Research Overview

This research project investigates the feasibility of automating the emergency forced landing procedure.    We have studied the issues relating to the forced landing process for a UAV and specifically focussed on the methods for making the initial landing site selection after a failure has occurred, based on machine vision techniques.  The technology that we have developed can be used on manned aircraft also, such as providing information on possible landing sites for pilots placed in emergency scenarios or for providing a number of landing sites for glider pilots.  (Future Research)

A number of algorithms have been proven to identify appropriate landing areas - large open spaces, free of obstructions, suitable for landing purposes.    Further algorithms have been shown to automatically classify the surface type of the landing sites using neural networks to allow a higher level process to select the most appropriate landing site from the ones available.    (Results)

This research is critical to allow UAVs to fly in civilian airspace above populated areas in case of unforseen emergencies.    Human pilots are trained in this skill and we believe that it is a fundamental capability that UAVs must have.  The technology can also add in an extra level of safety for piloted aircraft also - a forced landing is something that all pilots must train for, but something that very few practise until faced with the emergency.  (Background)

Finally, we have found no evidence of current techniques or commercial systems available for UAVs or piloted aircraft that perform the level of capability that we are working towards.    It is believed that this technology will have an enormous impact on the future systems that regulatory organisations such as CASA or the FAA will impose on UAVs operating in civilian airspace.     Our dominance in this research space can be seen by doing a simple Google search for "UAV Forced Landing".

The initial research was conducted as part of a 4 year PhD course at the Queensland University of Technology, Australia finished in 2006.    Current activities are focussed around trialing these algorithms on a small UAV platform as well as path planning and the high level decision making required to select the best landing site out of the available ones identified by the site selection algorithm. Visit the Video page to see the latest videos.