A Mixed Gas Recognition System Based On Embedded Chip

With the improvement of the speed and processing capability of integrated circuits, it is feasible to apply the embedded chips into mixed gas recognition systems. After a brief introduction of the characteristics of classic gas recognition technology and embedded systems, this dissertation reviewed the current status of the gas recognition technology and analyzed the feasibility of applying this technology into embedded systems. According to their applying characteristics, a gas recognition system based on wireless transmission was put forward.The gas recognition system presented in this dissertation includes data acquisition module, wireless transmission module, recognition module and data storage module. The data acquisition module converts the outputs of gas sensor into digital values, which are transmitted to the recognition module via wireless transmission module. Then the trained recognition module can identify target gas as the known gas or unknown gas based on the current data. Due to the separation between data acquisition module and recognition module, the system is suitable to be applied in harsh or dangerous conditions as well as common conditions.MCU which integrates an 8-channel A/D converter is used in this system for data acquisition and transmission. A ARM processor based on ARM7TDMI is used for the training of neural networks and gas recognition.Finally, The software for acquisition module and recognition module was developed. To improve the wireless transmission efficiency, Some characters was extracted in acquisition module. In addition, the software modifies the BP neural-network to improved its disadvantage, such as convergent speed too slow, easily oscillated.