在实际应用中,由于初始偏差、采样速率不同等原因,系统量测一般也非同步。多基地雷达数据通常面临异步数据融合问题。为了解决该问题,该文按照批处理的思路,联合一段时间内的多个异步数据对同一目标状态进行估计,并基于最优贝叶斯估计原理,提出了一种新的批估计数据融合准则;然后,依据该准则推导出了一种解析的分布式批估计方法;最后,针对非线性非高斯场景,提出了一套完备的粒子滤波实现方案。仿真结果表明,文中提出的方法相比现有方法具有跟踪精度高,计算量小等优点。

In practical applications, due to the different initial bias and sample rates of different multi-static radars, usually, the measurements of the system is non-synchronous. Therefore, there exists an annoying asynchronous data fusion problem in multi-radar data fusion. To address this problem, in this paper, we adopt the idea of batch processing, i. e. , utilizing the multiple asynchronous data during a period to estimate the same target state, and propose a batch estimation data fusion rule based on the optimal Bayesian estimation. Then, a distributed batch estimation approach is derived based on this rule. Finally, with regard to the nonlinear and non-Gaussian scenarios, a particle filtering based implementation of distributed batch estimation is derived. Simulation results show that our method has a better performance in tracking accuracy and computation cost comparing with the existing methods.

TN957.52