外测系统中传统的正交多项式微分方法和自然样条微分方法容错能力弱,无法避免截断误差、消除异常数据的影响,并且求得的微分结果对随机误差敏感。为了解决这些问题,提出容错样条微分技术,融合了自然样条微分算法与正交多项式中心平滑的优点,不仅增强了算法的容错性能、降低了结果对随机误差的敏感性,并且减小了截断误差。通过仿真试验,验证了该算法的实用性和有效性,为试验任务提供了高精度的弹道参数结果。
Traditionally, orthogonal polynomial and natural spline differential algorithms are used in data processing for trajectory measurement systems. But the algorithms have no ways to overcome the drawbacks of truncation error, weak fault tolerance and outliers or random-error influence. The paper proposes a fault tolerance spline differential algorithm to solve or minimize such problems. The new method has the advantages of natural spline, strong fault tolerance, weak sensitivity to random-error and less truncation error. Simulation and tests have verified the correctness and practicality of the new algorithm and it has provided high-accuracy trajectory data.