Monitoring workers on indoor construction sites using data fusion of real-time worker's location, body orientation, and productivity state

Systematic control of construction operations using a Real-Time Locating System (RTLS) is becoming common in providing support to the management of complex construction sites. This paper proposes a novel indoor construction site monitoring system to track workers' location, workspace, body orientation, and productivity state at the same time. The system works based on three modules: (I) workspace discovery; (ii) body orientation detection; and (iii) productivity state identification. The workspace discovery module integrates location data with the 4-Dimensional Building Information Model (BIM) of the job site. Then, the body orientation detection module identifies workers' field of view onsite using Received Signal Strength Indicator (RSSI) values between Bluetooth Low Energy (BLE) beacons. Last but not least, the productivity state identification module relies on an accelerometer sensor embedded in body-mounted beacons. The results of performed experiments in the laboratory show a mean error of less than 30° in detecting the worker's focus orientation for a period of eight minutes. For identifying direct work and support work states, the productivity state identification module yielded an average error of 14% and a maximum error of 20%.