International Journal on Advanced Science, Engineering and Information Technology, Vol. 8 (2018) No. 3, pages: 998-1007, DOI:10.18517/ijaseit.8.3.6086

Optimization of Fingerprint Indoor Localization System for Multiple Object Tracking Based on Iterated Weighting Constant - KNN Method

Abstract

Indoor localization promises a lot of benefits on the application in various fields. The fingerprint method is often used because it has high mobility, low network cost, and high compatibility. However, the distance and RSSI relationships are non-linear which decreases the accuracy of the system. KNN is required as a matching algorithm to solve the problem. The error result of Fingerprint-KNN system for indoor localization is still less satisfactory, therefore weighting factor is added in KNN algorithm as a modification to optimize the accuracy and precision of the localization system. The usual W-KNN is adding a value in form of the distance error from estimation result. In this paper, the constant as the result of iteration process within a range is multiplied by the error value which is added to the system as a weighting of KNN algorithm. The iterated weighting constant provides optimization on the system up to 25% better than the conventional system.

Keywords:

indoor localization; fingerprint; iterated weighting constant-KNN

Viewed: 1148 times (since abstract online)

• Show fulltext