Adaptive Filter in Hearing Aid System for Impaired People

Most of we humans are gifted with our five senses intact. However an unfortunate few continue to suffer in their own domain. Science has given them gadgets, which may help them to lead a near normal life. Most of these gadgets have shortcomings. For instance, we may consider the hearing aid. A gadget used widely by people who are hard of hearing. Through it improves their hearing perception, unwanted amplification of noise could cause irritation and even damage under extreme conditions. There are noise cancellation algorithms, which currently exist. But they act only upon noises of kind that they are designed for. There is no algorithm currently existing, which will cancel any kind of noise. In this paper we aim at presenting an algorithm, which enhances speech by attenuating any kind of noise. To achieve this we design an adaptive filter, which would adapt itself depending on the nature of noise. we compared with LMS Algorithm and proposed blind technique, since we do not know the parameters of the noise signal. The primary differentiation that we make between speech and noise is that speech is highly non-stationary in a time interval of 250ms whereas noise stationary. This characteristic is used to derive a cost functional using which we may achieve speech enhancement. Adaptation takes place by changing two sets of weights. Apart from hearing aids this algorithm also finds application in mobile phones (Hands free kit). The above algorithms are simulated with the help of mat lab program and to implement using TMS320C6X DSP processor for real time application.