Indoor location service in support of a smart manufacturing facility

• Published in: Computers in industry Vol. 103; pp. 132 - 140
• Main Authors: Carrasco, Ulises, Urbina Coronado, Pedro Daniel, Parto, Mahmoud, Kurfess, Thomas
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.12.2018
• Subjects: Android OS; Bluetooth; Cloud; Indoor localization; Industry 4.0; RSSI; RSSI; Bluetooth; Industry 4.0; Android OS; Indoor localization; Cloud
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/j.compind.2018.09.009

•System finds the nearest CNC machine to a user with a ERP app and Bluetooth beacons.•RSSIs are sent to the cloud, which hosts the RSSIs and returns the nearest machine.•The techniques of nearest neighbor (NN), WKNN and Bayesian inference were compared.•A combination of Kalman filter and Bayesian inference provided higher accuracy.•Having the same amount of Bluetooth beacons and machines increases accuracy. Location awareness in manufacturing facilities has high potential to produce information according to the space in which is important. This work presents a system that finds the nearest machine to a user. The system runs in an Android app which is part of a mobile enterprise resource planning system. The indoor localization system collects the received signal strength indicators (RSSIs) from low-cost Bluetooth beacons installed in the machines. The RSSIs are sent to the cloud which hosts the values and returns the name of the nearest machine. This work uses the fingerprinting algorithm to map the location of each machine with a set of four RSSIs. The results show the nearest neighbor, weighted k-nearest neighbor and Bayesian inference techniques, the latter presenting the best accuracy. Using the Kalman filter on the RSSIs reduces variability, which increases the correct machine coincidences. The use of one Bluetooth beacon per machine, guessed the correct machine around 89% of the time.