An Intelligent Microphone Array for Speech Enhancement

This paper describes the design of an intelligent microphone array system for speech enhancement. The proposed system consists of a non-linearly spaced array structure followed by a speech seeking unit, a dual-beamformer, a post-processing stage as well as an artificial neural network (ANN) central control unit. The system can seek and track the target speaker automatically and obtain an optimal output with the best quality and intelligibility. Extensive simulation results have shown that the proposed system is a useful device for speech signal acquisition in the presence of competing speech and/or background noise. The application of this technique include hands-free telephone and covert recording systems for forensic application.