Robust algorithm in distributed estimation fusion with correlation of local estimates

In distributed estimation fusion, locally obtained estimates are transmitted to the central processor via noisy channels. Traditionally, optimal linear methods are applied to solve the fusion problem under Gaussian noise assumption that can be severely violated in practise when channel noises are heavy-tailed. Hence, those methods should be replaced by robust analogs. M-estimates are well-known robust tools; however, when there is considerable correlation between local estimates, fusion accuracy may decrease. Thus, we propose a robust fusion algorithm based on a procedure for trimming outliers and the subsequent application of an optimal fusion method. Numerical experiments show that the proposed method is more accurate than conventional M-estimates, especially when there is a high degree of correlation involved.