基于模糊神经网络的驾驶警觉度识别方法研究

Recognition Method of Driving Vigilance Based on Fuzzy Neural Network

为了有效识别驾驶警觉度,构建了一种基于脑电信号的驾驶警觉度识别方法。首先,以主观疲劳测评、驾驶行为绩效作为量化指标,验证驾驶警觉度等级划分的合理性。在此基础上,对脑电信号数据进行小波变换提取特征参数,作为驾驶警觉度的识别特征指标,结合模糊神经网络构建了驾驶警觉度识别模型。最后,采用该模型对20名驾驶员连续驾驶3h的脑电数据进行试算。结果表明:通过对前后时段的主观疲劳与行为数据进行对比分析,两时段数据存在着显著差异性,说明驾驶警觉度等级划分具有合理性;采用模糊神经网络的识别结果优于BP神经网络,其模型识别正确率为81.29%～84.95%,且平均正确率为83.12%,该方法可用于驾驶警觉度的识别。

In order to recognize driving vigilance effectively,an recognition method of driving vigilance based on electroencephalogram(EEG)was constructed.Firstly,the driver′s subjective fatigue measure and driving behavior performances were used as quantitative indexes to validate the rationality of driving vigilance grade division.The wavelet transformation was used to extract the characteristic parameters for the EEG data,which can be used as the recognition characteristic indexes of driving vigilance.Meanwhile,combined with fuzzy neural network,the recognition model for driving vigilance was established.Finally,the model was used to calculate the EEG data of 20 drivers who drove 3 hours continuously.The results show that through comparing and analyzing subjective fatigue and behavioral data at early and later period,there are significant differences between the data of two periods,which shows that the partition of driving vigilance grade is reasonable;the recognition result of fuzzy neural network is better than that of BP neural network,the recognition accuracy rate of the model is between 81.29%~84.95%,and the average accuracy rate is 83.12%,which indicates the method can be used to recognize driving vigilance.