Students working in the Human Centered Machine Intelligence Laboratory under the direction of Dr. Michael Goodrich are exploring the potential of unmanned aerial vehicles (UAVs) in wilderness search and rescue.  In missing person scenarios, these miniature remote controlled planes, fitted with cameras and infrared sensors, can be sent out to survey rugged wilderness areas that could be hazardous for search and rescue workers to explore.  The goal is to create a mobile ground control station from which a single operator with minimal training could deploy, control, and recover a UAV equipped with video and infrared capabilities.

Although they look like roughly-assembled model planes, these miniature contraptions have the potential to save lives. Currently, search efforts are problematic.  They are extremely time consuming, costly, dangerous, and easily delayed by night, difficult terrain, and unfavorable weather conditions.  The HCMI lab believes UAVs to be part of the solution.  Time is a critical factor in search and rescue efforts.  With each passing hour, the search area grows exponentially, meaning that the longer the missing person is out in the wilderness, the more difficult it will be to predict his or her location.  UAVs can make quick aerial sweeps of the area without having to send in search and rescue teams.  In addition to saving time, it also spares rescue personnel from the potential dangers posed by rugged wilderness terrain.  Furthermore, by using UAVs to make initial searches of the missing person's probably location instead of sending in rescue workers, the costs associated wilderness search and rescue decrease, ensuring that funds for future searches will be available. 

The HCMI team is also studying the methods which are currently used in search and rescue operations.  They have devised ways to model the movement of the missing person and thereby locate the most probable search area.  Using historical data and case studies of actual rescues obtained from the National Association of Search and Rescue, Dr. Goodrich and his students developed three models which describe the most likely movements of a person who is lost in the wilderness and how those movements change over time.  Each mathematical model takes different factors into account.  The first model looks at the slope of the terrain.  Data shows that persons wandering in the wilderness are most likely to head downhill.  However, older, more experienced woodsmen, such as hunters, move downhill more quickly, whereas young children will meander before beginning a downhill course.  The model takes all of these details into account.  The second model analyzes the slope of the terrain as well as the possibility that the missing person will have strong indicators as to the direction in which they are heading.  The third model takes into the account the vegetation and trails in the area, as a missing person is likely to follow trails and avoid areas of dense trees or thick brush.  Having completed the three models, the HCMI lab is now attempting to validate its research by running simulations and comparing the results with actual case studies.