Robotics has dramatically increased our ability to gather data about our environments. This is an opportune time for the robotics and algorithms community to come together to contribute novel solutions to pressing environmental monitoring problems. In order to do so, it is useful to consider a taxonomy of problems and methods in this realm. We present the first comprehensive summary of decision theoretic approaches that are enabling efficient sampling of various kinds of environmental processes. Representations for different kinds of environments are explored, followed by a discussion of tasks of interest such as learning, localization, or monitoring. Finally, various algorithms to carry out these tasks are presented, along with a few illustrative prior results from the community.

翻译：机器人技术极大地提升了我们从环境中采集数据的能力。当前正是机器人与算法领域的研究者携手合作、为紧迫的环境监测问题贡献创新解决方案的绝佳时机。为此，有必要对该领域的问题与方法进行系统分类。本文首次全面总结了一系列决策理论方法，这些方法能够高效地对各类环境过程进行采样。我们探讨了不同类型环境的表征方式，随后讨论了学习、定位或监测等相关任务。最后，介绍了执行这些任务的多种算法，并附上该领域若干有代表性的前期研究成果。