The objective of this section is to present existing models of integration specifically proposed for speech perception. It is introduced by a discussion on how fusion of sensory captors relies basically on two disciplines: Cognitive Psychology and Engineering Sciences. This report finally proposes a taxonomy of the integration models that can be found in the literature. Later on, we will make another report on the quantitative implementation of such models. Here again, those techniques come from two scientific domains, namely Cognitive Psychology and Automatic Speech Recognition.