Failure detection of mobile nodes in medical asset tracking system

A medical asset tracking system monitors the in/out status of a medical asset with a mobile node which moves around a monitoring area, leaves it, and then returns to the region repeatedly. Due to failure of the mobile nodes, the system may determine that a mobile asset is outside the region despite the fact that the mobile asset is inside. In this paper, a novel failure detection method is proposed to resolve this problem. It uses two properties of the neighboring nodes of a mobile node disconnected from a sensor network. One is the trend of the neighbor counts; the other is that of the ratios of the boundary nodes contained in the neighbors. With these trends, the proposed method detects failed nodes from mobile nodes disconnected from a sensor network. The experimental results show that the proposed method can detect most failures of mobile nodes, including those that the conventional approaches fail to observe.