Adaptive clustering with transmission power control in wireless sensor networks

Abstract

Transmission power control allows a node to dynamically change its power level for energy saving. Many adaptive clustering algorithms propose to use different power levels for clustering. However, the transmission power control had never been integrated as a step in the algorithms. Analysis of the algorithm is done based on assumption that nodes are capable of switching between different power levels. This paper attempts to highlight the possible overhead incurred due to applying power control algorithm in an adaptive clustering in Wireless Sensor Networks. The side effects of executing power control algorithm every time cluster heads rotate can possibly cancel all performance gained if communication overhead is not taken into account. This paper identifies the energy overhead and delay time as two main factors to consider for integration to be successfully implemented. We perform analysis of these factors on existing clustering algorithms such as EECS and MOECS. The analytical results show that the energy overhead is dependent on network size and the number of cluster head candidates. We also show that the delay time involved in switching power levels has to remain low for effective clustering process. (6 pages)