Aerospace and Electronic Systems - March 2019 - 61

Khriji et al.
any additional hardware since nodes can measure their
RSSI values based on the received signal.

Algorithm 1. Fuzzy logic for reference node selection algorithm
Output: Reference node chance;
Input: RSSI, Node density, Battery level;
1) Select available nodes in the communication range
2) Generate Membership Function for Input RSSI
3) Generate Membership Function for Input Node
density
4)
Generate Membership Function for Input Battery level
5)
Generate Membership Function for Output Reference node chance
6)
Generate Fuzzy Rule (RSSI, Node density, Battery
level, Reference chance)
7) Apply Fuzzy Inference Engine with Aggregation
(RSSI, Node density, Battery level, Reference
chance)
8) Calculate Output Chances (w) from CoA
This approach consists of two main steps. First, reference nodes broadcast a message to all their neighboring
nodes. Once the message is received, the unknown node
measures the RSSI to the sending node and calculates
the distance based on the path loss model [26]. If a node
receives an RSSI with a higher value, the reference node is
close and must have a higher weight. In some cases,
different reference nodes could have the same weight. For
this, additional parameters are considered like the node density and battery level to ensure the closeness of the node as
well as the lifetime of the node itself. By selecting the node
with higher battery level, we maintain the energy balance
between nodes and reduce the workload.
During the localization, each node tries to look for all
its neighbors to select the most suitable reference nodes to
conduct the distance calculation. This choice is based
on the chances of a reference node to be in the communication range of the node in question. To predict such
value, the fuzzy system is introduced. The reference selection step is processed in two phases:
Initialization phase: In this phase, neighboring
information can be updated. Then, each node calculates independently its chance to be the nearest reference. A beacon message will be exchanged
between neighbors including the chance of being
the nearest reference among available ones.
Reference selection phase: Every node waits until
the expiration of its own time delay. If a node does
not receive the beacon message from any other sensor node during its time delay, it is assumed to not
be covered. Each sensor node announces its current
status to all nodes within the reference range.
MARCH 2019

The node density is defined as the number of surrounding nodes in the coverage area. Having more than three
nodes in the communication area ensure a reduced calculation error since nodes will be closer to each other. Node density estimates the number of surrounding references in the
communication range of the node in question, where RSSI
values define the quality of the received signal as presented
in the following list. Based on the system requirement, rules
between input variables and output results are defined
RSSI > À55dBm ! High quality > 75%
À75dBm < RSSI < À55dBm ! 45% < Medium quality < 75%
À89dBm < RSSI < À75dBm ! 25% < Low quality < 45%
RSSI < À96dBm ! 8% < Unusable < 25%:

Once the localization algorithm is applied, nodes in the
network will be aware of their geographical positions. The
lookup table will be then updated with estimated positions. Therefore, by applying the localization algorithm,
the online phase of our proposed routing algorithm starts
with nodes positions as an input.

ONLINE PHASE
After finishing the offline phase, a cluster formation phase
will take place. It includes two subphases, which are the
clustering and data transmission.

FUZZY-BASED ENERGY AWARE UNEQUAL CLUSTERING
ALGORITHM
All nodes within same ring have the same cluster radius,
which is well defined in [27] based on
RðCHk Þopt

sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
2k À 1
¼d
k ¼ 1; . . . ; L
mk

(4)

with mk the optimal number of cluster in the ring k, d the
width of a ring, and L the total number of rings in the
network.
Each node determines autonomously its ring based on
its position. To discover its neighbors, each node uses a
TDMA protocol to broadcast a NODE INFO message
including its ID, its ring ID, its position, and its residual
energy with a communication range equal to the optimal
radius determined in (4). When a node receives NODEINFO from its neighbors, it saves the information in its
routing table, add this node in its neighbor list, and calculates its distance from its neighboring nodes. Thus, the
routing table of each node includes node ID, node ring ID,
node positions, node distance to each neighbors, list of
neighbors, and node residual energy.
After discovering its neighbors, each node has to calculate independently its chance to be a CH based on three
parameters, which are residual energy, density, and centrality. These parameters are chosen because of their

IEEE A&E SYSTEMS MAGAZINE

Aerospace and Electronic Systems - March 2019

Table of Contents for the Digital Edition of Aerospace and Electronic Systems - March 2019