Changeset 2820

Timestamp:

10/30/09 20:35:59 (4 weeks ago)

Author:

jaein

Message:

 

Location:

HydroWatch/Tim/doc/ipsn10

Files:

• sec_conc.tex (modified) (1 diff)
• sec_eval.tex (modified) (1 diff)
• sec_related.tex (modified) (1 diff)

HydroWatch/Tim/doc/ipsn10/sec_conc.tex

@@ -3 +3 @@
 We have presented a framework for changing the way we believe environmental wireless sensor networks should be utilized for long-term deployments. Building on previous work in this area, we argue that the ``always on'' approach to environmental sensing places an inherent bias towards network responsiveness being a key performance feature. By turning \emph{off} the radio for long periods, we believe we can achieve a much greater network utility by redirecting energy away from idle listening towards increasing data fidelity at the same time as maintaining a reasonable level of network responsiveness.
 
-Given this model of thinking we propose a means by which nodes can adapt and optimize their operating point in order to maximize network utility given some prediction of future energy resources. As such we are advocating moving away from the ``slow and steady'' approach to long-term sensing where a node adopts a fixed operating point and attempts to consume as little energy as possible. Given we have the ability to estimate future energy resources, we believe nodes should consume as \emph{much} energy as possible in order to fully utilize all available harvested energy and maximize both the data fidelity and network responsiveness. Using retrospective environmental data we have shown how our system performs over extended periods in being able to sustain long-term operation whilst maximizing network utility given energy constraints. Finally we compare our approach with other proposed methods for adapting node operating point, and show that we are able to achieve greater levels of network utility at the same levels of energy consumption.
+Given this model of thinking we propose a means by which nodes can adapt and optimize their operating point in order to maximize network utility given some prediction of future energy resources. As such, we are advocating moving away from the ``slow and steady'' approach to long-term sensing where a node adopts a fixed operating point and attempts to consume as little energy as possible. Given we have the ability to estimate future energy resources, we believe nodes should consume as \emph{much} energy as possible in order to fully utilize all available harvested energy and maximize both the data fidelity and network responsiveness. Using retrospective environmental data, we have shown how our system performs over extended periods in being able to sustain long-term operation whilst maximizing network utility given energy constraints. Finally we compare our approach with other proposed methods for adapting node operating point, and show that we are able to achieve greater levels of network utility at the same levels of energy consumption.
 
 %By moving away from the ``sample-listen-send'' paradigm which has dominated so much of sensor network protocols to date, we see a wealth of new research questions for the community. 
-Moving forward we intend to further demonstrate the benefit of our proposed approaches via long-term field trials. Of particular interest will be cases where there is significant variation in the amount of energy nodes in the network harvest. Currently there is not an energy-aware component in the routing protocol, thus we believe the inclusion of such a parameter will useful in these scenarios. Finally, we believe many of the ideas presented in this paper will have broad applicability in the growing field of multimedia networks, where nodes can assume many operating points with high variations in power consumption and performance. Optimizing the operating point for these types of nodes will become crucial for long-term operation and will greatly change the ways these kinds of technologies can be used into the future.
+Moving forward we intend to further demonstrate the benefit of our proposed approaches via long-term field trials. Of particular interest will be cases where there is significant variation in the amount of energy nodes in the network harvest. Currently there is not an energy-aware component in the routing protocol, thus we believe the inclusion of such a parameter will be useful in these scenarios. Finally, we believe many of the ideas presented in this paper will have broad applicability in the growing field of multimedia networks, where nodes can assume many operating points with high variations in power consumption and performance. Optimizing the operating point for these types of nodes will become crucial for long-term operation, and will greatly change the way these kinds of technologies can be used into the future.

HydroWatch/Tim/doc/ipsn10/sec_eval.tex

@@ -235 +235 @@
 our algorithm without prediction and our algorithm with
 prediction. For these results, we can say that our algorithm
-achieves no zerodays and similar or higher level of utility as the
-Vigorito's algorithm and Hsu's algorithm while maintaining smaller
-excess energy in the battery. This results can be also
+achieves no zerodays and a similar or higher level of utility as the
+Vigorito's algorithm and Hsu's algorithm while maintaining a smaller
+excess energy in the battery. These results can also be
 shown as utility per duty-cycle, which is illustrated
-in Figure~\ref{fig:comparison_D_U}, where the slope for three
+in Figure~\ref{fig:comparison_D_U}, where the slope of the three
 graphs represents the utility per duty-cycle for each
 algorithm. The comparison of these slopes shows that our

HydroWatch/Tim/doc/ipsn10/sec_related.tex

@@ -81 +81 @@
 and sensing, whereas previous algorithms only focused on the abstract notion 
 of duty-cycle, not articulating how such duty-cycle is translated in a real system.
-Second, our work uses high-level environment prediction model while previous
+Second, our work uses a high-level environment prediction model while previous
 works estimate environment using exponentially averaged history of the solar
 energy supply or battery level increase. Our work can predict the energy flow 
 using atmospheric and weather models, thus achieving better adaptability even 
-under harvested solar energy variation under severe weather changes.  
+under harvested solar energy variation with severe weather changes.  
 In addition, while our energy architecture supports energy harvesting, 
 the main premise of our work is that node batteries will eventually expire,