Wireless sensor network (WSN) has recently received tremendous attention. The interest is growing due to the benefits WSN brings and the large number of unexplored applications. However, WSN faces challenges which limit their usability. One of the most notable is the energy limit. The wireless sensor node can only be equipped with a limited power source (< 0.5 Ah, 1.2V) [8]. In some application scenarios, replenishment of power resources might be impossible. In these cases, sensor node lifetime strongly depends on battery lifetime. In most of WSN, sensor nodes play as not only data originators but data routers. The loss of some sensor nodes in WSN may cause significant overhead of network topological reorganization and rerouting. Meanwhile, many emerging applications require processing which has considerable computation demands and needs to be done in the sensor network. For instance, in target tracking system [214], sensors collaboratively measure and estimate the location of moving targets or classify targets. Another typical application is the video sensor network [71]. Since the size of raw video data is too large to transfer in the network, sensors are required to process some computationally intensive operations.