相关量测噪声情况下多传感器集中式融合跟踪

在实际的多传感器融合目标跟踪中,各传感器之间的量测噪声往往是相关的。对于量测噪声相关的多传感器量测,利用Cholesky分解和单位下三角阵的求逆方法,将其转化为量测噪声互不相关的等价的多传感器伪量测,然后基于Kalman滤波,提出了一种解决量测噪声相关情况下多传感器融合目标跟踪问题的新算法。与已有的和直接利用原始传感器量测的集中式融合算法相比,三者在计算精度上完全等价,但新算法的计算复杂度却大大降低。数值仿真实验进一步验证了新算法的有效性。

量测噪声相关条件情况下的多传感器状态融合估计

对于量测噪声相关的多传感器系统,采用Cholesky分解和单位下三角阵的求逆方法,将其转化为量测噪声互不相关的等价模型,从而可以直接利用现有的融合算法进行状态融合,由此给出了量测噪声相关情况下多传感器的状态融合算法。该算法和已有的算法相比具有两大优点,该算法不仅考虑了量测噪声相关而且简单可行,而且该算法和已有的算法相比更具有一般性,对于有特征值相同的情形同样适用。数值仿真结果表明了该算法的有效性。

Optimal filtering for uncertain systems with stochastic nonlinearities, correlated noises and missing measurements

The globally optimal recursive filtering problem is studied for a class of systems with random parameter matrices,stochastic nonlinearities, correlated noises and missing measurements. The stochastic nonlinearities are presented in the system model to reflect multiplicative random disturbances, and the additive noises, process noise and measurement noise, are assumed to be one-step autocorrelated as well as two-step cross-correlated.A series of random variables is introduced as the missing rates governing the intermittent measurement losses caused by unfavorable network conditions. The aim of the addressed filtering problem is to design an optimal recursive filter for the uncertain systems based on an innovation approach such that the filtering error is globally minimized at each sampling time. A numerical simulation example is provided to illustrate the effectiveness and applicability of the proposed algorithm.