The success rates rise rapidly and give a good sense of
the algorithm??™s reliability. A more interesting relation can be seen by analyzing
the number of hops traversed. The nodes skip checkpoints if the algorithm
fails to locate a next hop, and so the hop count is variable. The plot shows
a strong correlation between node density and hop count. This can be used
to determine the granularity or resolution of the trajectory that the nodes
will be able to follow. The plot shown is only for the run when k = 10. The
other runs reached the maximum hops available very fast and did not reveal
information of much interest.
For trajectory based forwarding to be useful, the trajectory seen by the
expert and the path traversed by the nodes need to be close. Figure 5(b) can
379
Fernand S. Cohen, Joshua Goldberg, and Jaudelice C. de Oliveira
Fig. 6. CAD snapshot showing the desired curve and the selected nodes.
Fig. 7. CAD snapshot zoom.
be used to estimate an optimal resolution for the checkpoints that the nodes
will be able to follow.
Figures 6 and 7 show a snap shot of the CAD and a zoom into the picture,
respectively. The CAD trajectory is shown in white, the available sensor
nodes are shown as white dots, and the estimated routing path found by the
algorithm is shown in black. As we can see from the simulation results on that
snap shot, the estimated path follows very closely that of the CAD trajectory.
In the same figure, we can also see the osculating search window (triangle) as
it navigates from source to destination node.
Pages:
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606