Abstract
Indoor positioning systems (IPS) are used to locate people or objects in environments where the global positioning system (GPS) fails. The commitment to make bluetooth low energy (BLE) technology the leader in IPS and their applications is clear: Since 2009, the Bluetooth Special Interest Group (SIG) has released several improved versions. BLE offers many advantages for IPS, e.g., their emitters or beacons are easily deployable, have low power consumption, give a high positioning accuracy and can provide advanced services to users. Fingerprinting is a popular indoor positioning algorithm that is based on the received signal strength (RSS); however, its main drawbacks are that data collection is a time-consuming and labor-intensive process and its main challenge is that positioning accuracy is affected by various factors. The purpose of this work was to develop a semi-automatic data collection support system in a BLE fingerprinting-based IPS to: (1) Streamline and shorten the data collection process, (2) carry out impact studies by protocol and channel on the static positioning accuracy related to configuration parameters of the beacons, such as transmission power (Tx) and the advertising interval (A), and their number and geometric distribution. With two types of systems-on-chip (SoCs) integrated in Bluetooth 5 beacons and in two different environments, our results showed that on average in the three BLE advertising channels, the configuration of the highest Tx (+4 dBm) in the beacons produced the best accuracy results. However, the lowest Tx (−20 dBm) did not worsen them excessively (only 11.8%). In addition, in both scenarios, when lowering the density of beacons by around 42.7%–50%, the error increase was only around 8%–9.2%.
Highlights
Indoor positioning systems (IPS) have been the subject of much research in recent years, mainly because the global positioning system (GPS) signal suffers from attenuation in many indoor environments, rendering it impossible to use for positioning [1], and because they have multiple and useful applications in real life [2]
Due to its recent introduction, we can only mention that of Karvonen et al [54], whose main goal was to evaluate experimentally the communications range and throughput performance of bluetooth low energy (BLE) 5 coded version, and that of Pancham et al [55], which identified BLE as one of the technologies that promise an acceptable response to the requirements of the healthcare environment and investigated in that context the latest improvements with Bluetooth 5, especially with regard to its range when the signal penetrates through different types of multiple partitions
Euclidean was the distance used for the calculation of weights in the weighted k-nearest neighbor (WKNN) algorithm since in previous works it has been detected that the distance/similarity used is not critical [69]
Summary
Indoor positioning systems (IPS) have been the subject of much research in recent years, mainly because the global positioning system (GPS) signal suffers from attenuation in many indoor environments, rendering it impossible to use for positioning [1], and because they have multiple and useful applications in real life [2]. Those that attempt by some means to reduce the time and effort of the CPh, e.g., path surveys, in which, by means of Gaussian processes, a dedicated user attempts to construct a signal map from a sparse set of fingerprints collected while walking through a space [27], or interpolation-based methods, such as inverse distance weighted (IDW), radial basis function (RBF) or Kriging, which build a more populated RP-database from a few RPs [28,29] Those that are CPh-free, which are subdivided into various categories, such as those that use online RSS measurements only or those that, in some manner, merge CPh and PPh using people, e.g., explicit or implicit crowdsourcing, in which users are involved in data collection, differing between them in the user incentive and degree of participation [1,25].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
