仿生嗅觉感知系统气体识别和浓度估计模型  

Gas Recognition and Concentration Estimation Model for Bionic Olfactory Perception System

作　　者：相洪涛 张文文[2,3,4] 肖文鑫 王磊 王远西[6] XIANG Hong-Tao;ZHANG Wen-Wen;XIAO Wen-Xin;WANG Lei;WANG Yuan-Xi(School of Automation and Software Engineering,Shanxi University,Taiyuan 030006,China;School of Optical-electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;College of Science,University of Shanghai for Science and Technology,Shanghai 200093,China;School of Electrical and Electronic Engineering,Nanyang Technological University,Singapore 639798,Singapore;School of Computer Science,Peking University,Beijing 100871,China;College of Electronic and Information Engineering,Tongji University,Shanghai 201804,China)

机构地区：[1]山西大学自动化与软件学院,太原030006 [2]上海理工大学光电信息与计算机工程学院,上海200093 [3]上海理工大学理学院,上海200093 [4]南洋理工大学电气与电子工程学院,新加坡639798 [5]北京大学计算机学院,北京100871 [6]同济大学电子与信息工程学院,上海201804

出　　处：《自动化学报》2024年第4期812-827,共16页Acta Automatica Sinica

基　　金：国家自然科学基金(62203307)资助。

摘　　要：常用气体检测模型需要使用气体传感器阵列响应信号的稳态值对气体进行种类识别和浓度估计,而在实际环境中,气体一般处于动态变化的状态,气体传感器阵列响应信号难以达到稳态值或长时间维持稳定状态.针对上述问题,提出一种由动态小波残差卷积神经网络(Dynamic wavelet residual convolutional neural network, DWRCNN)子模型和权重信号自注意力(Weighted signal self-attention, WSSA)子模型组成的气体检测模型.该模型可以直接使用气体传感器阵列的原始动态响应信号对动态变化的气体进行成分识别,并进一步对每种成分气体的浓度在线估计.通过搭建的仿生嗅觉感知系统对模型的性能进行评估,实验结果表明,与常用气体识别模型相比, DWRCNN能获得接近100%气体识别准确率,且在线训练时间短,收敛速度快;与常用气体浓度估计模型相比, WSSA浓度估计模型能够大幅提高气体浓度估计精度,并能同时对不同气体都保持较高气体浓度估计精度,解决了动态环境中仿生嗅觉感知系统需要针对不同气体选择不同最优气体浓度估计模型问题.Conventional gas detection models require the use of steady-state values of the response signals of gas sensor arrays for gas recognition and gas concentration estimation,whereas in real environments,gases are typically in a state of dynamic change,making it difficult or time-consuming to achieve steady-state values for the response signals from gas sensor arrays.To address the aforementioned issues,a gas detection model consisting of dynamic wavelet residual convolutional neural networks(DWRCNN)sub-model and weighted signal self-attention(WSSA)sub-model is proposed.The model can directly identify the components of dynamically changing gases using the raw dynamic response signals from the gas sensor array,and it can also online estimate the concentration of each component gas.The performance of the model is evaluated by the self-built bionic olfactory perception system.The results demonstrate that DWRCNN achieves nearly 100%gas recognition accuracy in comparison to other prevalent gas recognition models,with a short online training time and a rapid convergence speed;The problem of selecting different optimal gas concentration estimation models for different gases in a dynamic environment for bionic olfactory perception systems is solved by the WSSA concentration estimation model,which can significantly improve the gas concentration estimation accuracy while simultaneously maintaining a very high gas concentration estimation accuracy for different gases.

关 键 词：气体识别 浓度估计 仿生嗅觉感知系统 注意力机制 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]