Research On The Strategies For Energy Management In Wireless Sensor Networks

The rapid developments of wireless sensor technology make the use of wireless sensor networks becomes reality. It can be deployed in the monitored region, and then endless relevant information of the target region will be returned. In most wireless sensor networks applications, sensor nodes'energy is limited. For the reasons of large number of nodes and environment, energy can not be added to the sensor nodes. Battery capacity of sensor nodes would not be significantly enhanced in the near future, so it is necessary to search for new strategies for energy management in wireless sensor networks.Energy management in wireless sensor network nodes involves three levels: first, the node-level energy management including the choice of low-power sensor node and dynamic voltage scaling (DVS); Second, the network-level energy management mainly concentrated in the data link layer and network layer protocol, including reducing the volume of data, avoiding conflicts and the "multi-hop" communications; Third, system-level energy management, which considers energy consumption from the point of entire network. And the strategy of dynamic power management (DPM) was used to ensure effective coverage of the target regions.Node-level and the network-level strategy of energy managements in wireless sensor networks are restricted by their inherent characteristics of hardware and network, so they can not maximize the life of sensor networks. Dynamic power management defines several modes based on the sensor node's divided states: In normal time, sensor networks work in a low-power mode; when something unusual monitored, sensor networks work in a real-time mode. The adjustment of modes is triggered by three events: something unusual is monitored by sensor nodes or something unusual is forecasted based on the stream (the historical data the networks collected) or a portent of unusual is obtained from the trend analysis on the stream. Wavelet analysis and auto regression are used in forecasting and trend analysis. Compared with traditional strategies of energy management, dynamic power management can maximize the life of wireless sensor networks by effectively reducing energy consumption of nodes and balancing energy consumption in the networks.