Distributed Unknown Input Filtering for Systems with Unknown Inputs

It is well known that, for model-free unknown input filtering (MUIF) the unknown inputs are generally reconstructed with a multi-step delay, which may yield multi-step delayed state and unknown input estimates. This certainly is unpractical for real-time state estimation problems. To solve this problem, the paper first proposes a distributed multi-step-delay state and input estimator (DMSDSIE) for a specific topology of network. Then, a real-time state estimator design of a certain MUIF application based on the DMSDSIE is developed. This research highlights a possibility to apply distributed filtering through a network to alleviate the time-delay estimation problem of a general simultaneous input and state estimation problem.