不等维数观测向量的复合量测融合

为了降低观测向量维数不等时进行融合估计的计算量，分析了扩维滤波法在构造全局观测空间时导致很大计算量的缺陷，提出按照观测数据的实际物理意义来构造维数较低的全局观测空间，将维数不等的各局部观测向量扩展成全局观测空间中的等效表示形式，使其可应用复合量测滤波法进行融合的方法，以克服复合量测滤波法要求观测向量等维的局限，并降低不等维数观测向量融合时的计算量。同时，还给出相关条件下多传感器最小方差融合估计的递推计算公式。仿真实验结果表明，与扩维滤波法相比，本融合估计精度更高，且计算量更低。

Combined Measurement Fusion of Unequal Dimensional Measurement Vectors

To reduce calculation cost of fusing unequal dimensions measurement vectors,the defect of augmented method which causes high calculation cost was analysed,and a method was presnted which constructs whole measurement space with lower dimensions by physical meanings of measurement data,extends local measurement vectors of unequal dimensions to equivalent in whole measurement space to be fused using combined measurement filtering method. By this method,the limit of combined measurement filtering method was overcomed and the calculation cost of fusing unequal dimensions measurement vectors was reduced.The recursion formula of multisensor minimize variance fusion with correlation was presented too. Simulation results showed that this method has higher fusion accuracy and lower calculation cost than augmented method.