An Improved Dv-Hop Localization Algorithm Based On Machine Learning For Wireless Sensor Network

The DV-Hoplocalization method has a series of superiorities, such as high distributiveness and expandability, which is perfectly fit for large-scale deployment. Moreover, it may lead to reasonable positioning accuracy merely in isotropous dense network. In practical environment, most scenes are anisotropic, with unevenly distributed nodes. In this paper, kernel PCR method is applied to collect and utilize the correlation between hop count and real distance, so as to build an optimal relationship model, converting hop count information between nodes into the value of real distance, so that DV-Hop method may be applicable to different environment. Compared with existing similar and typical methods, the method proposed in this paper has higher environment adaptability, as well as higher positioning accuracy and stability.