Modeling for Geo-Routing in Multi-Hop Wireless Networks - Application and Multidisciplinary Aspects of Wireless Sensor Networks - page 150

and comply with the specification of IRIS motes [ 16 ], as detailed earlier. Another
assumption is that the wireless network is in an isolated remote environment with
either slow moving or no mobility events. When a sending node tries to measure or
estimate channel parameters, it is assumed these channel parameters remain
unchanged for time periods longer than the typical packet transmission time as
discussed in [ 17 ]. The remaining parameters are summarized in Table 7.1 .
Before running the UDG model and the proposed correlated shadow fading
coverage model, Dijkstra's shortest path scheme was used to test the network
connectivity and only connected graphs were used in the simulations [ 13 ].
Furthermore, we discard network topologies that cause the simple UDG forwarding
model to fail to deliver data packets successfully. The proposed circular grid model
shown in Fig. 7.3 which contains an instance of the correlated shadow fading model
is produced for every transmitting node in the simulation area separately. Each node
is placed in the center of the circular field and the correlated shadow fading can be
produced as described in earlier.
Prior to discussing the results of simulations, we illustrate the path choice results
of each algorithm outlined in § 7.6, by describing the outcome of the operation of
all the algorithms described above in an example 50-node network over a 200 × 200
m 2 geographical region (Fig. 7.5 ).
The simulation environment is programmed in Matlab, where the first step is to
generate uniformly distributed random node positions with the help of the built-in
function “ rand. ” Rand command generates values, which lie in the range of “0” and
“1” and is initialized with a different seed in each subsequent simulation.
After the generation of the node positions, the transmission radius of UDG,
which is found on the basis of receiver sensitivity (−101 dBm), was calculated to be
approximately 61 m. All nodes are assigned Cartesian coordinates to represent their
locations and the distances to all other nodes are calculated by Euclidean equation.
All nodes that lie within the distance of 61 m, i.e., transmission radius, are one-hop
neighbors. All one-hop neighbors will have the “ ppr ” (probability of packet reception)
as “1” while outside the transmission radius the “ ppr ” for UDG is taken as “0.” In the
case of UDG and Dijkstra the one-hop neighbors are exactly the same.

Table 7.1 Chosen simulation parameters
Parameters Used in Simulations
Height of transmitter antenna

1.5 m

Gain of transmitter antenna 2.16 dBi

Height of receiver antenna 1.5 m

Gain of receiver antenna 2.16 dBi

Power transmitted from transmitter −27 dBW

Receiver sensitivity −131 dBW

Transmitter/receiver frequency 2.1 GHz

Packet length 128 bits

BER model Coherent QPSK

Neighbor selection threshold for probabilistic greedy algorithm 0.5

Degree of Reed-Muller forward error correction coding 3

Path-loss model [ 4 ]