Steering Wheel Angle Excursions As A Measure Of Fatigue-Related Driver Performance Impairment

Abstract Introduction Fatigue from sleep loss and circadian misalignment causes automobile driving performance impairment. Metrics based on steering wheel angle, which is straightforward to measure, could be used to quantify this impairment. As the tail of the distribution of steering wheel angles (absolute magnitude of deviation from center) increases with fatigue, we investigated whether driving performance impairment could be quantified based on the prevalence of steering wheel excursions beyond a given angle threshold. We used data from two published laboratory studies of simulated shift work, in which fatigue remained low during day shifts but increased across time awake during night shifts. Methods N=37 healthy adults (ages 26.8±5.2y; 25 men) were assigned to a simulated night shift schedule (awake 20:00-10:00) or day shift schedule (awake 08:00-22:00; study 1 only). After an adaptation period, participants underwent two 5-day shift cycles with an intervening rest period. Driving performance was measured on a high-fidelity simulator during adaptation (data not used) and four times at 3h intervals during each shift day. Every drive involved 30min driving at 55mph, including ten 0.5mi uneventful straightaways being considered here. Steering wheel angle was measured at 72Hz (study 1) or 60Hz (study 2). A total of 1,471 drives (31,394,498 angle measurements) were available for analysis. Results We investigated angle thresholds across 0.01-0.25rad in 0.01rad intervals and counted the corresponding number of threshold excursions in each drive for each participant. For study 1, we applied mixed-effects ANOVA with fixed effects for condition and time awake, and their interaction, and determined the local effect size for interaction. A 0.03rad (1.7°) threshold yielded the greatest effect size, f²=0.031 (small). For this threshold, we repeated the analysis using the data from both studies, controlling for study. The interaction was significant (F[3,1428]=13.23, p<0.001), showing low driving impairment across time awake during day shifts but increasing impairment across time awake during night shifts. Conclusion The prevalence of steering wheel excursions beyond a 1.7° angle threshold yielded sensitivity to fatigue-related driving performance impairment during simulated night shifts. Further research will extend our results to driving through curves and with greater fatigue levels. Support (if any) FMCSA DTMC75-07-D-00006