Abstract:

In order to more accurately assess the driving risk within the intersection, firstly, the driver ECG data was introduced and the approximated ideal solution ranking (TOPSIS) model based on the cosine similarity distance was proposed. Secondly, the rolling time window method was established to improve the traditional short-term frequency domain index calculation method. The HRV indicators such as low frequency (LF) variability index and low frequency to high frequency ratio (LF/HF) were calculated in the model. The time domain indicators of the model included the heart beat cycle (R-R interval) rate and its Standard Deviation Normal to Normal heart beat (SDNN). Then, the driving risk in the intersection area was assessed by ranking the ECG indicators in terms of their proximity to their resting state counterparts in a composite manner, according to the principle that the closer the indicators were, the smaller the driving risk was. Finally, 30 drivers were tested at 23 intersections of Xi'an to collect test data and to validate the method. The results indicate that the similarity of drivers' driving risk assessment results at the same intersection is higher than 90.1%, which indicates that the proposed method can be applied to evaluate the overall driving risk at the intersection even with sparse sample. The variance of the assessment of high driving risk intersections is 38.8% and 67.9% higher than that of medium and low driving risk intersections respectively, indicating that driving risk is more accurately assessed in low risk intersection areas.

Key words: intersection, driving risk assessment, electrocardiograms(ECG), heart rate variability(HRV), technique for order of preference by similarity to ideal solution (TOPSIS)