科研之友
微信
新浪微博
Facebook
分享链接
摘要
Localization is one of the challenges in wireless sensor networks, especially those without the aid of a global positioning system. Use of a dedicated positioning device incurs additional cost and reduces battery life; therefore, a range-free localization scheme is promising as a cost-effective approach. However, the main limitation of this approach is that the estimation precision can be affected by factors such as node density, sensing coverage, and topology diversity. Thus, this study investigates and proposes a method for improving a traditional range-free-based localization method (centroid) that uses soft computing approaches in a hybrid model. This model integrates a fuzzy logic system into centroid and uses an extreme learning machine (ELM) optimization technique to capitalize on the strengths of both approaches: the former is properly used with low node density and short coverage, while the latter is used for the opposite-to achieve a robust location estimation scheme. The ratios of known nodes within the sensing coverage range to the total known nodes and of the sensing coverage range to the maximum coverage range are used as adaptive weights for this hybrid model. To further improve the efficiency, especially in heterogeneous topologies, the concept of resultant force vectors is applied to this hybrid model over particle swarm optimization to mitigate the effects of irregular deployments. The performance of the proposed method is extensively evaluated via simulations that demonstrate its effectiveness compared to other state-of-the-art soft-computing-based range-free localization schemes (i.e., centroid, a fuzzy logic system, and a support vector machine with a traditional ELM).
- 出版日期2018-4
- 单位南阳理工学院
