Dynamic Cluster Formation and Management in a Factory Wireless Sensor Network

In this paper, we propose a distributed approach to manage wireless sensor node tracking in a factory environment. The approach uses a computationally efficient distributed k-means algorithm that divides the sensing environment into a set of Voronoi cells. We take advantage of the geometrical structure of the solution, using the resulting Voronoi cells to support inter-cluster trade-off decisions. To demonstrate the wireless sensor network´s adaptability to changing shop floor conditions, we report on a set of experiments performed in JADE (Java Agent Development Framework).