Mobile Sink Trajectory Determination for Wireless Sensor Networks Using Dependent Nonparametric Trees

The mobility of sink is one of the significant strategies to gather data in wireless sensor networks (WSNs). However, the reduction or limitation of the trajectory length is more significant issue. In this work, we recommend a novel algorithm to determine the mobile sink (MS) trajectory using hierarchical dependent nonparametric trees (DNT) for WSNs. According to the algorithm, sensor nodes are represented as sub-categories of hierarchical trees. We consider the tree-structured stick breaking process (TSSBP) in the algorithm. The prediction based distribution is performed on the trees, and thanks to the inter-node connections, the MS follows the trajectory hierarchically, starting from the root node. In this way, the MS collects data from all of the nodes. The proposed algorithm has been subjected to the performance comparison of k-means and energy density trajectory (EDT) methods under the same conditions. Simulation results clearly illustrate that the proposed scheme offers more useful contributions to the WSNs from other methods, in terms of network life, trajectory length, and energy requirements.