作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2019, Vol. 45 ›› Issue (2): 278-283. doi: 10.19678/j.issn.1000-3428.0049387

• 开发研究与工程应用 • 上一篇    下一篇

基于PSO混合特征选择算法在疲劳驾驶中的应用

林雨培,陈兰岚,邹俊忠   

  1. 华东理工大学 化工过程先进控制和优化技术教育部重点实验室,上海 200237
  • 收稿日期:2017-11-22 出版日期:2019-02-15 发布日期:2019-02-15
  • 作者简介:林雨培(1993—),男,硕士研究生,主研方向为电生理信号处理、模式识别;陈兰岚,副教授;邹俊忠,教授。
  • 基金资助:

    国家自然科学基金(61201124);中央高校基本科研业务费重点科研基地创新基金(222201717006)。

Application of Hybrid Feature Selection Algorithm Based on Particle Swarm Optimization in Fatigue Driving

LIN Yupei,CHEN Lanlan,ZOU Junzhong   

  1. Key Laboratory of Advanced Control and Optimization for Chemical Process,Ministry of Education,East China University of Science and Technology,Shanghai 200237,China
  • Received:2017-11-22 Online:2019-02-15 Published:2019-02-15

摘要:

基于多源生理信号的驾驶员疲劳检测研究存在特征信息冗余以及佩戴多种传感器影响驾驶员操作的问题。为此,提出一种结合粒子群优化算法和序列后向选择的特征选择算法。在适应度函数中加入信号源数的惩罚项,在降低特征维度的同时减少传感器的使用数量。根据所使用分类器的特点对适应度函数进行简化,提高特征选择算法的运行效率。在粒子定义中加入信号选择位,提高信号的筛选力度。实验结果表明,该算法平均使用2种信号和16.1种特征,能够获得95.3%的疲劳驾驶检测正确率。

关键词: 疲劳驾驶, 多源生理信号, 混合特征选择, 粒子群优化, 序列后向选择

Abstract:

There are some problems in driver fatigue detection research based on multi-source physiological signals,such as the redundancy of characteristic information and the influence of wearing multiple sensors on driver operation.Therefore,a feature selection algorithm combining Particle Swarm Optimization(PSO) algorithm and Sequential Backward Selection(SBS) is proposed.The penalty term of signal source number is added to the fitness function to reduce the number of sensors while reducing the feature dimension.According to the characteristics of the classifier used,the fitness function is simplified and the efficiency of the feature selection algorithm is improved.The signal selection bit is added to the definition of particle to improve the signal screening.Experimental results show that this algorithm uses an average of 2 signals and 16.1 features,and can achieve an accuracy of 95.3% in fatigue driving detection.

Key words: fatigue driving, multi-source physiological signals, hybrid feature selection, Particle Swarm Optimization(PSO), Sequential Backward Selection(SBS)

中图分类号: