A Link Quality Estimation Method for Wireless Sensor Networks Based on Deep Forest

Mu He,Jian Shu
DOI: https://doi.org/10.1109/access.2020.3047648
IF: 3.9
2021-01-01
IEEE Access
Abstract:In wireless sensor networks, sensor nodes, the miniature embedded devices, have limitation of energy, storage, computing, and etc. One of the tasks of the nodes is to use their limited resources to complete work efficiently. Choosing high quality link communication can effectively save energy. In this paper, we propose a link quality estimation model that is based on deep forest. To avoid a noise sample becoming a center point in the clustering, we use an improved K-medoids algorithm based on step increasing and optimizing medoids (INCK) when dividing the link quality grades. During the sample preprocessing stage, the Pauta criterion is used to delete the noise link samples, and we fill the mean value of each grade into the missing values. The feature extraction performance of deep forest is improved by combining the stratified sampling to change the unbalance distribution of link quality samples. And then the Stratified Sampling Cascade Forest link quality estimation (SCForest-LQE) is constructed by combining stratified sampling with cascade forest. The experiments are conducted in three real application scenarios. Compared with the existing six link quality estimation models, SCForest-LQE has better estimation performance and stability.
computer science, information systems,telecommunications,engineering, electrical & electronic
What problem does this paper attempt to address?