Online decision making for stream-based robotic sampling via submodular optimization

We consider the problem of online robotic sampling in environmental monitoring tasks where the goal is to collect k best samples from n sequentially occurring measurements. In contrast to many existing works that seek to maximize the utility of the selected samples online, we aim to find the cardinality constrained subset of streaming measurements under irrevocable sampling decisions so that the prediction over untested measurements is most accurate. Using the information theoretic criterion, we present an online submodular algorithm for stream-based sample selection with a provable performance bound. We demonstrate the effectiveness of our algorithm via simulations of information gathering from indoor static sensors.