Efficient power savings in wireless sensor networks with network coding and overhearing avoidance

Wireless sensor networks and embedded systems are becoming commonplace in many fields of research. However, sensor nodes are generally powered by batteries which provide a limited amount of power, and it is often difficult to recharge or replace the batteries. Therefore, power aware and power efficient protocols at each layer of the communications are very important for wireless sensor networks. Various schemes tried to achieve significant power saving, among them; the network coding is one promising technique. To the best of our knowledge, AdapCode is by now, the most promising network coding architecture for power saving in wireless sensor networks. However, the AdapCode has two clear limitations. First, the coding process of AdapCode works based only on partial neighbors of a node which may waste some potential coding opportunities. Second, the AdapCode did not consider overhearing and idle listening issues in packets forwarding, so, it may cause unnecessary battery power expenditure from receiving irrelevant packets. To address the earlier mentioned limitations, we first propose an improved AdapCode scheme (namely AdapCode+) by deploying a power efficient protocol to discover all neighbors of a node, such that the potential coding opportunities can be increased. To further avoid overhearing irrelevant packets, we then enhance AdapCode+ to AdapCode++. The basic idea of AdapCode++ is to compose new digest information about the subsequent packets to be sent out from current node, and then broadcast it together with the conventional wake up message to all the neighbor nodes. In this way, the neighbor nodes can intelligently determine receiving desired packets only. Our NS-2-based simulation indicates that the simple AdapCode+ scheme can improve the AdapCode in terms of power saving, and the AdapCode++ can further make this improvement much more significant.