ZRP creates a circular zone around each node of radius parameter
rzone. Within each zone, the center node knows the geographical topology of
the network within that radius (i.e., it knows the nodes located inside that
perimeter and the distance to these nodes). This knowledge could be based on
triangulation range measurements, dead reckoning, cooperative localization,
etc. Finding a routing path between each node and a destination node is based
on overlapping circles of radius rzone that eventually lead to the discovery of
the destination node. In actual implementation, in order to determine the
routing path between node I and node N, a broadcast message is sent from
node I within its own rzpne. The message is then repeated until the destination
node N receives the message. For simulation purposes, routes are chosen based
on a depth first search in which the children of each node are the nodes inside
rzpne, and by not allowing closed loops as well as terminated routes with more
than m (m = 8) hops to destination.
Figure 8(a) shows the beginning of a 20-node simulation run, generated
after the nodes have been placed randomly and the routing completed, but
before goals have been selected, with node 20 being a network sink, or data
collection node. In figure 8(b), a goal location of (7,8) is selected and a T =
0.01 Euler-step discretization of the potential field equations is to run for
100 iterations. The nodes within communication range of the goal are being
attracted to the goal location in order to collect samples at that location.
Pages:
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124