Construction workers are frequently exposed to safety hazards on sites. Wearable sensing systems (e.g. wearable inertial measurement units (WIMUs), wearable insole pressure system (WIPS)) have been used to collect workers' gait patterns for distinguishing safety hazards. However, the performance of measuring WIPS-based gait parameters for identifying safety hazards as compared to a reference system (i.e. WIMUs) has not been studied. Therefore, this study examined the validity and reliability of measuring WIPS-based gait parameters as compared to WIMU-based gait parameters for distinguishing safety hazards in construction.
Five fall-risk events were conducted in a laboratory setting, and the performance of the proposed approach was assessed by calculating the mean difference (MD), mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE) and intraclass correlation coefficient (ICC) of five gait parameters.
Comparable results of MD, MAE, MAPE and RMSE were found between WIPS-based gait parameters and the reference system. Furthermore, all measured gait parameters had validity (ICC = 0.751) and test-retest reliability (ICC = 0.910) closer to 1, indicating a good performance of measuring WIPS-based gait parameters for distinguishing safety hazards.
Overall, this study supports the relevance of developing a WIPS as a noninvasive wearable sensing system for identifying safety hazards on construction sites, thus highlighting the usefulness of its applications for construction safety research.
This is the first study to examine the performance of a wearable insole pressure system for identifying safety hazards in construction.
