Design And Implementation Of Temperature Compensation System For Go Coated Silicon Bridge Humidity Sensor Based On Fpga

As an ordinary physical, humidity take a very important role in our lives, factory production, farming and other fields. Currently there is a lot of research on humidity material and measurement methods at home and abroad. The laboratory study of graphene oxide found that the surface of graphene oxide material exists hydroxyl groups、carboxyl groups、 epoxy functional groups, which is having a strong hydrophilicity and larger specific surface area, so graphene oxide is the ideal choice of humidity sensitive materials. Semiconductor silicon bridge with high stability, sensitivity and other characteristics, can made into a humidity sensor which coated silicon bridge with graphene oxide to form Feeling Humidity Film. But both of the graphene oxide and semiconductor exists temperature characteristics, such silicon bridge performance of coated graphene oxide is easy changes by temperature affected. Therefore, it need to a variety of methods for temperature compensation.FPGA chip that produced by Altera has been highly integrated resources, in addition to abundant logic resources, it has been integrated a variety of application-specific integrates module, enables to make easier design and shorter develop-cycle; it has also been greatly improved with speed; And with low-cost for FPGA development, so it’s the best choice for a variety of designed applications processors. Radial Basis Function (RBF) neural network has a strong learning ability, and simpler the network structure、fast convergence、enable approximation arbitrary nonlinear function. Therefore, we can use FPGA to mplement RBF neural network to handle the humidity sensor temperature drift problem.In this paper, we take the FPGA chip as the core device, desiged RBF neural network to compensate for the humidity sensor temperature drift. The first, humidity sensor of coated graphene oxide characteristics experimental analysis, design and production of the humidity sensor measuring circuit, collecting humidity characteristics under different constant temperature and humidity sensors in the tank(ie, the relationship of between relative humidity and silicon bridge output voltage) experimental data. In MATLAB, the experimental data is trained repeatedly by using RBF neural network, get network model of2inputs、21neurons、a outputs.On the FPGA to get the RBF neural network, using SOPC Builder to design Hardware architecture implementations, by adding instructions, in NIOS environment using C language to design, This would resolve the problem of using Verilog or VHDL design which will cause more resources-consuming, and the problem of more time-consuming in software simulation. Finally, the system validation results shows that this method is simple to realize, has strong reliability, and will cause lesser resource-consuming. At the same time, it also designed and implemented IP core and read-write programe of A/D and D/A、serial debugger programe、LCD interfaces and display program、temperature read and write program.The analog signal outputted by the GO coated silicon bridge humidity sensor after filtered-amplified-filtered, then be send into the A/D converter chip, Send digital signal that has been converted and the temperature signal that has been collected into the FPGA simultaneously, using the RBF neural network for its temperature compensation, humidity signal that has been compensated would be dispose in two ways:1、using LCD1602to display;2、using the D/A chip converted it into differential analog signal to output. Meanwhile, in order to manually enter and modify the parameters of RBF neural network, also designed the serial circuit, thus it can reduce the hassle of different sensor data processing need to reconfigure the FPGA.Finally, the system of verification proved that the system can achieve the desired design goals.