A new data aggregation technique in multi-sink wireless sensor networks

Abstract

As an essential operation in multi-sink wireless sensor networks (WSNs), the main aim of data aggregation is collecting the data stored in the sensor nodes to the sink nodes and its key points are shortening the aggregation time and prolonging the network lifetime. We know that the nearest-sink (NS) scheme has the better performance in shortening the aggregation time but it does not do well in prolonging the network time, while another scheme called round-robin (RR) has opposite result. To balance the performance between aggregation time and network lifetime, a scheme called time-window and voronoi-partition based aggregation scheduling (TVAS) is proposed in this paper. Firstly, according to the Voronoi algorithm we divide the sensing area into numbers of partitions, each of which owns several sub-regions, in which all sensor nodes will choose the same scheme to forward the data. Secondly, Recourse-Arbitration- Response (RAR) mechanism is presented to dynamically achieve the destination sink node for the sensor nodes in each sub-region which reserves several candidate sink nodes. Finally, based on the data pressures of the candidate sink nodes, the aggregation scheduling of the sensor nodes in one sub-region should be switched between the NS scheme and the proposed TVAS scheme. A performance evaluation reveals that the proposed TVAS scheme can achieve the balance between reducing the aggregation time and prolonging the network lifetime of the WSN in an asymmetric data generation environment.