一种极低载噪比高动态信号跟踪的线性预测器设计

针对深空通信高动态低载噪比微弱信号跟踪精度低、稳态均方性能差的问题,提出了一种高动态低载噪比信号跟踪的分级异构凸组合线性预测器算法,该算法将归一化最小均方算法(NLMS)和最小二乘算法(RLS)进行分级异构凸组合,给出了算法的结构、跟踪精度联合优化设计方案和最优参数选择的方法。在载噪比(CNR)为10~30dB-Hz、载波多普勒频移为-300~300kHz、多普勒频率变化率为-800~1200Hz/s、码速率为20bps条件下对分级异构凸组合线性预测器跟踪性能及跟踪精度指标(归一化均方根误差)进行了仿真。理论分析和仿真结果表明,该线性预测器与分级同构凸组合线性预测器相比,算法运算量减少、信号跟踪精度提高、收敛速度加快。因此,该新型结构线性预测有着良好的应用前景。

The Design of an Adaptive Linear Predictor for the Tracking of Extremely Low CNR High dynamic Signals

To solve the low tracking precision and the poor steady state mean square performance of high-dynamic low-CNR weak signals in deep space communications, a convex combination of hierarchical heterogeneous linear predictor is proposed. The structure and the union design of optimal tracking accuracy are given, and the key parameters selection method of the linear predictor is presented accordingly. The linear predictor uses a hierarchical structure with the combination of different families filters, such as the normalized least-mean-squares (NLMS) algorithm and the recursive-least squares (RLS) algorithm. Simulations of the linear predictor tracking performance and tracking accuracy indexes are performed when Carrier-Noise-Ratio (CNR) varies from 10 to 30dB-Hz, Doppler frequency ranges from -300 to 300kHz, Doppler rate varies from -800 to 1200Hz/s and date rate is 20bps. Both theoretical analysis and simulation results indicate that, compared with the convex combination of hierarchical isomorphism linear predictor, amount of calculation is reduced, signal tracking accuracy is improved and convergence rate is accelerated, so that the linear predictor with new structure has prospective application potential．