Robust Sensor Placement for Signal Extraction

This paper proposes an efficient algorithm for robust sensor placement with the purpose of recovering a source signal from noisy measurements. To model uncertainty on spatially-variant sensors gain and spatially correlated noise, we assume that both are realizations Gaussian processes. Since to noise ratio (SNR) is also uncertain in this context, achieve extraction, propose new criterion based maximization probability SNR exceeds given threshold. can be easily evaluated using process assumption. Moreover, reduce computational complexity joint respect all positions, suggest sequential approach, where positions chosen one at time. Finally, present numerical results showing superior robustness proposed approach when compared standard criteria aimed interpolating spatial recently maximizing average SNR.