ARCAA Home Personnel Research Facilities Media News Links Partners

UAV Forced Landing Research

Research Programs
Research Projects
System Design and Optimization
Sensors and Advanced Flight Systems
  Guidance, navigation and control
  Sense-and-Act
  * Flight Termination
  Propulsion
Smart Payloads
Systems Integration, Regulation and Operation
Publications

A range of technology-based and operational-based mitigation strategies may need to be included as part of the safety-case justifying the operation of a UAS over a "populous area".

The UAV Forced Landing Research Program is providing unique research into the feasibility and development of a forced landing system capable of replicating the forced landing function as performed by a human pilot.

High-level illustration of the forced landing process
Illustration of the forced landing process

Specifically, the research program investigates:

It is important to note that this technology can also be used as a decision aid for piloted aircraft - a forced landing is something that all pilots must train for, but something that very few practice until faced with the emergency.

 

1. Candidate Landing Site Identification and Classification


This section of UAV Forced Landing Research Program focuses on the initial problem of finding a suitable place to land under a forced landing scenario.

Example Output from site identification algorithm

 

Description of Research

The research explores the application of machine-vision techniques to automatically identify and classify suitable landing areas.

The algorithms use image segmentation and pattern recognition techniques drawn from the machine-vision field to identify candidate landing sites from a moving aircraft. A neural-network is then used to classify candidate landing sites.

Current Research Programs

Publications

Related Sites

Research Personnel

Related Media

This short video shows the automatic forced-landing site identification algorithm in action!
 

Back to Top

 

2. Multicriteria Decision Approaches


Once candidate landing sites have been identified and characterised, a process is needed for selecting the most suitable landing site from the set of candidate sites identified.

 

Illustration of The Site Selection Process

 

Description of Research

This section of the UAV Forced Landing Research Program investigates the application of multi-criteria decision making techniques for addressing the landing site selection problem.

Site selection for forced landing involves Multiple Criteria Decision Making in real time, under conditions of uncertainty, and with constantly changing inputs. A number of potential landing sites would be identified in the previous phase through image processing, and one must be selected that would facilitate the highest chance of a successful landing. The decision making process will take into account various attributes such as wind velocity, candidate landing site geometry, proximity to population, obstacles, etc., while trying to minimize the risk to human life, property, and the UAV itself in this order of priority. The entire process will be implemented onboard the UAV, being fully automated, while allowing intervention from human operators if necessary.

Current Research Programs

Publications

Related Sites

Research Personnel

Related Media

   
 

Back to Top

 

3. Navigation and Trajectory Planning


Phase 3 of the forced landing problem requires the UAS to be able to determine a trajectory to reach an identified landing site. 

Illustration of forced landing trajectory problem

 

Description of Research

This section of the UAV Forced Landing Research Program focuses on enabling a UAS to intelligently construct its own landing trajectory to a desired landing site location, from any initial position, and also to be able to dynamically adjust that trajectory to deal with changing winds and uncertainties in the location of the landing site.

The approach is currently researching the use of motion primitives, where the final, planned trajectory can be idealized as a sentence that is constructed from different words, called primitives, obtained from a library of such primitives.

An additional component in the final landing approach ensures that the UAS is able to determine a suitable final landing approach. This component of the UAV Forced Landing Research Program focuses on the use of EO sensors and vision-based processing techniques to provide the necessary attitude and navigation information to allow a UAS to perform a landing at a landing site with no or little prior knowledge on the conditions of the site.

The research has successfully employed a range of vision-domain techniques including optical flow and horizon detection.

Current Research Programs

Publications

Related Sites

Research Personnel

Related Media

   
Here we demonstrate a number of forced landings with a Boomerang R/C plane in Queensland, Australia. This flight is manual R/C only, however the intent is to automate this process
 

Back to Top

 

Research Team:

Back to Top

 

.