The surface of the Earth and planetary neighbors like the Moon and Mars are blanketed with granular materials. Whether they are composed of soil, regolith, or frozen admixtures, our success or failure in exploring and manipulating them hinges on the reliability of models for granular rheology. In this dissertation, I rely on a combined approach of fieldwork and laboratory experimentation, and use a sensitive robotic rheometer — capable of measuring in situ material strength through a force based sensing technique — to investigate the dominant controls that influence the strength of granular materials. Using this method, I show that purely frictional granular materials can be collapsed onto a master curve that relates changes in friction to the distance from a critical volume fraction, that sand-rich soils are mechanically distinct from clay-rich soils, and that the rheology of frozen granular materials undergoes transitions in strength related to ice content. This thesis bridges concepts outlined in soft-matter physics, geotechnical engineering and geomorphology, and robotics, and in doing so provides novel insights into predicting granular rheology in unknown and complex environments — both on Earth and other planetary surfaces.