Mobility Aspects in WSN - Application and Multidisciplinary Aspects of Wireless Sensor Networks - page 126

b

a

c

b

c

Fig. 6.10 Filing-the-holes with sliding

relocation is performed. The proposed relocation algorithm, of filling-the-hole by
sliding model (Fig. 6.10 ), provides continuous connectivity of the sensor nodes
within each cluster by filling out the coverage holes using the sensor nodes within
the neighbor cells. When a master node dies in a cell, one of the redundant nodes (if
such a node exists) will become a master node. Otherwise, a coverage hole occurs.
In this case, a sensor node from the neighbor cell relocates to the coverage hole. If
there is a redundant node in the neighbor cell, then that node relocates. Otherwise,
the master node in the neighbor cell relocates. Hence, connectivity of all sensor
nodes within the cluster is satisfied continuously. Authors use the term sliding for
filling out the hole in the inner cell by a node from the outer cell. After consecutive
sliding relocations happen, a hole may occur in the outermost tier, and hence reloca-
tion from other clusters may be required to fill the holes in the outermost tier of the
cluster. The algorithm performs locally, and in dense and large networks a long time
can be required for the algorithm to terminate.
The TT layout is considered in [ 30 ], and two approaches for sensor self-deployment
are introduced - Greedy Advance (GA) and Greedy-Rotation-Greedy (GRG). An equi-
lateral TT is a planar graph composed of congruent equilateral triangles. It is an appro-
priate solution because it maximizes the coverage area of any given number of nodes
without a coverage gap when the nodal separation equals 3

R and when the nodes

are placed on the vertices of the layout.
The GA moves the nodes along the TT edges toward the given Point of Interest
(PoI) where the nodes move to the new vertex closer to the PoI (calculated in the graph
distance). Each vertex of TT is represented with three coordinates, and three moving
rules are further presented. The first rule is called Priority Rule and determines the
priority for the nodes' simultaneous movements to the same vertex. Forbiddance Rule
is introduced to avoid simultaneous movement, and the Innermost-Layer Rule allows
the three closest nodes to the PoI to move to the PoI, in which case after some node
reaches the destination, no other node will be allowed to move to it.