Composite Peer Hand-Shake Radio Map for Indoor WLAN Localization

Wi-Fi fingerprint localization using received signal strength indicator (RSSI) values has been widely adopted for indoor WLAN environments due to its accuracy and low deployment cost. However, the indoor environment comprises different schematic dimensions, e.g., for offices, corridors, and lobby spaces, under the same Wi-Fi coverage range. We propose a novel fingerprinting approach that leverages the schematic dimensional size of the target area to construct the radio map: The Peer Hand-shake (PHS) database. It deploys competitive pattern learning for representation between adjacent candidate reference vector points during offline calibration phase. Our extensive experimental analysis in a 15 m-by-9 m target area, using conventional nearest neighbor-based localization algorithms, show that PHS outperforms comparisons with average localization error below 2.7 m for all localization algorithms at affordable cost.