基于压电阵列与ACO-Elman神经网络的步态检测系统研究  

Research on Gait Detection System Based on Piezoelectric Array and ACO-Elman Neural Network

作　　者：陈美意 张加宏[1,2] 祁博宇[2] 孟辉 朱涵 CHEN Meiyi;ZHANG Jiahong;QI Boyu;MENG Hui;ZHU Han(Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China;School of Electronics and Information Engineering,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China)

机构地区：[1]南京信息工程大学,江苏省大气环境与装备技术协同创新中心,江苏南京210044 [2]南京信息工程大学,电子与信息工程学院,江苏南京210044

出　　处：《电子器件》2023年第1期261-267,共7页Chinese Journal of Electron Devices

基　　金：国家自然科学基金项目(41875035);江苏高校品牌专业建设工程二期项目(电子信息工程);江苏高校优势学科Ⅲ期建设工程资助项目(PAPD);教育部产学合作协同育人项目(202102281011)。

摘　　要：鉴于压电传感器能够感测压力动态变化,采用柔性聚偏氟乙烯(PVDF)压电薄膜阵列作为足底压力检测单元,并提取其测得的压力波形中周期、频率、强度等特征参数作为蚁群算法优化的艾尔曼(ACO-Elman)神经网络的输入信息,通过神经网络训练学习,进而实现人体站立、行走、跑步、跌倒等基本步态活动的分类检测。实验测试结果表明:不同的步态具有各自的特点,基于压电阵列与神经网络的步态检测系统对人体基本活动的预测分类具有较高的准确性,总体正确率超过85%。In view of the ability of piezoelectric sensor to sense the dynamic changes of pressure,the flexible polyvinylidene fluoride(PVDF)piezoelectric film array is used as the plantar pressure detection unit,and the characteristic parameters such as period,frequen-cy,and intensity of the measured pressure waveform are extracted as the input information of the ant colony algorithm optimized Elman(ACO-Elman)neural network,and then the classification and detection of basic gait activities such as standing,walking,running,and falling of the human body is achieved through the training and learning of neural network.Experimental test results show that different gaits have their own characteristics.The gait detection system based on piezoelectric array and neural network achieves high accuracy in predicting and categorizing basic human activities,and the overall accuracy rate exceeds 85%.

关 键 词：步态检测 压电薄膜阵列 足底压力检测 ACO-Elman神经网络 预测分类 

分 类 号：TN384[电子电信—物理电子学] TP212[自动化与计算机技术—检测技术与自动化装置]