This paper presents the problem of distributed estimation in an incremental network based on the family of normalized subband adaptive algorithms (NSAAs). The distributed NSAA (dNSAA), the distributed selective partial update NSAA (dSPU-NSAA), the distributed dynamic se
More
This paper presents the problem of distributed estimation in an incremental network based on the family of normalized subband adaptive algorithms (NSAAs). The distributed NSAA (dNSAA), the distributed selective partial update NSAA (dSPU-NSAA), the distributed dynamic selection NSAA (dDS-NSAA), and the dSPU-DS-NSAA are introduced in a unified way. The dNSAAs have better convergence speed than distributed normalized least mean square (dNLMS) algorithm especially for colored Gaussian input of the nodes. In comparison with dNSAA, the dSPU-NSAA, and dDS-NSAA have lower computational complexity and close performance to dNSAA. Also by combination of these algorithms, the dSPU-DS-NSAA is established which is computationally efficient. In addition, a unified approach for mean-square performance analysis of each individual node is presented. This approach can be used to establish a performance analysis of classical distributed adaptive algorithms as well. The theoretical expressions for transient, and steady-state performance analysis of the various dNSAAs are introduced. The validity of the theoretical results, and the good performance of these algorithms are demonstrated by several computer simulations.
Manuscript profile
