Research on the identification of temperature in intelligent building based on feed forward neural network and particle swarm optimization algorithm

Because the structure of intelligent building is complex and there are many kinds of building equipments in intelligent building, it is difficult to sample parameters about the variety of temperature in intelligent building in real time. To forecast the temperature of least future in observation position accurately, a feed forward neural network with one hidden layer is used as the identification structure for the identification of temperature in this paper and parameters of the identification structure is optimized with particle swarm optimization (PSO) algorithm in this paper too. In our experiment, the number of neurons of input layer and hidden layer of desired neural network are confirmed with BP neural network and experiment results show that the precision and stability of our proposed method are good enough for application based on temperature identification with time requirement satisfied.