一种无平方根运算的QRD-LSL插值算法研究

Research of QRD-LSL interpolation algorithm based on square-root-free

为了降低基于QR分解技术的最小二乘格型(QRD-LSL)插值算法的计算复杂度,提出一种无平方根运算的QRD-LSL插值算法,并将其用于对直扩系统中窄带干扰的抑制。该算法通过将无平方根运算的Givens旋转变换代入原QRD-LSL插值算法,避免了QR分解过程中的求平方根运算,降低了计算复杂度及硬件实施难度,提高了有限精度场合下算法的鲁棒性。仿真结果表明,改进算法较QRD-LSL插值算法具有更小的稳态误差以及更好的数值稳健性,并且保持了良好的NBI抑制效果。

To reduce the complexity of the QR-decomposition-based least-squares lattices（QRD-LSL） interpolation algorithm used in narrow-band interference（NBI） suppression, this paper proposed a square-root-free（SRF） QRD-LSL interpolation algorithm. The proposed interpolation algorithm, which was developed by applying SRF Givens rotation with feedback mechanism, could effectively avoid the square-root computations in QR-decomposition and improve QRD-LSL interpolation algorithm robustness in finite precision environment. The results of computer simulation confirm that the steady-state performance and the numerical robustness of the proposed algorithm outperforms the conventional QRD-LSL interpolation algorithm and holds good performance of NBI suppression in DSSS.