeneous Multi-hop No No Yes No
DMQOS[16] Homogeneous Multi-hop No No No No
ATTEMPT Heterogeneous Single-hop/Multi-
hop
Yes Yes Yes No
M-ATTEMPT Heterogeneous Single-hop/Multi-
hop
Yes Yes Yes Yes
41

reliability in term of delay. As the distance increases more energy is consumed to
send data. Thus, in multi-hop communication, energy consumption is very less.
Algorithm 9 Single-hop/Multi-hop traffic control algorithm
1: for i=1:1:n do
2: Initialization Phase
3: Tr←Transmission range of node i
4: if (Nodei<= Tr)then
5: Direct communication with Sink Node
6: else
7: if (Critical Data ==1 |On-demand==1) then
8: Send data to Sink Node
9: if (Li,j >CT)then
10: Send data to other route
11: end if
12: end if
13: end if
14: end for
When we are dealing with wireless communication around the human body, effects
of these sensor on human body can also be taken in to consideration. The most
important factor consider for this purpose is Specific Absorption Rate (SAR) and
heating effects of the implanted sensor nodes on human body. The purpose routing
protocol is design to work according to SAR and heating effects on human body.
As nodes implanted closer to sink node are forwarding data of their follower nodes.
Whenever these nodes reach their temperature threshold, these nodes break their
link to their neighbor nodes for few rounds. As their temperature become normal
these sensor nodes establish their previous routes. However, if a sensor node
receives a data packet and reaches its temperature threshold then it returns packet
to previous node. And previous node mark this link as Hot-spot as shown in Fig.
3. 3. When we are dealing with normal data the Delay Tolerant Network (DTN)
is supportive [1].
To calculate the energy consume during a multi-hop communication we assume a
42

Figure 3.3: Link Hot-spot Detection
linear network in which all nodes are deployed at equal distance from each other.
The loss of energy during multi-hop communication can be computed using the
following equations:
EM−HOP =Etransmit +Ereceived (3.5)
where, Ereceived is the energy loss for receiving data. If we are transmitting b-bits
to a distance of n-hops then the transmission energy will be n∗b∗Etransmit and
receiving energy will be (n−1)b∗Ereceived. Since the first node transmit only
and intermediate nodes first receive n-bits and then transmit these received bits.
43

Therefore, the energy consumed for multi-hop is:
EM−HOP =nbEtransmit + (n−1)bEreceived (3.6)
from equation (2), equation (6) becomes:
EM−HOP =nb(Eelec +Eampd2)
+(n−1)bEelec
(3.7)
=nbEelec +nbEampd2
+nbEelec −bEelec
(3.8)
The ATTEMPT routing is discussed in Algorithm 2. If two routes are available
the less hop-count route is selected. If two route have same hop-count, than route
selected which have less energy consumption to the sink node.
Algorithm 10 : ATTEMPT Routing
1: Routing Phase
2: if ( route 1 <route 2 ) then
3: route 1 = selected route
4: else
5: route 2 = selected route
6: if ( route 2 <route 1 ) then
7: route 2 = selected route
8: else
9: route 1 = selected route
10: if ( route 1 = route 2 ) then
11: Ehop−count ←Energy consumption for a route
12: if (Ehop−count 1< Ehop−count 2 ) then
13: route 1 = selected route
14: else
15: route 2 = selected route
16: end if
17: end if
18: end if
19: end if
44

= [2nbEelec +nbEampd2−bEelec ] (3.9)
Single hop and multi-hop communication of root node with sink is shown in Fig.
3. 4.
(a) Multi-hop communication for normal
data delivery
10 Kbyte
1 Kbyte
50 bytes
Data rates
(b) Single-hop communication for critical and
on-demand data delivery
10 Kbyte
1 Kbyte
50 bytes
Data rates Sink node
Parent node
First level child-node
Second level child-node
Figure 3.4: Energy Management for Single-hop and Multi-hop Communication
3.2.3 Scheduling Phase
After route selection phase, the sink node starts channel assignment using Time
Division Multiple Access (TDMA) schedule for communication with root nodes.
Sink node allocates time-slots to the root nodes for normal data delivery.
3.2.4 Data Transmission Phase
After time slots assignment to the root nodes, root nodes send their data to sink
node in their allocated time slot. After that the sink node received data it will take
some time to aggregate the received data. Sink node then send this aggregated
45

data to the out body server through wireless link. Flowchart of the ATTEMPT
algorithm is depicted in Fig. 3. 5.
Intialization
If (critical data=1| on-
demand=1)
Routing Phase
If(sensor temperature
< threshold temperature)
Scheduling Phase
Data Transmission
If (round end=1)
Select second
available path
No
Yes
Yes
No
No
Yes
Figure 3.5: Flow-chart of ATTEMPT
3.2.5 Mobility Support in ATTEMPT
WASNs are mobile in nature because of movements in human body. Our proposed
scheme supports mobility. To achieve this we propose a prototype for placing nodes
on human body and named it as Mobile-ATTEMPT (M-ATTEMPT). Nodes with
high data rate are placed at less mobile places on human body. Theses nodes are
parent nodes and directly connected to sink node. Parent nodes with 10Jgenerate
10Kbytes of data. The nodes directly connected to parent nodes are first level
46

child-nodes with 5Jgenerate 1Kbytes of data. The nodes which are connected to
first level nodes are second level child-nodes with 1Jgenerates 50bytes of data.
Parent nodes, first level child-nodes and second level child nodes placed on hu-
man body and their respective topology. Overview of existing WBASNs routing
protocols are discussed in [4] and are given in Table 3. 3. In next section, we
discuss complete working of M-ATTEMPT. Sequence of phases during a round of
M-ATTEMPT is depicted in Fig. 3. 6.
Setup Phase
Invitation
Phase Routing Phase Scheduling Phase Data Transmission
Time
Time
Joint-
Request
Advertisement Wait for
Registration
Figure 3.6: Sequence of Phases in Each Round
Table 3.3: Overview of WBASNs Routing Protocol
Cross
Layer
Temperature
aware
Cluster
based
WASP TARA HIT
CICADA (A)LTR x
Ruzzelli LTRT x
3.2.6 Invitation Phase
In this phase, we discuss how our proposed routing protocol support mobility. If
a node changes its position during a round, nodes have to pay lot of energy to
maintain link establishment. Sensor nodes only maintain its connection during
47

mobility if sensor nodes sending an emergency data. As, human body changes its
position, first level child node C4 disconnected from parent node P2 and entered
in communication range of parent node P1, as depicted in Fig. 3. 7 phase-I.
Now C4 will send joint-request to parent node P1, as shown Fig. 3. 7 phase-II.
Parent node will check its parent child list, if number of child nodes are less then
µmax = 3. Then parent node P1 will accept joint-request and register C4 it as a
child node as depicted in Fig. 3. 7 phase-III.
Sink Node
P1 P2 P3 P4
C1 C2 C3 C5 C6C4
C11 C31
Sink
Node
P1 P2
P3P4
C1
C11 C31
C3
C2 C4
C5
C6
Sink Node
P1 P2 P3 P4
C1 C2 C3 C5 C6C4
C11 C31
Sink
Node
P1 P2
P3P4
C1
C11 C31
C3
C2
C4
C5
C6
Sink Node
P1 P2 P 3 P4
C1 C2 C3 C5 C6C4
C11 C31
Sink
Node
P1 P2
P3
P4
C1
C11 C31
C3
C2
C4
C5
C6
(a) Phase-I (b) Phase-II (c) Phase-III
Figure 3.7: Link Establishment and Link Breakage due to Mobility Of Human body
48

The attenuation between root node and sink node which is proportional to d2.
The expression for this attenuation, A, as calculated in [17]:
Aj=
nj
X
i=1
d2
i(3.10)
Location of parent nodes can be computed from following equation:
Xj=1
nj
nj
X
i=1
Xj(3.11)
Yj=1
nj
nj
X
i=1
Yj(3.12)
Cost of energy paid by a root node during a round due to mobility of human body
position is computed as:
C=vi∗q(Xi−Xj)2+ (Yi−Yj)2(3.13)
vi


viif vi< vt
viotherwise
(3.14)
where, viis velocity of mobile node, Xjand Yjare defined in equation (19) and
(20) respectively. Some WBASNs sensor nodes with their data rate, bandwidth,
power discussed in [4], is given in Table 3. 4.
3.3 Simulation Results
We perform simulations to compare the performance of our proposed routing pro-
tocol with multi-hop communication in MATLAB. We take network size of 5m
x 5min which 10 nodes are randomly distributed and sink node is placed in the
center of the network. We take 5000rounds for these simulations. For mobility
support we change positions of first level child-nodes and second level child nodes
49

Table 3.4: Comparison of Different WBASN Wearable Sensors
Sensor nodes Data
type
Power con-
sumption
QoS Privacy Accuracy Band-
width
ECG 288 kbps Low Yes High 12 bits 100-1000
Hz
EMG 300 kbps Low Yes High 16 bits 0-10,000
Hz
EEG 43.2 kbps Low Yes High 12 bits 0-1 Hz
Blood Pres-
sure
16 bps High Yes High 8 bits 0-150 Hz
Temperature
sensor
120 bps Low Yes Extremely
Low
8 bits 0-1 Hz
50

after 5rounds. All parameters taken for these simulations are given in Table 3. 5.
Table 3.5: Simulation Environment
Parameters Value
Size of Net-
work
5 m x 5m
Number of
Nodes
10
Deployment Random
Sink Location (2.5, 2.5)
Initial Energy 0.5 J
Number of
Rounds
5000
Application
type
Periodic-base/
Threshold-base
Packet Size <= 64 Byte
Traffic Type CBR
Radio Range <= 10m
There is a complex trade-off between energy efficiency and fast routing in this
mobile network. A multi-hop routing is the best choice for WBASNs and that
one has to deal with a trade-off between energy efficiency and reliability. Reli-
ability is experimentally investigated by measuring the packet delivery ratio. A
multi-hop strategy turns out to be the most reliable. Fig. 3.8 depicts that the
reliability of ATTEMPT is almost 400% better in stable and unstable region than
M-ATTEMPT and multi-hop communication. The throughput of ATTEMPT is
greater because it is sending threshold data and periodic data. While in case of
M-ATTEMPT this packet drop also increased due to mobility of human body and
less number of packets are receiving at sink node. However, as the mobility in-
creased, packet drop rate increased. As a result of mobility fewer packets reached
at sink node. As the distance and mobility of human body is increased between
sink and deployed nodes the packet drop rate is also increased.
Stability period of a network is defined as when all nodes in a network are alive.
Network lifetime of M-ATTEMPT is 48% long as compared to multi-hop com-
51

munication and almost 35% greater lifetime, as compared to ATTEMPT. M-
ATTEMPT has 32% better stability period, as compared to Multi-hop routing
and 20% greater stability, as compared to ATTEMPT, as depicted in Fig. 3.9.
Fig. 3.10 shows instable period of network. M-ATTEMPT has 28% less insta-
ble period, as compared to multi hop communication. The energy consumption
of M-ATTEMPT is less and has better network lifetime, as compared to multi-
hop communication and ATTEMPT. Total energy consumption of M-ATTEMPT,
ATTEMPT and multi hop communication is depicted in Fig. 3.11.
52

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
5000
10000
15000
Number of Rounds
Reliability
ATTEMPT
M−ATTEMPT
Multi−hop
Figure 3.8: Packet Delivery Ratio
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
1
2
3
4
5
6
7
8
9
10
Number of Rounds
Number of Alive Nodes
ATTEMPT
M−ATTEMPT
Multi−hop
Figure 3.9: Number of Nodes Alive Over Time
53

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
1
2
3
4
5
6
7
8
9
10
Number of Rounds
Number of Dead Nodes
ATTEMPT
M−ATTEMPT
Multi−hop
Figure 3.10: Number of Dead Nodes Over Time
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Number of Rounds
Total Energy of Network
ATTEMPT
M−ATTEMPT
Multi−hop
Figure 3.11: Total Energy of the Network
54

Chapter 4
Conclusion
55

4.1 Conclusion
In this thesis, we review the current research on WBASNs. Adaptation of different
techniques and several proposed algorithms to handle problems for WBASNs.
WBASN is a useful technology with potential to over a large range of facilities
to patients, medical personnel and humanity through continuous monitoring and
early finding of problems. For the routing the possible network topologies for
WBASNs are discussed. Taking into consideration the required energy efficiency
and reliability. MATLAB simulations of proposed algorithms are analyzed for the
delay and energy consumption of disconnected nodes, and check the connectivity
between nodes and out body server for LOS and NLOS due to human movement.
We see some connections break due to absorption of cloth and human body, and
also take time to setup a connection, which causes delay and energy utilization.
In this thesis, we present an energy efficient routing algorithm for heterogeneous
WBASNs. For real-time and on-demand data traffic root node directly communi-
cate with sink node and multi-hop communication is used for normal data delivery.
The proposed routing algorithm is thermal-aware which senses the link Hot-spot
and routes the data away from these links. After selection of routes sink node
creates TDMA schedule for communication between sink node and root nodes for
normal data delivery. MATLAB simulations of proposed routing algorithm are
performed for lifetime and reliability in comparison with multi-hop communica-
tion. Topology and placement of nodes is described with single-hop and multi-hop
communication scenario. The results show that the proposed routing algorithm
has less energy consumption and more reliable as compared to multi-hop commu-
nication.
56

References
[1] M. Quwaider and S. Biswas, “On-body packet routing algorithms for body
sensor networks, in Networks and Communications, 2009. NETCOM 09. First
International Conference on, dec. 2009, pp. 171 177.
[2] S.-H. Seo, S. Gopalan, S.-M. Chun, K.-J. Seok, J.-W. Nah, and J.-T. Park, An
energy-efficient configuration management for multi-hop wireless body area
networks, in Broadband Network and Multimedia Technology (IC-BNMT),
2010 3rd IEEE International Conference on, oct. 2010, pp. 1235 1239.
[3] M. Quwaider and S. Biswas, “DTN routing in body sensor networks with
dynamic postural partitioning, Ad Hoc Networks, vol. 8, no. 8, pp. 824841,
Nov. 2010.
[4] Sana Ullah et al., “A comprehensive survey of Wireless Body Area Networks.”
Journal of Medical Systems (2010), Publisher: Springer Netherlands, Pages:
1-30, 2010.
[5] Yan Li Yuen et al., “Fast Synchronization Algorithms for GMSK at Low SNR
in BAN”. IEEE 13th international Conference on e-health Networking, Appli-
cation and services, 2011.
[6] Hassanpour, S.; Asadi, B.; Vejdanparast, Y.; Zargar, P.; , “Improving relia-
bility of routing in Wireless Body Area Sensor Networks using Genetic Algo-
57

rithm,” Computer Science and Automation Engineering (CSAE), 2011 IEEE
International Conference on , vol.2, no., pp.590-593, 10-12, June, 2011.
[7] Yanhong Liu; Veeravalli, B.; , “Heuristic-Path and Observer based Low-Energy
Scheduling Algorithms for Body Area Network Systems,” Biomedical Circuits
and Systems Conference, 2007. BIOCAS 2007. IEEE , vol., no., pp.167-170,
27-30, Nov, 2007.
[8] Annur, R.; Wattanamongkhol, N.; Nakpeerayuth, S.; Wuttisittikulkij, L.;
Takada, J.; , “Applying the Tree Algorithm with Prioritization for Body Area
Networks,” Autonomous Decentralized Systems (ISADS), 2011 10th Interna-
tional Symposium on , vol., no., pp.519-524, 23-27, March, 2011. 2011.
[9] Jung-Yeol Oh; Jae-Hwan Kim; Hyung-Soo Lee; Jae-Young Kim; , “PSSK Mod-
ulation Scheme for High-Data Rate Implantable Medical Devices,” Information
Technology in Biomedicine, IEEE Transactions on , vol.14, no.3, pp.634-640,
May 2010
[10] Zhang, A.; Hanlen, L.W.; Xiaojing Huang; , “Spreading with Walsh code for
p/2-DBPSK modulated WBAN systems,” Personal, Indoor and Mobile Radio
Communications Workshops (PIMRC Workshops), 2010 IEEE 21st Interna-
tional Symposium on , vol., no., pp.84-88, 26-30, Sept, 2010.
[11] Yan Li; Sumei Sun; Yuen Sam Kwok; , “Energy Efficient Low-Complexity
Symbol-by-Symbol GMSK Demodulator for BAN,” Vehicular Technology Con-
ference (VTC Spring), 2011 IEEE 73rd , vol., no., pp.1-5, 15-18, May, 2011.
[12] Hang Su, Xi Zhang, “Battery-dynamics driven tama Mac protocols for wire-
less body-area monitoring networks in health care applications.” Selected Areas
in Communications, IEEE Journal on, Volume: 27 Issue:4, Page no: 424 - 434.
58

May, 2009.
[13] D. Y. Kim, W. Y. Kim, J. S. Cho, and B. Lee, “Ear: An environment-adaptive
routing algorithm for WBANs, in Proceedings of International Symposium on
Medical Information and Communication Technology (ISMICT), March 2010.
[14] Q. Tang and N. Tummala, “Thermal-aware routing algorithm for im-planted
sensor networks, in Proceedings of the International Conference on Distributed
Computing in Sensor Systems, 2005, pp. 206217.
[15] B. Bart et al., “The Wireless Autonomous Spanning tree Protocol for
Multi-hop Wireless Body Area Networks.” Mobile and Ubiquitous Systems-
Workshops. 3rd Annual International Conference on , vol., no., pp.1-8, 17-21.
July, 2006.
59