Study On Multi-information Fusion Algorithm Of Indoor Environment

With improving of living standard, the requirements on the living environment are growing, too. Indoor environment is related closely to people's daily lives, and its quality will directly affect people's lives quality, even the subsistence of mankind. Therefore, it is significant and necessary to monitor the parameters of indoor environment, evaluate the comfort level, and produce a comfortable, quiet, clean room environment. However, indoor environment is integrated, the current researches on indoor environments are only focused on one aspect of heat, light, sound environment comfort level or air quality, instead of on these factors together. Aiming at this problem, the fusion architecture of NN (Feature Level)+ D-S(Decision Level)is presented, it makes the output of NN as the evidence of D-S, avoiding the complexity and subjectivity of basic probability assignment function selection.Main works are as follows.(1)The evaluation standards and grade divisions of indoor environment comfort level are identified, and it makes a basic evaluation standard for indoor environment comfort level fusion evaluation.(2)The indoor heat environment comfort level fusion evaluation algorithm based on GNN is proposed. Compared with BP, GNN uses GA optimizing the initial weight and threshold of BP, it is evaluated better with high accuracy and fast speed.(3)For the problem of that there are complicated nonlinear relationships among the parameters of heat comfort index PMV, PCA and KPCA are used to do the feature extraction. On the basis, PCA+BP, PCA+GNN, KPCA+BP and KPCA+GNN are utilized to forecast the heat comfort level. Simulation results show that KPCA can extract the nonlinear uncorrelated sample data, and KPCA+GNN is evaluated best with high accuracy.(4)The fusion evaluation algorithm of indoor light, sound environment comfort level and air quality based on FNN is proposed, and FNN and fuzzy evaluation are studied and compared. Simulation results show that FNN can effectively evaluate indoor light, sound environment comfort level and air quality.(5)Using D-S realize decision level fusion evaluation, and compared with fuzzy evaluation. Simulation results show that this fusion architecture of NN + D-S can effectively evaluate indoor environment comfort level.