本文利用信号的循环平稳性研究分离频谱重叠信号的频移滤波器。通过对循环平稳性和频谱相关性的分析,设计了一种频移滤波器,并采用变步长LMS(Least Mean Square)自适应算法计算多个频移分量的最优权值系数。仿真结果表明,该算法可有效分离频谱重叠信号,效果远好于传统滤波器。最后,在DSP平台上实现了设计的频移滤波器。
In the scenario of multiple wireless personal area networks （WPANs） accessing the home area network （HAN）, the frequency-shift （FRESH）filter based on the cyclostationary theory is applied for the anti-interference at the 2. 4 GHz spectrum. The main architecture of multiple WPANs accessing the HAN is proposed. The medium access control（ MAC） -level coordination solution applied in the access point （AP）for the coexistence of different communication protocols within WPAN is discussed. The diagram of the adaptive FRESH filter is described. The anti-interference models of the FRESH filter in the scenario of multiple WPANs accessing the HAN are proposed. The minimum mean square error （MMSE）convergence property of the FRESH filter with the change in the data point number is analyzed. The simulation results indicate that the FRESH filter can effectively extract the signal of interest （SOI） from the interference with the partly overlapped spectrum. Thus, the excellent anti-interference performance of the FRESH filter is validated. Moreover, by both theoretical analysis and simulation, the MMSE convergent property of the FRESH filter is also proven.
