阈值后处理是针对TD-SCDMA系统中B Steiner信道估计器受加性噪声影响而提出的一种去噪方法,该方法误滤除信息径,保留噪声径.本文提出一种基于阈值处理的多用户信道估计方法,它基于某一下行用户会收到所有其它下行用户的训练序列这一特性,设定一个值,与各用户幅度加权值比较,去掉信噪比很小的用户对应的信道估计窗,将其余用户的信道估计窗进行平均,然后将平均后的结果再进行阈值后处理,抑制由噪声贡献的信道响应.仿真结果表明,提出的方法与直接进行阈值后处理法在同—误码率下,所需平均信噪比(SNR)降低1～ 4dB；在SNR低于5dB时,与复杂度高的加权合并法具有相同的误码率.%Because channel estimator proposed by B Steiner for TD-SCDMA system is mainly affected by additive noise, those methods using threshold processing is likely to filter out the paths containing strong information whereas the path with noise is falsely kept. Based on the characteristics that a specific user can receive the training sequences of all the downlink users, a multi-user channel estimation method based on threshold processing was proposed. By comparing the amplitude weighted values with a given value,the method firstly removed some channel estimation windows whose Signal Noise Ratio(SNR) were too small,(hen averaged the rest of the channel estimation windows, and finally,processed the averaged results with threshold processing method to suppress the channel response contributed by noise. Simulation results show that at the same bit error rate, the average SNR required in the proposed method is 1 ~ 4dB lower than that in the direct threshold processing method;When SNR is lower than 5dB,the proposed method has the same performance as the weighted-unit method with high complexity.
展开▼
