论文部分内容阅读
针对差分方法和最小二乘法在处理连续波雷达跟踪数据的随机误差时，处理结果常常出现不一致的情况，对这两种方法的内在联系进行了深刻的研究。首次证明了“ｋ阶差分与ｋ＋１个点的ｋ－１次多项式的平滑残差具有完全相同的频谱特性”。针对雷达测量数据中随机误差时序相关性较强的实际情况，导出了相关条件下差分步长的计算公式。最后，提出了高精度分离数据中随机误差的非线性自由节点样条函数方法，该方法经仿真和大量实测数据处理检验是成功的。
In the light of the difference method and the least square method, the processing results often appear inconsistent when dealing with the random error of continuous-wave radar tracking data, and the intrinsic relations between the two methods are deeply studied. It is proved for the first time that “the k-order difference has exactly the same spectral characteristics as the smooth residuals of the k-1 degree polynomials of k + 1 points.” Aiming at the fact that the correlation of random errors in radar measurement data is strong, the formula of differential step under the relevant conditions is derived. Finally, a nonlinear free-spline spline function method for high-precision separation of random errors in data is proposed. The proposed method is proved to be successful by simulation and data processing.