Font Size: a A A

The Disquisition Of The Diabetes Diagnosis Based On Artificial Olfactory System

Posted on:2007-01-20Degree:MasterType:Thesis
Country:ChinaCandidate:X J MaFull Text:PDF
GTID:2132360182496778Subject:Biomedical engineering
Abstract/Summary:PDF Full Text Request
The ANN technology has got great development since 1940's,ithas been used in the system control, system detecting field widely.Artificial olfactory system is a aptitude technology combining ANNand sensor technology. In cause of the speed of reaction , detectingaccuracy and the facility of use , it is much better than the traditionalailment diagnose system. So the artificial olfactory has a wide marketif the ailment diagnose field.According to the requirement of clinic application, the systemdesign research put the importance on the follow points:1,The ANN has strong nonlinearity data processing ability, andhas advantage in the sensor system whose data has serious nonlinearity.The BP model is the most popular net model, but the BP net has its'own defect, for the purpose of getting rid of the affect of this kind ofdefect, the system designed in this article use a modified BParithmetic-PBF. The test result show that the identify ability to the typeof target gas and consistence of the target gas can get to 100% and97.5%, this result indicates that the system design has get the systemdesign purpose.2,cause the gas sensor's sensibility is affected by its' worktemperature greatly, so to get a more accuracy control result, theaccuracy temperature is necessary. In the old artificial olfactorysystem design, the sensor's work temperature is not much considered,so the whole system's accuracy is much affected for the above reason.In this system, I design a temperature control system using NTC andTEC to get a proper sensor work temperature. The system testingresult show that the use of temperature control system develop thewhole system's accuray. The temperature system's control accuracycan get ±4℃。3,The communication system between MCU and PCNow the system require the MCU's ability more and morepowerful, although there are a lot kinds of MCU in the market, andthese new MCU is more powerful than AT89C51, but it is not enoughfor the use of processing complex data. Because the ANN need a lot ofdata to train and learn, at the same time to develop the accuracy ofsystem, this system use AT89C51 as assistant processor and use PC asmain processor. The AT89C51 has the advantage of little bulk, agilityuse, so it is widely used in the field of control system. In this system,the AT89C51 take on the task of data acquirement, temperature control,keyboard input and display and communication with PC. The PC hasvery strong ability to process complex data, so the PC takes on the taskof train ANN and process ANN data. In this system, the RS232 serialport communication model is used, RS232 serial port communicationtechnology is a very popular way to achieve communication betweenMCU and PC. The system test result proves that the AT89C51 and PCcan communicate properly. To make the system work more efficiently,the task of these two parts needs to be a bit more optimized.4,Serial system designTo save the limited hardware resource, reduce the cost ofhardware, the serial system design is very popular. In the system, whichdoes not need very high speed and the hardware, resource is not veryrich, the serial system's design is necessary. In this system, the systemdoes not require very high speed and the MCU's I/O resource is notrich because the MCU need to control data acquirement system, inputand display system, temperature control system, so there are a lot ofserial part in this system what save a lot of I/O resource and make theMCU affording complex task come true.5,In the system work circumstance, there is a lot of interfere aroundthe system and this interference can make the system unstable , so inthis system ,the hardware and software disturb-free system is used, thesystem's test result indicate that the disturb-free system make thewhole system work more stable and satisfied the requirements of thesystem as a clinic diagnose device.
Keywords/Search Tags:sensor array, ANN, gas detection, temperature control, MCU
PDF Full Text Request
Related items