中文摘要：

采用基于匹配追踪（MP）-小波包（WP）原子分解的数据融合算法估计声表面波标签的特征参数．该方法采用MP-WP原子分解方法从包含噪声的接收信号中估计出声表面波标签的脉冲响应，利用估计值和原始值得到估计方差，并依此计算多次接收信号的权重分布．在此基础上，采用基于均方误差的权重数据融合算法融合接收到的标签信号．构建了相位检测系统并对融合方法进行了实验验证．结果表明，在相同的累计次数下，融合方法比算术平均在脉冲峰值和位置上都更接近于真值．

英文摘要：

Based on the matching pursuit （MP）-wavelet packet （WP） atomic decomposition method, a weighted data fusion algorithm which estimates the tag feature parameters of surface acoustic wave （SAW） was presented. First, using the MP-WP atomic decomposition method, noise-free pulse responses of SAW tag reflectors are estimated from its received tag signal. The weighted data fusion algorithm, which is based on minimum mean square error （MMSE）, is then applied to fuse the received signal. The variance of noise in each pulse response is then computed based on the noise pulse response and its estimation. Accord- ingly, the weight of each pulse response for data fusion is calculated using its noise variance. A phase detection system close to the optimal receiver with low signal noise ratio （SNR） was subsequently introduced. The final experimental result indicates that the weighted fusion presents an optimum estimation of the feature parameter of multiple pulse responses of SAW tag, as compared with the arithmetic mean process.