Energy-Driven Partitioning of Signal Processing Algorithms in Sensor Networks
In a sensor network, as we increase the number of nodes, the requirements on network lifetime, and the volume of data traffic across the network, it is often efficient to move towards hierarchical network architectures (e.g., see ). In such hierarchical networks, sensor nodes are clustered into groups, and their roles are divided into master and slave nodes for more efficient structuring of network traffic. The opera tional complexity of each sensor node and the amount of data to be transmitted across sensor nodes strongly influence the energy consump tion of the nodes, which ultimately determines the network lifetime. This paper provides a new way of reducing data traffic across nodes by determining and exploiting the lowest data token delivery points within an application graph that is distributed across a network. The technique divides an application graph into two sub-graphs and then distributes each divided subgraph over a master node and its associated slave nodes. The buffer costs of the graph edges over the cutting line corre sponds to the amount of data to be transmitted between nodes after allo cating the two partial subgraphs such that one subgraph executes on a master node, and the other subgraph is distributed across the associated slave nodes. Since the energy consumption on each node is dominated by the transceiver, the reduced data traffic allows for reducing the turn-on time of the transceivers, and thereby leads to high energy savings. This technique also distributes the workload of sensor nodes in a sys tematic manner. The more balanced workload also contributes to effi cient battery usage, and also improves the latency for processing the data frames captured by the sensor nodes.
KeywordsSensor Network Sensor Node Network Lifetime Data Frame Master Node
Unable to display preview. Download preview PDF.
- 3.Buck, J.T., Lee, E.A.: Scheduling Dynamic Dataflow Graphs with Bounded Memory using the Token Flow Model. In: Proc. ICASSP (April 1993)Google Scholar
- 4.Eker, J., et al.: Taming heterogeneity — the Ptolemy approach. Proceedings of the IEEE (January 2003)Google Scholar
- 5.Heinzelman, W., Chandrakasan, A., Balakrishnan, H.: Energy-efficient communication protocol for wireless microsensor networks. In: Proceedings of the Hawaii International Conference on System Sciences (2000)Google Scholar
- 6.Kalavade, A., Subrahmanyam, P.A.: Hardware / Software Partitioning for Multi-function Systems. In: Proc. International Conference on Computer Aided Design, November 1997, pp. 516–521 (1997)Google Scholar
- 7.Ko, D., Bhattacharyya, S.S.: Dynamic configuration of dataflow graph topology for DSP system design. In: Proceedings of the International Conference on Acoustics, Speech, and Signal Processing, Philadelphia, Pennsylvania, March 2005, pp. 69–72 (2005)Google Scholar
- 8.Kumar, R., Tsiatsis, V., Srivastava, M.B.: Computation Hierarchy for In-network Processing. In: The 2nd ACM international conference on Wireless sensor networks and applications, pp. 68–77 (2003)Google Scholar
- 10.Lindsey, S., Raghavendra, C., Sivalingam, K.: Data Gathering in Sensor Networks using the Energy Delay Metric. IEEE Transactions on Parallel and Distributive Systems, special issue on Mobile Computing, pp. 924–935 (April 2002)Google Scholar
- 11.Shih, E., Cho, S., Ickes, N., Min, R., Sinha, A., Wang, A., Chandrakasan, A.: Physical layer driven protocol and algorithm design for energy-efficient wireless sensor networks. In: Proc. ACM MOBICOM 2001 (July 2001)Google Scholar
- 12.Singh, M., Prasanna, V.K.: System-Level Energy Tradeoffs for Collaborative Computation in Wireless Networks. Kluwer, Norwell (2002)Google Scholar