The Research Based On Neural Network In Measurement And Assessment Of Atmosphere Quality

With the development of social economy and modern industry, the environment of city becomes more and more serious. We must give much attention to the pollution of environment if we keep carrying out substantial development policy. In the last years, the air pollution is one of problems, which are motioned by more and more people. In order to control air pollution efficiently, we should make scientific assessment of air quality. So the research on this aspect is very important which we study the current situation of air pollution objectively and accurately in city and predict the development trend, it also controls the air pollution efficiently to make the strategy of the sustainable development.This paper is about 'the research based on neural network inmeasurement and evaluation of atmosphere quality'. By studying of multi-sensor data fusion, genetic algorithm (GA), fuzzy logic rule, how to get rule and RBF neural network, the paper includes the two main topics as below.(1) Multi-sensor data fusion is a rising technology in the recent years. It brings about advanced and reliable method for the problem of information processing and decision system in the information age. The paper discusses the feasibility of applying artificial neural network (ANN) into the data processing and the improved solution for it, demonstrates the advantage of the combination of ANN and GA and realizes the computer simulation of the combination of ANN and GA. By analyzing the simulation results, obtained the conclusion that the GA & ANN is better than traditional ANN. Thus constitutes a measurement model of atmosphere quality based on genetic neural network.(2) In recent years the assessment of atmosphere quality has got the extensive attention in international and domestic countries. Though there are various methods of atmosphere quality evaluation, how to reflect the results accurately and objectively,which affected by many factors in total environment, is related to reasonable choice and establishment of assessment method and model. The theory of neural network is an advanced filed in artificial intelligence and a kind of important reasoning method. The paper combines the advantage and disadvantage of fuzzy system and neural network with the characteristic of RBF neural network, designs fuzzy RBF neural network by constitution of fuzzy system with neural network. The application of this simulative system to assessment of atmosphere quality not only considers various uncertain factors, but also resolves the problem more efficiently from the view that the evaluation of atmosphere quality is essentially a kind of mode identification.By many simulation experiments, which are based on collected atmospheric monitoring data and correlative information in the north industry area of Tai Yuan, obtains a model for measurement and assessment based on neural network. The result proves this method is objective, scientific, simple and accurate compared with other assessment methods. The conclusion is close to the reality. It is applied successfully to analyze and evaluate theatmosphere environment quality. This creates best foundation for the realization of intelligent and real time system in measurement and assessment of atmosphere quality.