针对非数据辅助前向反馈(NDA—FF)算法在低滚降系数的成形与匹配滤波器的条件下定时误差大,并且可能出现周期跳转的现象,提出了一种改进的NDA—FF符号定时恢复算法。在对平方定时恢复器的结果进行Kalman滤波的整体结构上,利用两次锯齿波调整,使得估计的延时结果可以直接用于插值。仿真结果表明该算法基本消除了可能出现的周期跳转现象,有效地改善了定时跟踪的性能。
In view of that the Non Data-aided Feedforward (NDA-FF) timing recovery algorithm produces high estimation variance, anu the phenomenon of cycle slip when the rolloff factors of the pulse shaping filter and the matched filter are low, an improved NDA-FF timing recovery algorithm is proposed in this paper, which bases on the square timing recovery and Kalman filter. Two sawtooth functions are added, then the output can be directly used to interpolate. Simulating results show that this algorithm basically eliminates the phenomenon of cycle slip and efficiently improves the tracking performance.