Anchor Selection Algorithm for Mobile Indoor Positioning using WSN with UWB Radio

Positioning a person or an object has become essential in many applications. It already exists solutions for outdoor positioning such as satellite based techniques (i.e. GPS) but indoor positioning still remains a great challenge for applications like sports monitoring, contextual visits of museum, Building Information Modeling (BIM) or automated drone missions. Classical approaches using radio communication such as WiFi, Bluetooth, ZigBee only give an accuracy of approximately 2.5 meters when the mobile is static, of course worse when moving. Recently some new radio communication chipsets have emerged based on Ultra Wide Band (UWB) communications. UWB allows accurate Time Of Flight (TOF) measurements, and thus distances estimations between nodes equipped with. A positioning algorithm named Best Anchor Selection for Trilateration (BAST) based on position prediction and noise estimation is proposed. Then a wearable, light and low power Wireless Sensor Network (WSN) prototype (named Zyggie) including an UWB chipset has been developed for algorithms comparison. Finally, an experimental testbed using real cases experiments shows that BAST can give from 1.26 up to 4.17 times better accuracy than low complexity state of the art algorithms when the mobile/person is in movement (e.g. tennis player).