Multitask Simplified Diffusion Orthogonal Gradient Adaptive Algorithm

In this paper, we introduce how to simplify the orthogonal gradient adaptive (OGA) algorithm for distributed network. The diffusion strategies compose with two stages as: adaptation stage and combination stage. Based on multitask criterion, we propose another stages to combine between clusters and the intra-clusters at the combination stage. As concerned with the multitask scheme, the global and local cost functions are defined to investigate with the diffusion strategy. For the fast convergence rate, the OGA algorithm is furnished to diffusion approach with the extending stage of combination between clusters. Simulation results show that the proposed algorithm can examine to confirm competitive results for system identification.