An improved self-localization algorithm for Ad hoc network based on extreme learning machine

Wireless sensor network (WSN) is a typical application of Ad hoc network in autonomous system (AS). It has attracted considerable attention in the past. Recent years have witnessed a growing interest in the study of localization algorithm for WSN. Self-localization of nodes is one of the key technologies for application of WSN. The localization accuracy is a significant criterion to evaluate the practical utility of localization algorithm. In most of the localization algorithms, increasing the density of anchor nodes is one of the main strategies to improve the localization accuracy. But the number of anchor nodes is always limited due to the hardware limitations, such as energy consumption, cost, and so on. In this paper, a novel machine learning algorithm is developed to improve the accuracy. Specifically, we propose a DV-Hop self-localization algorithm using extreme learning machine (ELM), called DV-Hop-ELM, which achieves the objective by virtually increasing the number of anchor nodes. The ELM-based single-hidden layer feedforward network (SLFN) is firstly designed to find the appropriate sub-anchor nodes. Moreover, the improved DV-Hop localization algorithm utilizes those anchor nodes and sub-anchor nodes to localize remaining sensor nodes. We test the algorithm DV-Hop-ELM to demonstrate its performance. Compared to classic DV-Hop algorithm, the proposed method improves the localization accuracy while reducing the costs of WSN.