Improved Bias Cancellation and Header Communication for Random Gossiping-based Static Wireless Sensor Networks

Abstract:
Random gossiping provides a communication paradigm for wireless sensor networks so that all sensors can aggregate messages from the entire network without specifying a routing tree or a central unit. Random gossiping leads to a robust aggregation, however, it also causes biased aggregation and long aggregation time in terms of the number of communications among sensors. In our previous work, a scheme is proposed to reduce and even eliminate the aggregation bias and to reduce the number of communications by introducing indicating headers to the messages that are communicated in the network. In this paper, we extend our work assuming a static wireless sensor network. Exploiting that the topology remains constant in static networks, we introduce an improved bias cancellation method which increases the probability to achieve a bias-free aggregation and an algorithm to reduce the number of communications for indicating headers. Simulation results show the reduction of both the number of communications for indicating headers and the bias in comparison to our previous work.