Research Article
RoomE and WatchBLoc: A BLE and Smartwatch IMU Algorithm and Dataset for Room-Level Localisation in Privacy-Preserving Environments
@INPROCEEDINGS{10.1007/978-3-031-34776-4_15, author={Ada Alevizaki and Niki Trigoni}, title={RoomE and WatchBLoc: A BLE and Smartwatch IMU Algorithm and Dataset for Room-Level Localisation in Privacy-Preserving Environments}, proceedings={Mobile and Ubiquitous Systems: Computing, Networking and Services. 19th EAI International Conference, MobiQuitous 2022, Pittsburgh, PA, USA, November 14-17, 2022, Proceedings}, proceedings_a={MOBIQUITOUS}, year={2023}, month={6}, keywords={indoor localisation semantic mapping smartwatch ambient IoT device BLE IMU probabilistic graphical models HMM dataset}, doi={10.1007/978-3-031-34776-4_15} }- Ada Alevizaki
Niki Trigoni
Year: 2023
RoomE and WatchBLoc: A BLE and Smartwatch IMU Algorithm and Dataset for Room-Level Localisation in Privacy-Preserving Environments
MOBIQUITOUS
Springer
DOI: 10.1007/978-3-031-34776-4_15
Abstract
The increasing at-home way of living, which became essential due to the COVID-19 pandemic, yields significant interest in analysing human behaviour at home. Estimating a person’s room-level position can provide essential information to improve situation awareness in smart human-environment interactions. Such information is constrained by two significant challenges: the cost of required infrastructure, and privacy concerns for the monitored household. In this paper, we advocate that ambient bluetooth signals, from IoT devices around the house, and inertial data from a smartwatch can be leveraged to provide room-level tracking information without additional infrastructure. We contribute a comprehensive dataset that combines real-world BLE RSSI data and smartwatch IMU data from two environments, which we use to achieve room-level indoor localisation. We propose an unsupervised, probabilistic framework that combines the two sensor modalities, to achieve robustness against different device placements and effectively track the user around rooms of the house, and examine how different configurations of IoT devices can affect the performance. Over time, through transition-events and stay-events, the model learns to infer the user’s room position, as well as a semantic map of the rooms of the environment. Performance has been evaluated on the collected dataset. Our proposed approach boosts the localisation accuracy from(67.77\%)on average in standard BLE RSSI localisation, to(81.53\%).
