Joint distributed scheduling and tree formation for heterogeneous wireless sensor networks

Data collection is very basic application of Wireless Sensor Networks. Sensor nodes are deployed in region to be observed. They sense quantity of interest. It may be temperature, pressure, humidity, solar radiation and so on. The sensed readings are transferred to a central node known as sink or base station. Often networks are heterogeneous in nature i.e. more than one types of sensors are deployed in given region. In homogeneous networks, only one type of nodes are present. Once nodes are deployed, a tree is formed rooted at sink. It is difficult to achieve perfect aggregation in heterogeneous networks. Unlike homogeneous networks, nodes may have more than one types of packets coming in and going out. To maximize chances of aggregation, node should forward every packet to a parent where it can be aggregated. Thus a node may have to select more than one parents. This means that tree formation process should be done keeping heterogeneity in consideration. We have extended joint distributed scheduling and tree formation proposed in literature for heterogeneous networks. It is found that the extended algorithm maximizes aggregation, requires less number of time-slots to schedule the network and conserves energy. It is also found that scheduling heterogeneous network results in more control overhead than homogeneous network.