Akustische Echokompensation bei Mikrofonarrays kleiner Apertur

Full-duplex hands-free applications require acoustic interfaces that provide acoustic echo cancellation, local interference and environmental noise suppression. In this work the interaction of the different algorithmic parts of such a system should be considered in depth, in particular the combination of adaptive beamforming and acoustic echo cancellation (AEC). In the first part of this work a mathematical description of diverse algorithms for single-channel acoustic echo cancellation is given and their performance is compared. Furthermore, with the implementation of an individual step-size-control for each filter-tap a higher convergence rate and better tracking capability could be achieved. This shows in particular positive effects for sparse impulse responses and strong time-varying acoustic environments. By simulations in MATLAB, the behaviour of variable step-size-control strategies in different acoustic conditions is evaluated for a single-channel AEC system. In the second part of this work, the interaction of adaptive beamforming, based on the generalized sidelobe canceller (GSC), and algorithms for acoustic echo cancellation is studied. With the integration of the AEC into the structure of the GSC and a joint optimization criterion, the system shows a good behaviour for strong time-varying room impulse responses, whereas the computational complexity is relatively low. The implementation of such a system and a simulation in MATLAB with measured room impulse responses are further aims of this project.