Distributed State Estimation for Multi-Area Data Reconciliation

Victoria Erofeeva, Sergei Parsegov, Pavel Osinenko, Shyam Kamal

Abstract: Data reconciliation is an essential tool in data processing in various industries. It helps to improve accuracy of decision-making algorithms by reducing the influence of random errors in measurements. In this paper, we consider large-scale data reconciliation problems in which multiple areas communicate over a network to obtain an optimal solution of the centralized problem. Our proposed approach accounts for the boundaries between different areas avoiding a mismatch and sub-optimality as well as reduces computational and communication complexities. The proposed distributed data reconciliation method is compared to a centralized reference in different scenarios.