基于FPGA面向口臭检测的电子鼻系统研究

发布时间：2018-01-31 01:19

本文关键词： 电子鼻传感器阵列神经网络模式识别　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：现实生活中,口臭是经常困扰人们的问题。口臭问题已经严重影响了人们的正常工作和生活。随着人们生活质量的提高,口臭已经引起越来越多的人关注。研究表明,口腔内可挥发性硫化物是口臭的主要来源。获知口腔内可挥发性硫化物的种类和浓度,对明确的诊断及进一步治疗有重要的意义。因此,提出了基于金属氧化物半导体传感器阵列、FPGA、BP人工神经网络的面向口臭检测的电子鼻系统,这种系统可以用于辅助医生诊断口臭病情。本文中使用硫化氢和甲基硫醇与空气充分混合模拟口臭患者的呼气,通过电子鼻系统对模拟的呼气进行定性和定量分析。首先,对典型的电子鼻系统进行研究,一个典型的电子鼻系统一般包括三个部分:传感器阵列,数字采集系统和模式识别算法。根据电子鼻的结构及工作原理,设计了面向口臭检测的电子鼻系统的系统框架及工作流程。系统包括气体采集系统和模式识别系统。其次,气体采集系统包括传感器电路和数据采集电路两个部分。传感器阵列和小信号放大电路被组装在一块印刷电路板上构成传感器电路,被放置在玻璃容器内,在恒温恒湿条件下对气体进行采样。A/D转换电路和通信电路构成的数据采集电路通过电缆与传感器电路相连,完成数据采集。采用UART(universal asynchronous receiver/transmitter)实现RS232通信方式将数据传至上位机。最后,对BP神经网络的工作原理及硬件实现进行了研究。采用MATLAB上位机软件,通过BP神经网络算法对来自气体采集系统的传感器阵列响应信号进行数据处理,分别建立气体与种类和气体与浓度之间的数学模型,然后通过这些模型和待测气体的响应信号,得到待测气体的种类和浓度信息,与实际种类和浓度进行对比,验证系统的准确性及可靠性。通过实验证明,采用气体传感器阵列和软件模拟的模式识别算法的电子鼻系统能够完成硫化氢和甲基硫醇的辨别以及硫化氢单一气体的浓度检测。综上所述,本文研制的基于FPGA的面向口臭检测的电子鼻系统在进行单一气体的定性定量检测方面已经证明有一定的可行性。电子鼻的实用化还需要更多的研究工作,具有很重要的研究意义和广阔的发展前景。
[Abstract]:In real life, halitosis is a problem that often puzzles people. The problem of bad breath has seriously affected people's normal work and life. With the improvement of people's quality of life. Halitosis has attracted more and more attention. Studies have shown that volatile sulfides in oral cavity are the main source of halitosis. Therefore, an electronic nose system for halitosis detection based on metal oxide semiconductor sensor array FPGABP artificial neural network is proposed. This system can be used to assist doctors in the diagnosis of halitosis. In this paper, hydrogen sulfide and methyl mercaptan are used to fully mix with air to simulate the breath of a patient with bad breath. Through the electronic nose system to simulate the qualitative and quantitative analysis of breath. First of all, the typical electronic nose system is studied. A typical electronic nose system generally consists of three parts: sensor array. Digital acquisition system and pattern recognition algorithm. According to the structure and working principle of electronic nose. The system framework and workflow of the electronic nose system for halitosis detection are designed. The system includes gas acquisition system and pattern recognition system. Secondly. The gas acquisition system includes two parts: the sensor circuit and the data acquisition circuit. The sensor array and the small signal amplifier circuit are assembled on a printed circuit board to form a sensor circuit and placed in a glass container. Under the condition of constant temperature and humidity, the data acquisition circuit composed of sampling. A- / D conversion circuit and communication circuit is connected with the sensor circuit through the cable. Complete data acquisition using UART(universal asynchronous receiver / transmitter). RS232 communication mode will be implemented to transmit data to the host computer. Finally. The working principle and hardware realization of BP neural network are studied. MATLAB software is used. The data of sensor array response signal from gas acquisition system are processed by BP neural network algorithm, and the mathematical models between gas and category and gas and concentration are established respectively. Then, through these models and the response signals of the gas to be tested, the information of the kind and concentration of the gas to be tested is obtained, and compared with the actual type and concentration, the accuracy and reliability of the system are verified. The electronic nose system using gas sensor array and pattern recognition algorithm simulated by software can distinguish hydrogen sulfide from methyl mercaptan and detect the concentration of single gas of hydrogen sulfide. The electronic nose system based on FPGA for halitosis detection has been proved to be feasible in the qualitative and quantitative detection of single gas. More research work is needed for the practicality of electronic nose. It has very important research significance and broad development prospect.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R781.9;TP212

【相似文献】