摘要
针对机动目标跟踪问题,提出了一种IMM-RDCKF算法。首先充分利用量测方程中只有部分状态变量是非线性的特点,对于非线性的量测方程采用降维滤波方法,可以在保障跟踪精度条件下减小计算量。其次,对IMM算法中的转移概率矩阵进行实时估计,提高了模型匹配概率。再次,滤波过程中由于误差累积可能导致协方差矩阵失去正定性,对算法进行了优化,确保了滤波过程中协方差矩阵的正定性,提高了算法稳定性。Monte-Carlo仿真结果表明,与CKF算法相比,该算法的跟踪精度有明显的提高,计算效率提高了一倍。
A tracking algorithm based on IMM-RDCKF is proposed to solve the maneuvering target tracking.Firstly,the reduced dimension cubature Kalman filter is used in the measurement equation for the characteristics that only part of the state variables are nonlinear,which can decrease the amount of calculation while ensure the tracking accuracy.Secondly,the transfer probability matrix in IMM algorithm is estimated in real-time to improve the model matching probability.At last,the error accumulation in the filtering process may lead to the loss of positive definite property of the covariance matrix,so the proposed algorithm is optimized to ensure positive definite property of the covariance matrix in the filtering process,which can improve the stability of the algorithm.Compared with CKF algorithm,the Monte-Carlo simulation results show that the algorithm improves the tracking accuracy,and the computational efficiency is twice as large as the CKF algorithm.
作者
周昆正
ZHOU Kunzheng(The 20th Research Institute of China Electronics Technology Group Corporation,Xi’an 710068,China)
出处
《雷达科学与技术》
北大核心
2018年第6期656-660,666,共6页
Radar Science and Technology
关键词
机动目标跟踪
交互多模型
时变Markov转移概率
降维容积卡尔曼滤波
协方差矩阵正定
maneuvering target tracking
interacting multiple model(IMM)
time-varying Markov transition probability
reduced dimension cubature Kalman filter
positive definite property of covariance matrix
