Remote, unknown environments are also some of the most compelling for scientific exploration because of the lack of data from such regions. Reliance on mobile robots for exploration of these areas reduces cost and risk, and autonomous mobile robots that are capable of adapting their sampling strategies increase the amount of information gained through exploration. These robots rely on dense, predictive mapping of large-scale regions to determine where to take measurements. Current approaches to mapping commonly rely on Gaussian process regression to spatially correlate data, extrapolate from sparse samples, and estimate uncertainty. However, these approaches do not incorporate meaningful information about physical processes that contribute to the scientific observations being recorded, instead focusing on predicting measurements only or relying on black box expert-based systems to incorporate physically meaningful information.

In this work, we introduce physical process-informed mapping, a framework for integrating scientific knowledge about physical processes into mapping for mobile exploration robots. This approach provides two advantages for exploration. First, the hierarchical statistical model underlying the mapping framework is capable of providing predictions and uncertainty estimates for both the sensor measurements and the latent natural phenomena driving these measurements in a dense, probabilistic manner. Second, various components of the mapping approach provide methods to measure the importance of models of scientific knowledge, enabling learning about the conditions under which these models are more or less relevant to the phenomenon. Taken together, these methods increase the ability of robotic mapping for exploration to model complex physical phenomena, enabling future robotic explorers to take advantage of a more in-depth representation of information, and can provide increases in the predictive accuracy of mapping.