Data Aggregation In WSNs : State Of The Art And A Multi-Criteria Proposal.
Résumé
The basic role of Wireless Sensor Network (WSN) is to collect data from the environment by many sensor nodes. The WSN could be constituted by hundreds or thousands of sensor nodes; hence, it could generate a huge amount of data. The communication of all the sensed data causes high power consumption limiting, as a result, the lifetime of the sensor nodes and consequently the WSN lifetime. Whereas, processing data locally within the sensor node consumes much less than its communication. Thus, it is possible to reduce the power consumption of the sensor nodes and, hence, to extend the network lifetime, by reducing the number and size of the communicated data. This reduction of communicated data is possible, if the sensor nodes' data are aggregated instead of being sent directly to the sink (the base station in WSNs). This chapter studies several methods of data aggregation in WSNs. It classifies them into three categories. The first category represents structure-free mechanisms which abstract the organization of the nodes within the network. It allows the end-user of the network to query each sensor node, if needed. In the second category, we find the structure-based mechanisms that solve the structure-free problems in terms of scalability. It organizes the sensor nodes into a well structured form, such as a tree, or a set of groups. It defines aggregator nodes to aggregate other nodes data. This aggregator node may be one head-node per group or one parent for several children in a tree. Then, the end-user of the network has the aggregated data through the aggre-gator nodes. The third one is a Multi-criteria Data aggregation (McDa) mechanism. It combines characteristics from the structure-free and the structure-based categories. It does not suppose any network structure. However, spontaneously during the same data aggregation session, it is possible to have cooperation between single nodes and