In this paper, the problem of characterizing adaptive output feedback control laws for a general class of unknown MIMO linear systems is considered. Specifically, the presented control approach relies on three compone...
详细信息
Acoustic fingerprinting technique with multiple microphones is proposed as an effective acoustic localization method in 3D cluttered environments. A sound source generates a unique signature in the form of a spectrum ...
详细信息
ISBN:
(纸本)9781627485609
Acoustic fingerprinting technique with multiple microphones is proposed as an effective acoustic localization method in 3D cluttered environments. A sound source generates a unique signature in the form of a spectrum at each location in a static environment. The sound pressure spectrum recorded by the microphones dispersed in the environment are compared to a dataset of previously determined spectra (fingerprints) produced by candidate source locations. It is, however, impractical to experimentally measure all these fingerprints for large areas. The use of an energy-based acoustic propagation model to predict signal fingerprints is an effective alternative used. The source location is then estimated by finding the best match between the measured fingerprints by the microphones and the reference fingerprints stored in the database. A numerical evaluation of the method is performed for 3D cluttered environments where finite-different, time-domain (FDTD) is used to simulate the microphone signals. Fingerprints are then predicted using a propagation model based on the diffusion equation. The effect of the accuracy of the fingerprints is also addressed using the exact fingerprints based on the FDTD acoustic model. The results show the fingerprints do not need to be very accurate if enough microphones are dispersed in the environment.
.jpg?site=8fa65c6403e3f8b5b3923814eda18bd0)
暂无评论