Investigation of methods for reducing calibration effort in RSSI-based indoor locating systems

Abstract

Indoor locating systems (ILSs) for locating assets (e.g. work pieces or tools) in manufacturing enterprises increase transparency, reduce search times on the shop floor and help to optimize intra-logistics of the enterprises. However, despite the potential for efficiency gains the adoption of ILSs in small and medium enterprises (SMEs) is restrained by the associated efforts. For SMEs an ideal ILS would be based on a standard low-cost technology for decreasing hardware costs, offer a simple way of introduction and minimal operational effort. Therefore, the thesis investigates design and associated localization methods of a Bluetooth low energy (BLE)- and received signal strength indicator (RSSI)-based ILS that possesses simplified calibration procedure or even totally eliminates it and, thus, decreases efforts related to its initial deployment. The thesis proposes two research concepts. The first concept is intended for simplification of the calibration procedure (in fact, acquisition of fingerprints) and leverages automatic movement that is available in industry (e.g. a conveyor or a robot) for this purpose. The second concept proposes an idea of spatial distribution of fixed nodes into emitters and receivers. For investigation of the named research concepts a BLE- and RSSI-based ILS has been implemented. As a localization approach, the core idea of the ring overlapping based on comparison of received signal strength indicators (ROCRSSI) range-free localization technique is used in the system. Tests of the implemented ILS conducted in two test venues have shown that fixed nodes spatially distributed into emitters and receivers perform better than the merged ones (both in terms of localization accuracy and in terms of deviation of localization errors from their means). Preliminary calibration of the system with the help of an automatic movement (in this particular case, with the help of a line-following robot) has led to ambiguous results and demands additional investigations.