中文摘要：

针对非合作目标跟踪问题,为解决无线传感器网络有限带宽和相关噪声造成的精度影响,在集中式融合框架下提出了三种基于量化信息的目标跟踪算法。首先,局部传感器节点采用自适应的量化策略将观测值量化成消息,并发送到融合中心;然后,融合中心利用状态方程恒等变换和Cholesky分解技术解除任意噪声的相关性;最后,引入强跟踪滤波技术、矩阵求逆引理和顺序滤波技术设计融合方法。几个仿真实验表明,三种新方法的估计精度完全等价,新算法还具备应对目标状态突变等不确定因素的能力,增强了算法的鲁棒性。

英文摘要：

For non-cooperative target tracking problems, this paper put forward three algorithms based on quantized informa- tion in centralized fusion framework, in order to solve the problem of poor tracking accuracy which was caused by the limited bandwidth and correlated noises in wireless sensor networks. First of all, each local sensor adopted an adaptive quantization strategy to quantize its observations into a bits message, and sent them to the fusion center （ FC ）. Then, it performed arbitrary noises decorrelations by using Cholesky factorization technology and identical transformation of state equation. Subsequently, it introduced strong tracking filtering, matrix inversion lemma and sequential filtering technique to design fusion algorithms. Final- ly, several computer simulation experiments show that three methods possess completely equivalent estimation accuracy. In ad- dition, the proposed algorithms have the capability to deal with uncertainly factors such as sudden changes in target state and thus the robustness of these methods is enhanced.