Musculoskeletal disorders (MSDs) are common among working populations, especially manufacturing workers, with exposure to non-neutral postures frequently cited as a risk factor. However, the magnitudes and precision of risk estimates vary between field-based studies, as it is difficult to continually follow and sample large study samples with time-varying exposures to non-neutral postures. Development of a low cost location-tracking system may help overcome this methodological limitation. The purpose of this thesis was to explore the utility of radio frequency identification (RFID) technology for extracting task-specific exposure data from full-shift measurements of upper arm posture as machine-paced assembly workers rotated job tasks.
Full-shift upper arm posture and movement velocities were recorded using inertial measurement units (IMUs) across up to 15 consecutive working days from among a sample of 8 participants. Workers scanned RFID tags with RFID readers at job task workstation as they started and finished performing a task, effectively measuring task duration. At the end of each shift, workers self-reported task duration estimates in a diary. Self-report and RFID-based measurement bias and agreement range were estimated using Bland-Altman analyses. Fully nested, random-effects analysis of variance (ANOVA) models were used to estimate the relative contribution of components of exposure variance to overall posture and movement exposure variance.
The study observed a slight measurement bias for self-reported task duration estimates when comparing both incomplete (i.e., single measurement from either self-report or RFID methodology) and complete task observation data (i.e., measurements from both methodologies), while the RFID system displayed a similar bias when comparing only complete task observation data. However, regardless of the data set, a large measurement agreement range was observed. The between-subjects and between-tasks-within-day (and within-subject) variance components generally contributed the most to total exposure variance, with the between-day-within-subject component contributing little if nothing at all. Depending on velocity level summary measure, between 65.7% and 84.5% of the total exposure variance was associated with the between-tasks-within-day (and within-subject) component. The RFID system did prove useful in extracting task-specific exposure data from full-day IMU measurements. However, there were unexpected instances in which workers failed to follow RFID system user protocol and generate irregular timestamp sequences. Future research and development is encouraged to refine the pairing of RFID technology with IMUs for ergonomic exposure assessment. Specifically, an active RFID system with adjustable read range could potentially overcome the limitation of requiring that a worker place the RFID tag within inches of the low frequency RFID reader to perform a scan.