无源雷达系统监测通道中含有直达波干扰,导致系统无法检测到目标。基于传统频域块最小均方(frequency-domain block least mean square,FBLMS)的直达波干扰抑制算法,无法同时兼顾收敛速度与稳态误差,影响了直达波干扰抑制效果和目标回...无源雷达系统监测通道中含有直达波干扰,导致系统无法检测到目标。基于传统频域块最小均方(frequency-domain block least mean square,FBLMS)的直达波干扰抑制算法,无法同时兼顾收敛速度与稳态误差,影响了直达波干扰抑制效果和目标回波信噪比。针对这一问题,提出一种改进的变步长FBLMS直达波干扰抑制算法。算法通过引入箕舌线函数建立步长因子与误差之间的非线性关系,同时引入变参数函数替代箕舌线函数中的常值参数协同控制步长因子。实验结果表明,与定步长FBLMS算法和基于箕舌线的FBLMS算法相比,改进算法具有更快的收敛速度,直达波干扰抑制效果更好,目标回波信噪比得到提高。展开更多
A novel method based on zoom fast Fourier transform（FFT） is proposed for minimizing the burden processing of cross-ambiguity functions without affecting performance. The low-pass anti-aliasing filter in zoom FFT is ...A novel method based on zoom fast Fourier transform（FFT） is proposed for minimizing the burden processing of cross-ambiguity functions without affecting performance. The low-pass anti-aliasing filter in zoom FFT is realized by using the multistage filtering technique and the weighting processing is employed in the first stage filter to get rid of the redundancy of the computation. In practical systems, the input data is divided into overlapped data frames to avoid loss of detection ability which results in the rapid increase of computational complexity. A segment technique is also proposed in which CAF calculation of overlapped data frames is viewed as slide window FFT to decrease the computational complexity. The experimental results show that compared to the conventional methods, the proposed method can lower computational complexity and is consistent with the real time implementation in existing high-speed digital processors.展开更多
文摘无源雷达系统监测通道中含有直达波干扰,导致系统无法检测到目标。基于传统频域块最小均方(frequency-domain block least mean square,FBLMS)的直达波干扰抑制算法,无法同时兼顾收敛速度与稳态误差,影响了直达波干扰抑制效果和目标回波信噪比。针对这一问题,提出一种改进的变步长FBLMS直达波干扰抑制算法。算法通过引入箕舌线函数建立步长因子与误差之间的非线性关系,同时引入变参数函数替代箕舌线函数中的常值参数协同控制步长因子。实验结果表明,与定步长FBLMS算法和基于箕舌线的FBLMS算法相比,改进算法具有更快的收敛速度,直达波干扰抑制效果更好,目标回波信噪比得到提高。
基金Sponsored by the Excellent Young Scholar Research Fund of Beijing Institute of Technology (000Y01-5)BIT(UBF 200501F4208.4)
文摘A novel method based on zoom fast Fourier transform（FFT） is proposed for minimizing the burden processing of cross-ambiguity functions without affecting performance. The low-pass anti-aliasing filter in zoom FFT is realized by using the multistage filtering technique and the weighting processing is employed in the first stage filter to get rid of the redundancy of the computation. In practical systems, the input data is divided into overlapped data frames to avoid loss of detection ability which results in the rapid increase of computational complexity. A segment technique is also proposed in which CAF calculation of overlapped data frames is viewed as slide window FFT to decrease the computational complexity. The experimental results show that compared to the conventional methods, the proposed method can lower computational complexity and is consistent with the real time implementation in existing high-speed digital processors.
