驾驶员脑电模糊熵与酒后驾车交通事故倾向相关性研究

Study on Correlation between Fuzzy Entropy of Drivers' EEGs and Traffic Accident Proneness to Drunk Driving

少数驾驶员适度饮酒后安全驾驶能力不降反升和不同人对酒精作用存在个体差异的事实表明:利用血液酒精浓度(BAC)判定酒后驾车行为的技术方案存在不足。为探索新型量化参数,结合酒后状态诱发试验和模拟驾驶测试,测取12位驾驶员不同程度饮酒后的脑电(EEG)和规范化交通事故倾向指标。结果发现:随着饮酒量的增加,驾驶员左额叶区EEG的瞬时复杂度和长时周期度都增加。结合模糊熵算法构造并计算EEG特征参数。相关性分析表明:在驾驶员从未饮酒到重度饮酒的过程中,其左额叶区EEG模糊熵和规范化事故倾向指标间的平均相关系数为0.76,二者正相关。因此,根据该模糊熵变化能判断其酒后安全驾驶能力的变化。

Using blood alcohol concentration as a reference to recognize drunk driving behaviors shows inherent deficiencies, since there exist facts that drinking alcohol does not necessarily to reduce but increase the ability to drive safety and different people may get different alcohol sensitivities. In order to explore a new substitutive technology, drunken state inducing experiments and driving simulation were designed so as to measure 12 qualified drivers＇ EEGs and traffic accident proneness. Experiment results showed that there were an increase in both instant complexity and long-term periodicity of EEGs measured from left frontal lobe as drivers drunk more. A characteristic parameter for EEG was constructed and calcu- lated on basis of the fuzzy entropy algorithm. Correlation study indicated that average correlation coefficient between fuzzy entropy of EEGs and normalized traffic accident proneness was 0.77, which proved they were positive correlated, which means that fuzzy entropy of EEGs from left frontal lobe could be taken as a novel for to recognizing a change in ability to drive safety.