中文摘要：

针对传感器对目标跟踪时观测噪声非高斯问题,提出了一种基于关系矩阵的主、被动传感器量测统计融合算法。采用方差加权距离处理传感器量测噪声非高斯问题,运用传感器综合融合度构建关系矩阵,并且在门限附近采用椭圆模糊处理技术,利用Perron-Frobenius定理计算量测融合过程中每个传感器的权重。仿真结果表明当传感器观测噪声具有非高斯特性时,基于关系矩阵的主被动传感器统计融合算法和传统的融合算法相比扰动较小,具有较好的稳定性,可用于改善跟踪系统的抗干扰能力。

英文摘要：

A measurement statistical fusion algorithm based on relation matrix is proposed when the measurements from sensors do not follow Gaussian distribution in target tracking system. Variance weighed distance is used to solve the problem of non- Gaussian distribution and integrated fusion degree of sensors to build the relation matrix where eclipse curve fuzzy technology is adopted nearby threshold, and weights of each senor is obtained based on Perron-Frobenius theory in data fusion process. Simulation shows that the proposed algorithm is lower perturbation and more stable compared to the traditional fusion algorithm when the measurements from multi-sensor are non-Gaussian distribution, and can enhance the anti-jamming capacity of target tracking system.