A survey of adaptive analog filters

Adaptive systems, a subject of active research during the last several years, have been mainly implemented in a digital way. However the analog realization forms may present several advantages over the digital ones in applications requiring high convergence speed with lower power consumption, and small sizes. This paper proposes analog adaptive filters based on gradient search based and modified Hopfield networks adaptation methods. Proposed systems provide solution to some of the problems still exiting in the analog adaptive filtering field such as: the distortion due to imperfect integration, increasing the convergence speed of the analog system and the development of analog adaptive system with low sensitivity to additive noise without requiring an explicit matrix inversion. Computer simulation results show that the proposed systems present much better convergence performance than conventional ones in several practical applications