Networked aerial robots for atmospheric sensing

According to the National Weather Service, 438 fatalities, close to
3000 injuries, and $11.2 billion in damages were incurred due to severe
weather in the United States in 2003. These losses could be
dramatically reduced with effective advanced prediction and warning
systems. Tornadoes are especially violent members of the severe storm family and thus the study of tornado formation and evolution is a
public safety necessity. There was an immediate improvement in tornado
warning capabilities with the introduction of Doppler radar. However,
primitive sensing methods such as human storm chasers and spotters
remain the most vital part of the public safety system. The inability
to determine the volumetric thermodynamic state of the atmosphere
between the ground and the base of the mesocyclone remains a major
barrier towards a deeper understanding of tornado genesis. The
limitations of remote sensing are evident; one cannot remotely sense
the thermodynamic field, these data can only be obtained with in situ
sensing.

A distributed sensing system is being developed by the Research and Engineering Center for Unmanned Vehicles (http://recuv.colorado.edu) at the University of Colorado to probe severe convective storms using an Unmanned Vehicle System (UAS) in conjunction with simultaneous dual-Doppler radar and other sensors. In support of this
effort, a suite of software was developed to allow for real time
visualization of radar and UA information. Through this interface, operators are able to control a UA to an area of
interest based upon meteorological information. An existing ad-hoc
network was augmented to allow for the effective dissemination of
telemetry, sensor data, and control throughout the multi-user network.

Additional information on this project can be found at http://tornadochaser.colorado.edu/.