中文摘要：

人体热舒适度模型对于建筑综合节能研究具有重要意义，生理参数能直接、有效地反映人体热舒适变化，是建立人体热舒适度模型的一种有效手段。血压信号是表征人体热反应状态的主要生理参数，对于建立热舒适模型具有重要的价值。为了提高模型的准确性，通过分析血压信号非稳态的特点并且结合应用背景，提出了具有任意多尺度分解特性的小波包方法进行特征提取和不易受干扰的能量作为信号特征用以构建模型。通过实验与传统特征提取与选择方法进行比较，验证了该方法的有效性。

英文摘要：

Physiological parameters can reflect the change of human thermal comfort directly and effec-tively .Establishing a mapping model of physiological parameters and thermal comfort is significant to building energy -saving in comfort indoor circumstances .Blood pressure is an important and useful physiological parameter reflecting human thermal response status .In order to improve the accuracy of the model ,by analyzing the unsteady characteristic of blood pressure signal and combined application background ,proposed wavelet packet method with arbitrary decomposition of multi-scale and energy feature w hich is less susceptible in the progress of model establishing as the signal characteristics to build the model .Through experiments compared with conventional feature extraction and selection methods to verify the effectiveness of new method .