基于信息融合的驾驶警觉度识别方法研究

Recognition Method of Driving Alertness Based on the Fusion of Information

为了对驾驶警觉度水平进行有效识别,基于眼动信号和心电信号特征指标的信息融合构建了一种驾驶警觉度水平的识别方法。通过驾驶行为绩效验证了驾驶警觉度等级划分的合理性,在此基础上对眼动信号和心电信号进行特征参数的提取和筛选,结合支持向量机(support vector machine, SVM)以融合眼动信号和心电信号为输入构建驾驶警觉度识别模型,并通过30名驾驶员的实验数据对模型进行检测。结果表明不同等级警觉度下的驾驶行为存在显著差异性,验证了驾驶警觉度等级划分的合理性;基于信息融合构建的模型识别效果更优,其识别准确率为89.23%,使用单模态眼动和心电指标分别构建的模型识别准确率为84.36%和81.65%,该方法可用于驾驶警觉度识别以提高识别准确率。

In order to effectively identify the level of driving alertness, a recognition method of driving alertness level is constructed based on the information fusion of eye movement signal and ECG signal. Firstly, the rationality of the classification of driving alertness is verified through driving behavior performance. On this basis, the feature parameters of eye movement signal and ECG signal are extracted and filtered, and then a driving alertness recognition model is developed by combining the support vector machine(SVM) with fusion eye movement signals and ECG signals as inputs. Finally, the proposed model is tested based on the experimental data with 30 drivers. The results show that the driving behavior under different levels of alertness is significantly different, which verifies the rationality of the classification of driving alertness. Moreover, the model based on information fusion has a greater recognition accuracy, since its accuracy rate is 89.23%, while those of the models constructed using single-mode eye movement and ECG indicators are 84.36% and 81.65%, respectively. Overall, the proposed method can be used for driving alertness recognition to improve accuracy.