4.4 Article

Canine-inspired Unidirectional Flows for Improving Memory Effects in Machine Olfaction

期刊

出版社

OXFORD UNIV PRESS INC
DOI: 10.1093/icb/icad016

关键词

-

类别

向作者/读者索取更多资源

A dog's nose allows air to flow in a unidirectional path, creating stagnant zones of air that contribute to its physical memory. In this study, experiments and simulations were conducted to compare the effects of bidirectional and unidirectional flows on odor detection. The unidirectional setting exhibited a slow return to baseline levels, indicating memory effects. The findings suggest that these memory effects can potentially enhance the sensitivity and utility of electronic noses.
Synopsis A dog's nose differs from a human's in that air does not change direction but flows in a unidirectional path from inlet to outlet. Previous simulations showed that unidirectional flow through a dog's complex nasal passageways creates stagnant zones of trapped air. We hypothesize that these zones give the dog a physical memory, which it may use to compare recent odors to past ones. In this study, we conducted experiments with our previously built Gaseous Recognition Oscillatory Machine Integrating Technology (GROMIT) and performed corresponding simulations in two dimensions. We compared three settings: a control setting that mimics the bidirectional flow of the human nose; a short-circuit setting where odors exit before reaching the sensors; and a unidirectional configuration using a dedicated inlet and outlet that mimics the dog's nose. After exposure to odors, the sensors in the unidirectional setting showed the slowest return to their baseline level, indicative of memory effects. Simulations showed that both short-circuit and unidirectional flows created trapped recirculation zones, which slowed the release of odors from the chamber. In the future, memory effects such as the ones found here may improve the sensitivity and utility of electronic noses.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.4
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据