基于驾驶人瞳孔直径的追尾冲突辨识方法研究

A Rear Collision Identification Method Based on Drivers' Pupil Diameter

以提高交通冲突辨识与驾驶人认知一致性为目的,以驾驶人心理负荷对瞳孔直径变化的影响为理论基础,以追尾冲突下瞳孔直径在时序上的变化为研究对象,使用傅里叶变换在频谱层面对数据进行逐点扫描并完成模板瞳孔直径数据曲线和全程瞳孔直径数据曲线的匹配,最终提取出与模板数据对应时间车辆运行状态相似度较高的时段,辨识结果相对人工同步录像辨识结果的准确率为71.4%。证明了基于驾驶人瞳孔直径的追尾冲突辨识方法的可行性和有效性,为提高交通冲突辨识与驾驶人认知的一致性提供了崭新而有效的方法。

The temporal variation of the pupil diameters in rear collision is analyzed. The test data are point-wise scanned on the spectrum level by the Fourier transformation. The template and the total pupil diameter curves are matched based on the effect of drivers＇ mental load on their pupil diameters to improve the consistency of the traffic colli- sion identification and the drivers＇ cognition. The periods similar to the period corresponding to the template＇s vehicle mo- tion state are extracted finally. The accuracy rate of the identification result in contrast with the artificial synchronous vid- eo identification is 71.4%. The feasibility and effectiveness of the collision identification method based on drivers＇ pupil di- ameters are verified. A new and effective way is provided to improve the consistency of traffic conflict identification and drivers＇ cognition. However, the identification accuracy is affected by template selection and illumination, which will be further studied in the future.