An iteratively tuned fuzzy logic movement model in WSN using particle swarm optimization

In contrast to adding new nodes, relocation of deployed nodes in mobile wireless sensor networks seems to be an effective solution to cope with undesirable, unpredictable and uncontrolled network topology changes due to nodes´ drift and failure. At the price of less global control, there is a trend in recent years towards giving nodes more autonomy and devising localized relocation algorithms to address challenges of network topology control in harsh and hostile environments in the absence of centralized control. Inspired by laws of nature, a large variety of distributed node relocation algorithms have been designed to alleviate undesirable oscillations caused by local interactions and uncertainties among autonomous nodes as they reach their desired formations. Force-based distributed relocation algorithms governed by virtual push-pull forces among autonomous nodes are among such aforesaid algorithms. Adapting fuzzy logic model in exerting proper amount of forces to reduce node movement oscillation seems to be promising as its conforms well with uncertainties and interactions of autonomous nodes. However, parameters of fuzzy logic relocation model should be tuned so to enable nodes to exert proper amount of forces among their in-range neighbours. In this paper, by using particle swarm optimization, parameters of fuzzy relocation model are obtained based on the desired combinations of performance metrics within nodes´ range in each movement iteration. The result shows that our model either outperforms or matches DSSA movement model.