Acoustic echo cancellation using blind source separation

Abstract

We investigate an acoustic echo cancellation method using a blind source separation technique. In a hands-free mobile system, when the large acoustic noise picked up by the microphone is mixed with the echo, the echo cancellation system shows poor performance. We use a blind source separation algorithm to control the acoustic noise which has to be suppressed. The blind source separation algorithm separates the original echo signal from the observation signal which is mixed with acoustic noise. The separated echo signal is applied to the reference signal of the adaptive filter; this approach gives good performance of the adaptive algorithm for echo cancellation. Computer simulation results show the efficiency of the suggested method.