Effects of subjective driving workload on hazard perception performance in older drivers: A driving simulation study.

Abstract

Objectives: The aim of this research was to explore the changes in eye movement characteristics, driving reaction, and operation performance of older drivers under different intersection conflict scenarios and to investigate the impact of subjective workload on hazard perception performance of older drivers.

Methods: A driving simulation experiment was conducted to simulate various intersection conflict risk scenarios, and data on driving behavior and visual characteristics were collected from 47 middle-aged and older drivers. The NASA-Task Load Index (TLX) scale was utilized to assess subjective workload across six dimensions: Mental Demands, Physical Demands, Temporal Demands, Own Performance, Effort, and Frustration. This study analyzed drivers' subjective workload, eye movement characteristics, driving responses, and performance under different conflict risk scenarios. The rank sum ratio method was employed for a comprehensive evaluation of subjective workload, categorizing participants into low-, moderate-, and high-workload groups. Differences in eye movement patterns, driving responses, and performance across these groups were subsequently examined. Furthermore, the XGBoost-SHAP model was applied to explore the impact of subjective workload on hazard perception performance.

Results: The results show significant differences in Mental Demands, Physical Demands, Temporal Demands, and Frustration levels between older drivers in different conflict scenarios (such as pedestrians, cars, and motorcycles), with lower subjective workload reported in car conflict scenarios compared to others. In motorcycle and pedestrian conflict scenarios, older drivers exhibited heightened visual alertness and attention, while their driving responses and operational behaviors were notably more cautious. Additionally, in left-turn scenarios, drivers displayed more frequent visual searches and driving maneuvers. As subjective workload increased, older drivers' average fixation time and scanning intensity in conflict risk areas showed a marked increase, while vertical eye movements and pupil diameter exhibited a downward trend. Simultaneously, indicators such as brake response time, average speed, steering wheel angle, lateral acceleration, and pedal opening and closing degree all showed a decreasing trend. Importantly, the analysis revealed that the three dimensions of Own Performance, Mental Demands, and Physical Demands in the subjective workload assessment positively influenced older drivers' hazard perception performance, with Own Performance having the greatest impact. In contrast, the effects of Effort and Frustration were relatively smaller, and Temporal Demands had the least influence.

Conclusions: Results show a significant association between driving subjective workload and hazard perception performance among older drivers. It is possible to consider how to utilize subjective workload assessment to identify drivers with poor hazard perception performance and potentially help older drivers improve driving safety.