Sensor Localization in Digital-Twin Representations using Mixed Reality and IoT for Smart Buildings

Sammanfattning

Smart buildings equipped with Internet of Things (IoT) sensors have significant potential for optimizing energy use, enhancing safety, and improving occupant comfort. However, their effectiveness is fundamentally limited by the challenge of accurately localizing sensors within the physical environment. Without precise spatial context, smart building systems cannot make informed decisions or provide reliable environmental monitoring. This thesis presents the design, implementation, and evaluation of a novel indoor localization and visualization system that integrates Bluetooth Low Energy (BLE)-based Angle of Arrival (AoA) positioning with extended reality (XR) platforms. This research addresses the challenge of accurate tracking of IoT sensors indoors, where traditional Global Positioning System (GPS) is ineffective, and camera-based solutions face limitations such as occlusion and lighting sensitivity. The proposed system leverages the u-blox XPLR-AOA-3 evaluation kit in combination with Visual-Inertial Odometry (VIO) for enhanced spatial awareness. BLE tags, placed on objects, transmit signals to anchor devices strategically positioned in various configurations. These anchors compute the AoA, enabling 3D position estimation. Data is streamed to a Unity-based 3D environment, synchronized via TCP/UDP communication protocol, and visualized on Microsoft HoloLens 2 and Magic Leap 2 headsets. Experiments were conducted in two distinct indoor environments, employing multiple anchor geometries, tag heights, and static and dynamic tag scenarios over 24 trials. Results demonstrated sub-meter (< 1 m) accuracy under controlled conditions, with performance influenced by room size, anchor placement, tag height, and environmental factors. The study also identified practical challenges, including system efficiency and limited flexibility due to the reliance on wired connections. Local network restrictions drove this choice, as the UDP-based wireless option could not be made to function reliably in the available network environment. In addition, the environmental sensor data (temperature, humidity, pressure) was reused from a previous project and integrated into the Unity application to provide contextual overlays, while all localization accuracy results (RMSE, stability) were produced using newly collected BLE–AoA position data. This research contributes a practical framework for accurate indoor sensor positioning that bridges the gap between physical IoT deployments and their digital representations. The findings demonstrate the feasibility of BLE-AoA localization for smart building applications and provide insights into optimal deployment strategies for real-world implementations. This approach of radio-based localization with immersive visualization establishes a foundation for more intelligent and spatially aware building management systems.