Proposal of Separate Channel Fingerprinting Using Bluetooth Low Energy

Abstract

BLE (Bluetooth Low Energy) based localization is a next candidate for indoor localization. In this paper, we propose a new BLE-based fingerprinting localization scheme that improves localization accuracy. BLE is a narrow band communication that is highly affected by frequency selective fading and channel gains. We employ channel-specific features to location fingerprint to improve localization accuracy. Current BLE standards provide no API to retrieve an advertising channel number. We therefore developed a separate channel advertising scheme to measure RSS (Received Signal Strength) on each advertising channel. To demonstrate the feasibility of the separate channel fingerprinting, we conducted initial evaluations. Experimental evaluations showed that the separate channel fingerprinting improves accuracy by approximately 12%.