多渐消因子平方根容积卡尔曼滤波算法

Multiple Fading Factors Strong Tracking Square Root Cubature Kalman Filter

摘要:

针对平方根容积卡尔曼滤波（square root cubature kalman filter，SCKF）在系统模型不准确和状态突变情况下鲁棒性差的问题，提出了一种多渐消因子平方根容积卡尔曼滤波算法（multiple fading factors strong tracking SCKF, MSTSCKF）。MSTSCKF引入强跟踪思想，通过多渐消因子实时调整增益矩阵，建立多渐消因子数值求解方法，克服多渐消因子求解依赖先验知识的不足；采用假设检验理论对系统异常进行检测，降低误判概率，提高滤波稳定性。通过仿真分析，比较了SCKF、单渐消因子平方根容积卡尔曼滤波（single strong tracking SCKF，STSCKF）和MSTSCKF的算法性能，实验表明MSTSCKF具有更好的跟踪精度和鲁棒性。

Abstract:

In order to solve the problem of poor robustness in the case of inaccurate system models and abrupt state transitions for square root cubature kalman filter (SCKF), a multiple fading factors strong tracking SCKF (MSTSCKF) is proposed. Combining with strong tracking filter idea, MSTSCKF adjusts the gain matrix in real-time through multiple fading factors, establishes a multiple fading factors numerical solution method and overcomes the dependent on prior knowledge. Meanwhile, MSTSCKF uses hypothesis testing theory to detect system anomalies, reduces the probability of misjudgment and improves filter stability. At last, SCKF、single strong tracking SCKF(STSCKF) and MSTSCKF is compared in numerical simulation experiments under different conditions. The simulation results declare that MSTSCKF has better performance on tracking accuracy and robustness.