Cooperative Training in Wireless Sensor and Actor Networks

Abstract

Exploiting features of high density wireless sensor networks represents a challenging issue. In this work, the training of a sensor network which consists of anonymous and asynchronous sensors, randomly and massively distributed in a circular area around a more powerful device, called actor, is considered. The aim is to partition the network area in concentric coronas and sectors, centered at the actor, and to bring each sensor autonomously to learn to which corona and sector belongs. The new protocol, called Cooperative, is the fastest training algorithm for asynchronous sensors, and it matches the running time of the fastest known training algorithm for synchronous sensors. Moreover, to be trained, each sensor stays awake only a constant number of time slots, independent of the network size, consuming very limited energy. The performances of the new protocol, measured as the number of trained sensors, the accuracy of the achieved localization, and the consumed energy, are also experimentally tested under different network density scenarios.