State Fusion Estimation for Asynchronous Multirate Multisensor Systems with Heavy-tailed Noises

This study investigates fusion estimation towards an asynchronous in-multisensor multirate dynamic system with unreliable observations and heavy-tailed noises. Our dynamic system at most detailed level is known. Numerous sensors independently observe a same target, each having distinct sample frequencies, while the observations are unreliable and asynchronously acquired. Those measurement noises and system noise are both non-Gaussian yet heavy-tailed. The sequential fusion estimation approach is formulated. As an efficient extension of traditionary Kalman filter-grounded sequential fusion method with Gaussian noises, both theoretical analysis and experimental evaluations demonstrate the efficacy of our derived approach.