Advanced Study of Odor Communication

近年来，人们对气味传播的研究逐渐进行了尝试。由于目前的设备无法检测到气味，因此我们的目标是在本研究中开发一种气味交流工具，其中由传感系统或质谱仪检测到的气味由远程气味发生器再生。首先，我们开发了一种由便宜的电磁阀组成的气味混合器。虽然电磁阀只有ON或OFF两种状态，但中间值可以用阀门快速开启或关闭时的开启时间来表示。采用delta-sigma调制方法控制电磁阀，成功研制出8组分气味混合器。其次，利用最优控制输入理论等自适应反馈控制算法，利用传感器阵列和所研制的混合器对气味进行记录。我们成功地记录了由五种成分组成的苹果香精的配方。此外，我们将目标气味近似为仅由少数成分气味组成的气味。将目标9组分气味近似为5组分气味后，人体感官测试显示，近似气味与目标气味足够相似。此外，还进行了以质谱计为传感器的实验。发现在注入体积一定的情况下，叠加定理是成立的。如果建立了自动进样的方法，质谱分析是有前景的。最后，还制作了“气味广播网站”。任何人都可以访问该主页，闻闻远程站点记录的气味。我们成功地利用传感器阵列在远程站点再生了双组分气味。

Study of odor communication has been recently and gradually attempted. Since scents cannot be detected by the present devices for that purpose, we aimed to develop an odor communication tool in the present study, where a smell detected by a sensing system or a mass spectrometer is regenerated by a remote odor generator.First, we developed an odor blender just composed of cheap solenoid valves. Although a solenoid valve takes just two states, ON or OFF, the intermediate value can be expressed by the period of opening valve when it is rapidly opened or closed. We adopted delta-sigma modulation method to control solenoid valves and succeeded in developing the odor blender with 8 component odors.Next, an odor was recorded by the system with a sensor array and the developed blender using an adaptive feedback control algorithm such as theory of optimal control input. We succeeded in recording the recipe of the apple flavor made up of five components. Moreover, we approximated the target odor as the smell composed of just a few component odors. After the target 9-component odor was approximated as 5component one, the human sensory test revealed that the approximated odor was sufficiently similar to the target odor.Moreover, the experiment on using the mass spectrometer as a sensor was performed. It was found that the superposition theorem was valid when the injected volume was kept constant. The mass spectrometer is promising if the method of the automatic injection is established.Finally, the homepage for broadcasting an odor over Internet has been made. Anyone can access that homepage to sniff the smell recorded at the remote site. We succeeded in regenerating 2-component odor at the remote site using the sensor array.