Research And Design On Embedded Platform For Odor Analysis

Smells,sounds and images are the three main media for information exchange between human and material components.Odor,as an important feature of the components,is closely related with the variety and quality of the material,so it can be used to identify and classify the components by means of odor analysis.Traditional odor analysis by smelling usually needs to rely o n professional person to carry on the assessment.This type of discrimination method has some limitations,and the assessment result is always subjective and unreliable.In order to overcome the difficulties with human physiological and sensory evaluation of material quality,an alternative use of chemical method has been used(e.g.gas chromatography)to analyze the odor.However,the chemical detection method is normally resulted in destruction of the samples and it would be difficult to assess the overal l properties of materials.Recently,with the advancement of electronic nose technology,the material odor can be analyzed and detected more accurately.However,due to the complexity and high cost of such electronic devices,there is room for more research and development for better odor analysis platforms.Most electronic noses today can detect odor but fail in accurate odor analysis and classification task.An embedded platform can be designed and employed for the integration of electronic nose data detection and embedded platform data analysis,which is of great significance to the development of electronic nose as applied to the identification of material components.The odor analysis platform based on embedded platform and electronic nose for odor connecting can be used to analyze and process the data collected intelligently and identify mix odor of much components,which has broader area of applications.Its main programming language is the C language.The hardware of the platform adopts LCD screen of ARM9 as the result print tool,and the sofeware is the Linux operation system and improved analysis algorithm.The odor analysis platform is constituted by four modules of the ARM9 core controller,the Linux operating system,smell analysis algorithm,and LCD display.The realization of odor analysis platform requires the data processing of Electronic Nose function to be implemented in an embedded chip.The relevant odor data collected by the electronic nose will be transferred via the serial port of an e mbedded microcontroller to realize the analysis of data and perform further classification of the processed odor.An algorithm was developed and implemented into the embedded chip to analyze,process and reduce the dimension of the odor data and display th e classification results on the LCD screen of the embedded platform.Finally,we adopt the “yujin”,“duhuo”,“yizhi” and “cangzhu” as the samples to acquire their data.The result of odor analysis by the embedded system was compared with the same algorithm running in Matlab to verify the accuracy.The experiment shows that the platform can analyze the data of different odor types and identify different types of odor.