多通道数据融合的飞行员操作意图分析与触觉反馈技术  

Pilot Operation Intention Analysis and Tactile Feedback Technology Based on Multi-Channel Data Fusion

作　　者：黄文博 王长元[2] 贾宏博 周高豪 HUANG Wenbo;WANG Changyuan;JIA Hongbo;ZHOU Gaohao(School of Optoelectronic Engineering,Xi’an technological University,Xi’an 710021,China;School of Computer Science and Engineering,Xi’an Technological University,Xi’an 710021,China;Air Force Medical Center,PLA,Air Force Medical University,Beijing 100142,China)

机构地区：[1]西安工业大学光电工程学院,西安710021 [2]西安工业大学计算机科学与工程学院,西安710021 [3]空军军医大学中国人民解放军空军特色医学中心,北京100142

出　　处：《西安工业大学学报》2022年第2期178-187,共10页Journal of Xi’an Technological University

基　　金：国家自然科学基金资助项目(52072293)。

摘　　要：为了解决传统意图推断方法不能很好地利用多生理源数据融合优势的问题,提出一种通过融合脑电信号、眼动信号和触觉反馈的人机交互意图推断方法,识别被试发出的预期命令。通过对脑电、眼动特征的提取,根据特定于预期最终用户的输入数据对CNN进行不同的预训练,对每个试验者的数据进行特征级数据融合得出推断结果,最后通过触觉反馈输出。研究结果表明:相同任务下,该系统的分类性能比基于传统几何方法提高了9.1%;多通道数据融合的意图推断中加入触觉反馈信号可以将意图识别准确率提高至71.99%。Traditional methods for intention inference do not make good use of the advantages of data fusion of multiple physiological sources.To solve this problem,the paper presents a method for inferring the human-computer interaction intentions by fusing EEG,EM signals and tactile feedback.The method has been trained to recognize the expected commands given by the experimenter.Through the extraction of EEG and EM features,different pre-trainings of CNN were carried out according to the input data specific to the expected final experimenter.The results were obtained by performing the feature-level data fusion to each experimenter's data,and then were output through the tactile feedback equipment.The study shows that for the same task,the classification performance of the system is improved by 9.1%compared with the traditional geometric method and that the addition of tactile feedback signals to multi-channel data fusion can increase the accuracy of intention recognition to 71.99%.

关 键 词：模拟飞行驾驶 眼动特征 脑电信号 数据融合 触觉反馈 

分 类 号：V328.1[航空宇航科学与技术—人机与环境工程]