Validating ambulatory gait assessment technique for hazard sensing in construction environments

Abstract Wearable sensing technology offers the opportunity to collect an individual's physical response data in occupational settings; recent studies highlight the usefulness of such data for safety management on a construction site. The authors previously studied a worker's gait movements (measured by Wearable Inertial Measurement Units, WIMUs), providing evidence for the pertinence of collecting workers' gait information to identify fall hazards in a steelwork environment. However, gait parameter measurement accuracy has not been examined in settings with different walk-path conditions and hazards, like those in construction environments. To address this knowledge gap, this study evaluates the accuracy of a WIMU-based gait parameter measurement and examines each gait parameter's ability to distinguish hazardous conditions. Data collection was performed in a laboratory environment with both a WIMU-based system and a marker-based motion tracking system (i.e., the reference system). The presented WIMU-based approach resulted in a 6.48 mean absolute percentage error (MAPE) in a non-hazard condition and manifested comparable performance for distinguishing hazardous environments compared to the reference motion tracking system. The results of this study indicate the validity of an ambulatory gait parameter measurement and demonstrate an opportunity to measure changes in gait parameter values caused by hazards.