International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B1. Istanbul 2004 
  
velocity vector and etc.) for producing the integrated maps. One 
possible solution can be to have computational nodes inserted in a 
Ring topology to minimize the data transfer rates and also 
improve the products availability time. The Ring configuration 
can have some advantages in data flow rates and overall operating 
efficiency. 
s4 
Hub Controller 
Cy? 
sn) Q) I Os 
© Oa 
4 
s1 Sensor 
  
  
  
  
Figure 4. Star concept communicating via single hub. 
Star Topology: In this configuration shown in Figure 4, all nodes 
are directly linked to a central node. The central node serves as a 
hub controller to provide bi-directional communication pathway 
to all the radial distributed nodes. This configuration offers some 
disadvantage i.e., it limits the number of nodes due to its push 
through capacity since each respective node has to connect via 
this hub. On the contrary it puts burden on the hub handler in 
terms of its overall push through capacity and efficiency. This 
type of architecture is beneficial when shorter turn around times 
are required and observing environment is somewhat confined in 
terms of its geographic boundaries. One can deploy multiple Star 
links over a larger area and then connect them via a super hub 
handlers which is very similar to the computer clusters for large 
distributed networks. A good example depicting the use of this 
configuration can be a Star configuration in an inclinde LEO 
(Low Earth Orbit) collecting precipitation data or continuously 
looking for other disasterous events such as wild fires or volcanoe 
eruptions. The sensors can be spread out over a large distance to 
provide a wider spatial coverage. The central hub can be located 
in a geostationary orbit or on the ground. This configuration can 
be further augmented with ground based network of sensors as 
well. The ground sensors can be provided in fixed locations to 
closely monitor the localized area and the space based sensors can 
provide a broader perspective. In case of a suspect situation, 
either sensor i.e., ground based or space based may trigger an 
event signal for a more focused view or an analytical feedback to 
the decision makers. 
Chain Topology: This configuration allows sensors to link 
linearly in an open path. This is rather a simple if individual 
sensors are linked with each other. However, it can be very 
complex arrangements if multiple Star or Ring topologies are 
s1 s2 s(n-1) sn 
SO 
~ \ 4 7 
v \ / Sta 
= * 4 L^ 
Communication Node 
Figure 5. Sensors in linear connectivity. 
linked with each other. This scheme is useful when large areas 
need to be covered in shorter time and it should be applicable to 
both space and ground deployments. In fact, it can be an 
interesting concept to install such arrangements along the 
earthquake fault lines around the globe and link them all further 
with a space based observer to continuously monitor the tectonic 
movements and deformation. It is conceivable that this may not 
give as comprehensive Earth's surface deformation maps as 
synthetic aperture radar (SAR), but can be a less expensive 
alternate to a constellation of SARs. 
Hybrid Topologies: lt is possible to mix network structures 
combining any of the above configurations. In fact, this probably 
would be a common practice and practical approach in the future. 
However, this can be a communications nightmare if all the 
collected information ends up going through some main 
communication nodes. In fact, an integrated web composed of 
many architectures can be more efficient if it has intelligent 
processing and its own communications paths for extensive data 
flow and frequent controls. The main communication nodes 
should only be used for the integration of the critical information 
and dissemination. 
7. SUMMARY 
As described above, the sensor web observing architecture can 
employ large number of agile sensors operating from multiple 
vantage points to simultaneously collect suites of observations 
from multiple regions. This further can improve temporal, spatiale 
and spectral resolutions. However, it is still crucial to have a 
strong tie to the applications and societal benefits as one of the 
most critical justification for such endeavors. No doubt, there is 
lot to gain in the science arena and answer many unknown 
phenomena as they command our planet, but such knowledge 
gain must be applied to more accurately predict hurricanes, air 
quality, water resources and others. There is also a great potential 
in understanding and predicting earthquakes and volcanic 
eruptions. This naturally requires many more observations to feed 
into the analytical engines and sensors webs may be the way to 
proceed. 
REFERENCES 
S. Habib, “Technology Thrust for Future Earth Science 
Application,” Optical Remote Sensing of the Atmosphere and 
Clouds, Second International Asia-Pacific Symposium on Remote 
Sensing of the Atmosphere, Environment, and Space, Sendai, 
Japan, 9-12 October 2000. 
M. Schoeberl, “The Afternoon Constellation: A Formation of 
Earth Observing Systems for the Atmosphere and Hydrosphere,” 
IEEE International Geoscience and Remote Sensing Symposium, 
Toronto, Canada, 24-28 June 2002. 
S. Habib and P. Hildebrand, “Sensor Web Architectural Concepts 
and Implementation Challenges - An Heuristic Approach,” SPIE 
International Symposium on Remote Sensing, Crete, Greece, 
September 22-27, 2002. 
S. Habib and N. Nokra, “Transferring Knowledge from 
Observations and Models to Decision Makers,” SPIE Europe, 
Barcelona, Spain,September 12-16, 2003. 
S. Talabac, “Sensor Node Aggregate and Topology 
Considerations,” submitted to SPIE Asia Pacific Symposium, 
Hawaii, Nov 2004. 
LU te fg" .. 1 ee Ld 
— 4 
mM CS 5 Ce