Block Diffusion Adaptation over Distributed Adaptive Networks under Imperfect Data Transmission

A distributed Block LMS estimate strategy is developed by appealing to collaboration techniques that exploits both space and time structures of data. In diffusion strategies, information are exchanged among the nodes, usually containing noisy links. The weight combination of the neighboring nodes play a crucial role in adaptation and tracking ability of the network. The paper investigates for general adaptive diffusion algorithm, in presence of various sources of imperfect information exchange, then moves on to investigate for BLMS diffusion method, which plays a crucial role in reducing the burden of communication by a factor of block length.