Design of fingerprinting technique for indoor localization using AM radio signals

摘要

Due to lack of GPS signals' indoor coverage, the implementation of Signals of Opportunity (SoOp) such as FM signals, Wi-Fi, Bluetooth, RFID, and Cellular Networks for indoor localization has been considered. Not all SoOp are good performers because of some constraints; for instance, lack of accuracy, higher cost in deployment, additional hardware requirements. In this research, a new method of indoor localization is developed using AM radio signalsand thereceived signal strength fingerprinting technique. The fingerprinting technique in this study uses the deterministic approach and three algorithms, Nearest Neighbour, K-Nearest Neighbour, and K-Weighted Nearest Neighbour to measure the mean distance error as the indicator of positioning accuracy. The outcome shows that the minimum mean distance error at the ground and first floor is less than 3m and the Nearest Neighbour algorithm performed best with K=1.