本文设计了一种以单层或多层前馈神经网络为核心的组合自适应滤波器。该滤波器能够通过离线及在线学习了解输入信号与噪声的某些特性，从而把两者区别开来，从中检出信号。仿真研究表明，本文提出的方法直接用于雷达导引头角跟踪系统的滤波比通常的方法具有更好的性能和效果。 In this paper, we design a combined adaptive filter with single-layer or multi-layer feedforward neural network as the core. The filter can learn some of the characteristics of the input signal and noise through offline and online learning, so as to distinguish the two, from which the signal is detected. Simulation results show that the method proposed in this paper has better performance and efficiency than the usual method, which is directly applied to the filtering of radar seeker angle tracking system.