The Research Of Indoor Comfort Evaluation Methods Based On Fuzzy Data Fusion

Indoor comfort evaluation is a precondition for improving indoor environment. At present, there are various ways to evaluate indoor comfort and it can mainly be attributed two major categories: applying comfort index to formulate directly and processing the measured data by comprehensive methods. Multi-sensor data fusion is a rising technology in recent years, it's a reliable method for information processing and decision making system. This paper discusses the feasibility of applying fuzzy logic and neural network into the evaluation of indoor comfort and the improved solution for it based on analysis of indoor environment characteristics and evaluation methods deeply to overcome the defects of the index complicated and dynamic change in the evaluating process.Fuzzy comprehensive evaluation which possesses the ability of synthetic decision make is a more mature and effective method in the area of quality evaluation, but the comfort index weights usually given by people or some experts which usually lead to subjective one-sidedness easily. To avoid this, in the comprehensive method this paper focused on improving index weights, using the improved analytic hierarchy process and gray relational analysis method to evaluate the subjective weights and objective weights respectively and finally establish an evaluation model based on combined weights. Considered the advantages and disadvantages of the fuzzy logic system and artificial neural network, and then combine them organically to establish a fuzzy neural network evaluation model , neural network has the abilities of self-learning and self-organization and could be used to adjust optimally of networks parameters.At last, Collecting indoor environment index of temperature, relative humid and illuminance under HKCAN system, The result of simulation shows that fuzzy data fusion is suitable for indoor environment comfort evaluation and it's more reliability and effectively.