Embedded Reasoning for Atmospheric Science Using Unmanned Aircraft Systems

The ability to understand and predict the dynamic behavior of our planet’s environment over multiple spatial and time scales remains an outstanding challenge for science and engineering (National Research Council 2007). For the past 35+ years, sensors on spaceborne platforms have increasingly surveilled most of the Earths surface and atmosphere. However, a summary of US severe weather data over the past 12 years shows no sustained reduction of fatalities or property losses (NOAA National Weather Service 2007). To reduce losses, in situ sensing and forecast systems must be improved. In particular, this paper addresses the use of a particular form of unmanned aircraft system to provide embedded reasoning for atmospheric science. The development of mobile systems for complex sensing applications has motivated the cooperative control of heterogeneous robots in general and unmanned aircraft systems in particular. Combining large numbers of robotic sensors leads to improved flexibility compared to single-robot systems. While heterogeneity expands the set of possible actions, solutions for general teaming are difficult, especially when robot capabilities vary drastically. This paper focuses on a specific form of heterogeneous unmanned aircraft system (UAS) comprised of two classes of aircraft with significantly different, though complementary, attributes: 1.) miniature daughterships (DS) that provide improved flexibility and spatio-temporal diversity of sensed data and 2.) larger motherships (MS) that carry and deploy the daughterships while facilitating coordination through increased mobility, computation, and communication. A fundamental drawback of miniature unmanned aircraft (MUA) is their short range, low endurance, and small payload capacity compared to larger systems. On the other hand, MUA can be deployed in large numbers, allowing groups to achieve significant spatio-temporal sensing of an environment. Because these systems are small, they are often safer close to terrain and obstacles, have smaller weight and speed so the kinetic energy released in a collision (should it unfortunately occur) is low, and are inexpensive and expendable. The motherships extend the capability of the UAS in sev-