基于压电振动能量采集器的信息感知与模式识别系统  

Information perception and pattern recognition system based on piezoelectric vibration energy harvester

作　　者：颜佟佟 鲁征浩[1] 徐大诚[1] YAN Tongtong;LU Zhenghao;XU Dacheng(Micro-Nano Sensor Technology Research Center,Soochow University,Suzhou 215006,China)

机构地区：[1]苏州大学微纳传感技术研究中心,江苏苏州215006

出　　处：《传感器与微系统》2022年第7期37-39,43,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金重点资助项目(61834007)

摘　　要：压电振动能量采集器不仅能高效收集环境振动能量,同时也可感知振动信息,通过进一步信息处理也可识别出振动的模式。在充分研究振动能量采集器输出特征的基础上,构建了一种振动信息感知以及振动模式识别系统。根据能量采集器输出的信号特征,采用卷积神经网络(CNN)算法给出了振动模式识别方法,并通过现场可编程门阵列(FPGA)的算法运行实现了振动模式的实时快速识别。实验结果表明:采用卷积神经网络算法的模式识别准确率可达96.7%,基于FPGA的识别系统能在能量采集器触发后的0.6 s内完成振动模式的快速识别。The piezoelectric vibration energy harvester can not only collect the environmental vibration energy efficiently,but also sense the vibration information.Through further information processing,the vibration pattern can be identified.On the basis of fully studying the output characteristics of vibration energy harvester,a vibration information perception and vibration pattern recognition system is constructed.According to the output signal characteristics of energy harvester,a vibration pattern recognition method based on convolutional neural net-work(CNN)algorithm is proposed.By running the algorithm on field programmable gate array(FPGA),the real-time and fast identification of vibration pattern is realized.The experimental results show that the recognition accuracy of CNN algorithm can reach 96.7%,and the recognition system based on FPGA can complete the rapid recognition of vibration pattern within 0.6 s after the energy harvester is triggered.

关 键 词：压电振动能量采集器 信息感知 卷积神经网络 模式识别 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置] TP183[自动化与计算机技术—控制科学与工程]