Topology control and data handling in wireless sensor networks

Our work in this thesis have provided two distinctive contributions to WSNs in the areas of data handling and topology control. In the area of data handling, we have demonstrated a solution to improve the power efficiency whilst preserving the important data features by data compression and the use of an adaptive sampling strategy, which are applicable to the specific application for oceanography monitoring required by the SECOAS project. Our work on oceanographic data analysis is important for the understanding of the data we are dealing with, such that suitable strategies can be deployed and system performance can be analysed. The Basic Adaptive Sampling Scheduler (BASS) algorithm uses the statistics of the data to adjust the sampling behaviour in a sensor node according to the environment in order to conserve energy and minimise detection delay. The motivation of topology control (TC) is to maintain the connectivity of the network, to reduce node degree to ease congestion in a collision-based medium access scheme; and to reduce power consumption in the sensor nodes. We have developed an algorithm Subgraph Topology Control (STC) that is distributed and does not require additional equipment to be implemented on the SECOAS nodes. STC uses a metric called subgraph number, which measures the 2-hops connectivity in the neighbourhood of a node. It is found that STC consistently forms topologies that have lower node degrees and higher probabilities of connectivity, as compared to k-Neighbours, an alternative algorithm that does not rely on special hardware on sensor node. Moreover, STC also gives better results in terms of the minimum degree in the network, which implies that the network structure is more robust to a single point of failure. As STC is an iterative algorithm, it is very scalable and adaptive and is well suited for the SECOAS applications.