有色量测噪声下带广义未知扰动的递推上限滤波

针对博弈对抗环境下利用快速采样雷达进行非合作目标跟踪带来的有色噪声和未知干扰共存问题, 本文提出有色量测噪声下带广义未知扰动的随机动态系统递推上限滤波. 这里, 有色量测噪声用于描述由于快速采样或持续干扰带来的噪声相关性, 广义未知扰动用于建模博弈对抗对雷达观测带来的异常影响(先验信息缺失). 针对所考虑系统, 通过参数优化实现状态估计误差协方差上限(而不是理论值)的在线递推, 提出有色噪声下上限滤波(CU-BF), 给出状态估计误差协方差最小上限的近似实现, 讨论了所提CUBF的存在性条件. 在具有时变未知扰动和有色量测噪声的目标跟踪仿真中验证了所提方法的有效性.

Recursive upper bound filtering with generalized unknown disturbances in the presence of colored measurement noises

For dealing with co-existence of colored noise and unknown interference when tracking the non-cooperative target with the high-speed sampling radar in gaming confrontation, a recursive upper bound filter for stochastic dynamic systems with generalized unknown disturbances in the presence of colored measurement noises is proposed. Here, the colored noise is utilized to depict the correlation at adjacent epochs caused by high-speed sampling or persistent pertur-bation, while the generalized unknown disturbance is adopted to model these abnormal effects without prior knowledge caused by gaming. By calculating the upper bound of the state estimate error covariance instead of the theoretical one via parameter optimization recursively, the upper bound filter with colored noises, abbreviated as CUBF, is proposed, and the corresponding achievable minimum upper bound filter with colored measurement noises is further put forward. Meanwhile, the existence condition of the proposed CUBF is also discussed. A target tracking example with time-varying unknown disturbances and colored measurement noises is simulated to validate the proposed method