 | |||||||||||||||||||||||
| Site Index | |||||||||||||||||||||||
|
Overview | Results | Publications | Software | People | References
Overview:
Wireless sensor networks (WSNs) are characterized by limited amount of energy supply at sensor nodes. Hence, energy efficiency is an important issue in system design and operation of WSNs.
One of the key tasks performed by any WSN is the collection of sensor data from the sensors in the field to the sink for processing. This task is also referred to as data gathering. In this project, we consider the problem of data gathering in environments where the data from the different sensors are correlated. Such correlation of the data being collected can be leveraged by appropriately fusing the data inside the network to the best extent possible, thereby reducing the number of transmissions and hence energy consumption, for the gathering process. Such correlation aware data gathering strategies typically shift the aggregation structure from a default shortest-path tree (SPT) to a steiner minimum tree (SMT) in order to achieve the required efficiency.
We study the energy efficiency of correlation aware data aggregation trees under various sensor network conditions and the tradeoffs involved in using them. Comprehensive simulations are used to evaluate and derive inferences about the energy efficiency of correlation aware data aggregation trees and theoretical analysis are conducted to justify the inferences.
Based on the insights gained through the above investigation, we propose a simple, scalable and distributed correlation aware aggregation structure that achieves good energy performance across a wide range of sensor network configurations, at the same time addresses the practical challenges of establishing a correlation aware data aggregation structure in resource constrained WSNs.
Results / Status:
In the first part of the project, we study the tradeoffs involved in using correlation aware structures in practical sensor applications. Specifically, we investigate sensor applications with delay tolerance and without delay tolerance, and explore the energy benefits of correlation aware structures in data gathering processes, as well as the tradeoffs for obtaining these energy benefits.
The contributions of this part of research is summarized as follows:
(1) We characterize through quantitative analysis how the energy improvement of a correlation aware aggregation structure is impacted by different network parameters. We show that the energy improvement tends to be bounded by a small constant under many network scenarios. Furthermore, the improvement corresponds to when the additional cost of establishing a correlation aware structure is not taken into account, in the presence of which the improvement will be further reduced.
(2) We characterize the energy performance of SMTs under different sensor data correlations. We show that SMT is an energy efficient aggregation structure only when the correlation degree is relative high. For low to medium correlation degrees, correlation structures that ensure low average aggregation delay are more energy efficient.
(3) We also characterize what the maximum useable delay bound is for achieving the maximum energy efficient structure. We show that the maximum useable delay bound is a small constant times the delay along the maximum length shortest-path in the default shortest path tree.
From the first part of the research, we conclude that theoretically, correlation aware data aggregation does not provide significant energy improvement under many circumstances, and the centralized coordinations usually required for its construction and maintenance may even offset the energy benefit. However, from a practical standpoint, the energy improvement possible is still non-ignorable. If we can design a correlation aware data aggregation structure that achieves close to ideal performance improvement and is simultaneously good for a wide range of node density, source density, source distribution and correlation degree without incurring the overhead of centralized coordination, the benefit of correlation aware data aggregation can be fully utilized, and energy efficient data gathering can be realized.
Based on this observation and insights we obtained from the previous investigation, we design a simple, scalable, and distributed data aggregation approach called SCT (Semantic/ Spatial Correlation-aware Tree). This approach achieves energy improvement close to the practical bounds we established in the previous study with relatively low communication overhead. The SCT structure is instantaneously constructed during the course of a single query delivery and is a fixed structure that is efficient for wide range of network parameters. The SCT approach, with its highly manageable structure, ensures low maintenance overhead of the aggregation structure. Problems such as load balancing, data aggregation synchronization and node failures are properly resolved in SCT, and at the same time efficient aggregation in most general sensor network settings is achieved.
Publications & Presentations:
- Y. Zhu, K. Sundaresan and R. Sivakumar,
"Practical Limits on Achievable Energy Improvements and Useable De lay Tolerance in Correlation Aware Data Gathering in Wireless Sensor Networks, "
IEEE Communications Society Conference on Sensor and Ad Hoc Communications and Networks (SECON), Santa Clara, California, USA, September 2005 (Best Paper Award). - Y. Zhu, R. Vedantham, S.-J. Park and R. Sivakumar,
"A Scalable Correlation Aware Aggregation Strategy for Wireless Sensor Networks ,"
IEEE International Conference on Wireless Internet (WICON), Budapest, Hungary, July 2005. - Y. Zhu, K. Sundaresan and R. Sivakumar,
" Exposing Two Critical Myths about Correlation Aware Data Aggregation,"
Poster Presentation, ACM International Symposium on Mobile Ad Hoc Networking and Computing (MOBIHOC), Urbana-Champaign, IL, USA, May 2005.
Software Downloads:
People:
- Yujie Zhu (Student)
- Ramanuja Vedantham (Student)
- Karthikeyan Sundaresan (Student)
- Seung-Jong Park (Alumnus)
- Raghupathy Sivakumar (Professor)
References & Related Work:
- R. Cristescu, B. Beferull-Lozano, and M. Vetterli, "On Network Correlated Data Gathering", in INFOCOM, Hong Kong, Mar. 2004.
- Pattem S., Krishnamachari B., and Govindan R., "The impact of spatial correlation on routing with compression in wireless sensor networks", in International Symposium on Information Processing in Sensor Networks, April 2004, pp. 28 - 35.
- Bhaskar Krishnamachari, Deborah Estrin, and Stephen B. Wicker, "The impact of data aggregation in wireless sensor networks", in Proceedings of the 22nd International Conference on Distributed Computing Systems, 2002, pp. 575 578.
- C. Intanagoniwawat, D. Estrin, R. Govindan, and J. Heidemann, "Impact of Network Density on Data Aggregation in Wireless Sensor Networks", in International Conference on Distributed Computing Systems (ICDCS 02), Vienna, Austria, July 2002.