Recursive distributed fusion estimation for multi-sensor systems with missing measurements, multiple random transmission delays and packet losses

This paper is concerned with the recursive distributed fusion estimation problems for networked multi-sensor systems with missing measurements, multiple random transmission delays, and packet losses. There exist missing measurements in the observation equations due to unpredictable sensor faults. Moreover, there are often random delays and losses during data transmissions from sensors to the fusion center due to limited communication bandwidths of the network. The Kalman-like recursive distributed fusion predictor and filter in the linear unbiased minimum variance (LUMV) sense are, respectively presented based on local estimates, cross-covariance matrices between local estimates, and cross-covariance matrices between the prior fusion estimate and local estimates. The stability and steady-state property of the proposed algorithms are analyzed. Their estimation accuracy is better than that of local estimates and distributed fusion estimates by matrix-weighting local estimates. A simulation example shows the effectiveness of the proposed algorithms.