Novel Dynamic KNN with Adaptive Weighting Mechanism for Beacon-based Indoor Positioning System

Abstract

This work proposes a novel dynamic K Nearest Neighbor (KNN) with an adaptive weighting (DKNN-AW) mechanism that performs beacon-based indoor positioning. Four cases are used to prove that DKNN-AW (Dynamic KNN algorithm with Adaptive Weight algorithm) is better than KNN (k-Nearest Neighbors algorithm), KNN-W (KNN with Weight algorithm), DKNN (Dynamic KNN algorithm) and DKNN-W (Dynamic KNN with Weight algorithm). The experimental results demonstrate that, in terms of approximate positioning accuracy, the proposed mechanism outperforms exiting mechanism such as KNN, DKNN, KNN-W and DKNN-W.