Modified Belief Propagation Algorithm for Energy Saving in Wireless and Sensor Networks

In this paper we develop a simple variant of the belief propagation (BP) algorithm targeted towards reducing the communication cost of broadcast supporting networks such as sensor networks. Our modified algorithm can be proven to achieve identical computation results on each node, while at the same time significantly reducing the number of transmitted messages. Furthermore, the number of computation steps in each node is not increased. We pay in additional storage space, as each node saves the last message received from each of his neighbors. Many recent works in the field of sensor networks use the BP algorithm for solving various problems. Our variant could be easily applied to any of these proposed systems, thus improving energy consumption and running time. One of the important issues in sensor networks is how to minimize of the sensors’ energy consumption, reducing the number of transmitted messages by a large factor significantly improves the energy consumption. Another advantage of our algorithm is a decrease in the number of transmission slots needed, since in wireless environment acquiring a transmission slot can be costly in terms of time and power.