Distributing omnidirectional captors over the space, and using some processing power they should be able to: - Locate spatially the sound sources. - Extract the sound emitted at a specific point of the space.
This has many application in noisy ambient sound recognition, real ambients automation systems, teleconferencing and others.
Some papers can be found using the terms:
and other stuff like that.
wikipedia.org dumped 2003-03-17 with terodump