基于自适应系统辨识的收发隔离技术研究

A Study on Transmitter-receiver Isolation Based on Adaptive System Identification

收发隔离是机载干扰机不可避免的难题。如果收发隔离问题解决不好,轻则削弱干扰机效率,重则造成自发自收,形成自激励。固定步长的归一化最小均方误差(NLMS)算法在解决基于自适应系统辨识的收发隔离的问题时,由于精度不够,隔离效果很不理想。针对此问题提出一种基于先验误差的变步长NLMS算法,该算法依据相邻时刻先验误差的相关系数改变步长因子,改变后的步长因子能够在算法收敛过程中削弱噪声的影响,提高算法精度。理论分析和仿真结果证明:基于文中的变步长NLMS算法的收发隔离方案与基于其他最小均方误差算法的隔离方案相比,隔离性能有较大的改善。

Transmitter-receiver isolation is a inevitable problem for airborne jammer. Lower efficiency of jammer and even self-exci- tation might be caused by lower degree of transmitter-receiver isolation. Transmitter-receiver isolation is not satisfactory when the fixed step size normalized least mean square （ NLMS } algorithm is implemented because of worse estimation accuracy. To solve this problem, a variable step-size NLMS algorithm based on prior estimation error is proposed in this paper. The correlation coefficient of the adjacent moment prior estimation error is employed in this NLMS algorithm to control the step size adaptively. The estimation accuracy as well as the isolation performance is improved by weakening the impact of syetem noise in this step size update scheme. The theory analysis and simulation results demonstrate that the performance of transmitter-receiver scheme based on the proposed variable step-size NLMS algorithm is better than schemes which are based on other least mean square algorithms.