We consider the inverse problem of discovering the location of a source from very sparse point measurements in a bounded domain that contains impenetrable (and possibly unknown) obstacles. We present an adaptive algorithm for determining the measurement locations, and ultimately, the source locations. Specifically, we investigate source discovery for the Laplace operator, though the approach can be applied to more general linear partial differential operators. We propose a strategy for the case when the obstacles are unknown and the environment has to be mapped out using a range sensor concurrently with source discovery.

@InProceedings{BLTT08,
  author       = {Burger, M. and Landa, Y. and Tanushev, N. and Tsai, R.},
  title        = {Discovering point sources in unknown environments},
  booktitle    = {Proceedings of WAFR 2008},
  year         = {2008},
  url          = \{/2008/BLTT08},
}