The method of periodic partial updates is evaluated. Cost, performance, portability and physical size considerations compel serious resource limitations on adaptive signal processing systems. Partial-Selective coefficient updates techniques can be used to reduce the resource utilization and hardware complexity in practical applications at the probable expense of higher steady-state MSE error and lower convergence speed. Periodic partial update is one of partial-selective coefficient update techniques by which, instead of updating the whole coefficients, a subset of the coefficients is updated at each iteration. The performance, convergence speed and MSE error of periodic partial updates is evaluated in the presence of white and colored Gaussian input. It is concluded that in some practical applications, this method could be used instead of full-update algorithm with some penalties in quality.

Login Required

Please login to access the full PDF of this article.

🔑 Author Login