摘要

In wireless sensor networks, sensor nodes providing small and simple services are usually needed to collaborate with each other to perform complex user tasks. Due to the lack of infrastructure in such a network, service collaboration should be done with a decentralized manner, but existing approaches do not fully take resource conservation and excessive control overhead into account. This paper proposes a novel distributed service collaboration algorithm for wireless sensor networks, with the goal of saving sensor resources along with low control overhead. Based on ideas of swarm intelligence, we exploit autonomic ants encapsulating service collaboration requests to selectively visit the more potential sensor nodes and services depending on pheromone tables rather than blind flooding. A service collaboration recovery mechanism is taken to handle the failures, such as link breakage, node departure or energy depletion. Our simulation results demonstrate the effectiveness of the proposed algorithm.

全文