Source Localization in Reverberant Environments: Part I -modeling

Source localization using microphone arrays has for some time been an active and challenging research area. The main obstacle in designing robust practical systems is the e ects of room reverberation. This paper, along with a companion, analyzes the in uence of reverberation. In this paper, a new statistical model is developed that explains the e ects of reverberation. The model is based on the theory of statistical room acoustics and is a generalization of the Ricean model used to model channels in digital communication. The properties of the statistical model is evaluated with both simulations using the image method, and with real data measurements. The agreements are quite good lending support to the model. The model is further validated in the companion paper where it is used to understand performance limits of source localization estimation accuracy in reverberant environments, and to understand the performance of source localization algorithms.