Method for RSSI heat map estimation using Radial Basis Functions for fingerprinting localization

Abstract

Radio Frequency (RF) based indoor localization is a challenging task due to the dynamic and multipath factors although many research proposed different improvement strategies in their indoor localization works. This paper discusses an alternative method for RF-based mapping estimation for Received Signal Strength Indicator (RSSI) fingerprinting indoor localization. The RSSI fingerprinting mappings are estimated by different Radial Basis Function (RBF) interpolation methods. The RBF interpolation methods may reduce the number of fingerprint locations required for fingerprinting localization. The predicted RSSI and actual RSSI radio mappings are compared to determine the accuracy of each RBF. The best mean error of 2.008dBm in Inverse Multiquadric RBF is achieved between the actual and predicted mappings in a complex laboratory environment. The RF-based mapping estimation techniques presented in this paper have shown that the effectiveness to fulfil the perquisite of fingerprinting localization with lower time and labour requirement.