Bioinspired early detection through gas flow modulation in chemo-sensory systems
Bioinspired early detection through gas flow modulation in chemo-sensory systems
复制标题
DOI：
10.1016/j.snb.2014.09.001
复制
发表时间：
2015-01-01
影响因子：
8.4
通讯作者：
Perera, Alexandre
中科院分区：
文献类型：
--
作者：
Ziyatdinov, Andrey;Fonollosa, Jordi;Perera, Alexandre
仿生化学传感系统的设计是机器嗅觉研究领域的一个热点。这条线的发展可以增加人工化学传感系统的寿命和灵敏度。这种方法是基于生物体中已知的感觉系统，并且由此产生的人工系统的目标是在一定程度上再现生物机制。嗅闻行为，积极采样气味，最近已经在神经科学中进行了研究，并且已经表明呼吸频率是嗅觉系统的重要参数，因为气味感知，特别是在复杂的场景中，例如新的气味剂探索，取决于刺激身份和采样方法。在这项工作中，我们提出了一个化学传感系统的基础上，我们结合外部机械呼吸机模拟生物呼吸循环的16个金属氧化物气体传感器的阵列。测试的气体类别形成了两种分析物，丙酮和乙醇，在二元混合物中的相对广泛的组合。从采集的信号中提取了低频和高频两组特征，表明高频特征包含与气体类别相关的信息。此外，这种信息在测量的早期阶段就可以获得，这可以使该技术适用于早期检测场景。完整的数据集向社区公开。(C)© 2014 Elsevier B.V.保留所有权利。
The design of bioinspired systems for chemical sensing is an engaging line of research in machine olfaction. Developments in this line could increase the lifetime and sensitivity of artificial chemo-sensory systems. Such approach is based on the sensory systems known in live organisms, and the resulting developed artificial systems are targeted to reproduce the biological mechanisms to some extent. Sniffing behaviour, sampling odours actively, has been studied recently in neuroscience, and it has been suggested that the respiration frequency is an important parameter of the olfactory system, since the odour perception, especially in complex scenarios such as novel odourants exploration, depends on both the stimulus identity and the sampling method. In this work we propose a chemical sensing system based on an array of 16 metal-oxide gas sensors that we combined with an external mechanical ventilator to simulate the biological respiration cycle. The tested gas classes formed a relatively broad combination of two analytes, acetone and ethanol, in binary mixtures. Two sets of low-frequency and high-frequency features were extracted from the acquired signals to show that the high-frequency features contain information related to the gas class. In addition, such information is available at early stages of the measurement, which could make the technique suitable in early detection scenarios. The full data set is made publicly available to the community. (C) 2014 Elsevier B.V. All rights reserved.