论文部分内容阅读
通过对注意涣散时驾驶员头部运动及面部表情变化特征的分析,系统实时监测驾驶员眼睛、嘴巴位置和运动状态信息,构建驾驶员注意涣散特征表征参量,实现对驾驶员注意涣散状态信息的检测与提取。驾驶员注意涣散表征量具有复杂的非线性特征,利用BP神经网络非线性识别的优势对驾驶员注意特征进行模式分类,实现驾驶员不同注意涣散状态下的特征捕捉。同时采用Dempster-Shafer证据推理技术,对驾驶注意涣散多源表征信息进行决策融合,实现对驾驶员注意涣散状态的判断。结果表明,BP神经网络与D-S规则多源信息决策融合技术的运用提高了驾驶员注意涣散特征检测的准确性和可靠性。
By analyzing the characteristics of driver’s head movement and facial expression changes, the system real-time monitoring driver’s eyes, mouth position and movement status information to build the driver’s attention to the characteristics of the dispersion parameters to achieve the driver’s attention status information Detection and extraction. The driver noticed that the amount of scattered measurement has complex non-linear features, and used the advantages of BP neural network to identify the driver’s attention patterns and classify the driver’s attention features. At the same time, Dempster-Shafer evidence reasoning technology is used to fuse the multi-source characterization information of driver’s attention and make judgment on driver’s distraction status. The results show that the combination of BP neural network and D-S rule-based multi-source information decision fusion technique improves the accuracy and reliability of the driver’s attention spat feature detection.