Recovery of Hop Count Matrices for the Sensing Nodes in Internet of Things

The hop count matrices (HCMs) are very helpful in obtaining the location information of sensing nodes in Internet of Things (IoT). However, in some scenarios, the HCMs cannot be completely observed due to abnormal termination of the flooding process, or some of the entries are contaminated by false information in external malicious attacks. Therefore, it is very important to recover the missing HCMs. However, to the best of our knowledge, there is no specific algorithm used in the current research to recover the HCMs, which would cause the positioning accuracy to be seriously deteriorated. In this article, for the scenarios of the entries partially observed in the HCMs, the HCMs recovery schemes, namely, HCMR-NBC and HCMR-MC, are proposed. The former, HCMR-NBC, is to learn the internal relations of different sensing node pairs in the HCMs. It is a simple and fast approach which utilizes the feature with a single dimension to predict the missing hop count values between the sensing nodes. The latter, HCMR-MC, is to transform the problem of the matrices recovery to the one of matrices completion. Compared with the previous SVT and BLMC algorithms, the proposed algorithms have great advantages in terms of the reconstruction performance and the computation complexity.