The Autonomous Ocean Sampling Network -II (AOSN-II) and Adaptive Sampling And Prediction (ASAP) projects aim to develop a sustainable, portable, adaptive ocean observing and prediction system for use in coastal environments. These projects employ, among other observation platforms, autonomous underwater vehicles that carry sensors to measure physical and biological signals in the ocean. The measurements from all sensing platforms are assimilated in real-time into advanced ocean models. The objective is to coordinate the mobile assets in order to collect data of highest possible utility. Critical to this effort are reliable, efficient and adaptive control strategies to enable the mobile sensor platforms to collect data autonomously. In this paper, we summarize feedback control strategies that enable us to gather useful information over a wide spectrum of spatial and temporal scales. First, we design formation control strategies to sample small spatial scale processes (less than 5 km). In this framework, the feedback control laws maintain a desired formation of vehicles and allow the group to locate interesting features in the ocean. Some of these control strategies were implemented on a group of underwater gliders in Monterey Bay in August 2003, as part of the AOSN-II project. Second, we direct mobile sensor networks to provide synoptic coverage to investigate larger scales ($5\!-\!100$ km). Coordinated vehicle trajectories are designed according to the spatial and temporal variability in the field in order to keep sensor measurements appropriately distributed in space and time.