中文摘要：

提出了采用每个线性相关通道组共享一个监视通道的简化对称归一化方法，并对标准归一化、对称归一化和简化对称归一化的估计偏差进行了理论分析。理论分析表明，对称归一化和简化对称归一化有相近的估计偏差，标准归一化方法的估计偏差较大。采用三种归一化方法，对90nm，1000nm及90nm与1000nm混合的标准聚苯乙烯乳胶颗粒进行了实验。实验结果表明，标准归一化自相关函数平均估计偏差和最大估计偏差分别是对称归一化和简化对称归一化的2．5～3．5倍和4．5～8．5倍；与对称归一化方法相比，简化对称归一化方法只需增加少量的附加通道资源，即可有效解决在多tau相关结构中存在的大延迟时间的归一化光强自相关函数偏差增大问题。

英文摘要：

A reduced symmetric normalization procedure was presented to reduce the requirement for hardware channels, which needs only one monitor channel in each linearity module. Theory analysis proves that the standard deviation of the reduced symmetric normalization approximates to that of the symmetric normalization and far smaller than that of the standard normalization at large delay time. Measurements with polystyrene particles of 90 nm, 1000 nm and 90 nm mixed with 1000 nm show superiority of both symmetric and reduced symmetric normalization over standard normalization. Calculated average standard deviation of standard normalization estimator is 2.5 to 3.5 times larger than that of the symmetric and reduced symmetric ones while the maximum deviation is 4.5 to 8.5 times. Reduced symmetric normalization achieves high statistical accuracy at the expense of obviously reduced additional monitor channels compared with symmetric normalization in multi-tau photon correlator.