Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Double-Proportionate Uncertainty-Aware Diffusion Algorithm for Distributed Estimation
Ist Teil von
IEEE transactions on circuits and systems. II, Express briefs, 2024-03, Vol.71 (3), p.1-1
Ort / Verlag
New York: IEEE
Erscheinungsjahr
2024
Quelle
IEEE/IET Electronic Library (IEL)
Beschreibungen/Notizen
In this paper, we consider an uncertainty in the linear measurement model of the distributed estimation problem. To deal with this uncertainty, the uncertainty vector is estimated along with the unknown vector itself. Hence, the cost function is assumed to be dependent on both unknown vector and uncertainty vector and these two vectors are separately estimated in an iterative manner. First, we update the unknown vector assuming uncertainty vector is known and then, we update the uncertainty vector assuming the unknown vector is given. Then, to achieve lower error, we consider the double proportionate scheme in the uncertainy-aware algorithm. Two diagonal gain matrices are obtained mathematically with one degree of freedom, i.e., one gain matrix is obtained in closed-form assuming the other gain matrix is known. Simulation results demonstrate superior performance of the proposed algorithms as compared to the classical diffusion least mean square (LMS) algorithm and achieve near the performance of a case without uncertainty under extreme noise conditions.