Skip to main content

helicoPter and aEronef naviGation Airborne SystEms

Final Report Summary - PEGASE (helicoPter and aEronef naviGation Airborne SystEms)

The PEGASE project was a feasibility study of a new navigation system which allows a three dimensional (3D) truly autonomous approach and guidance for airports and helipads and improves the integrity and accuracy of GNSS differential navigation systems. This new navigation system relied on three key technologies:
- specification of a reliable ground reference database
- innovative correlation techniques between sensors outputs and onboard ground database
- a robust servoing algorithm for the management of the trajectories of both fixed wing and rotary wing aircraft.

The main objectives of the study were:
- to assess the feasibility of the autonomous, all weather conditions, localisation and guidance system;
- to determine the performance required for the vision sensors (visual, infrared, electromagnetic) and the ground reference database in order to provide an accurate guidance from the final approach leg to full stop on the taxiway / helipad and for the take-off run from alignment up to the final take-off segment.

Two different threads of activity have been defined:
- how to create an onboard database safe and secure enough to allow the tracking and servoing;
- Assuming we are able to build such a database, what would be the candidate algorithms that would allow the tracking and servoing to work?

The study was carried out with existing simulation tools and man machine interface (MMI) developed for fixed wing and rotary wing aircraft.

The activity concerning the sensors was carried out by EADS and achieved the partners to have:
- a clear view of what are the requirements on the visual sensors when working for the landing / take-off functions;
- a clear view of the needs to develop new sensors and mainly a front line radar imaging system;
- a software package able to integrate in the shared simulation framework in order to create 'real' sensors using the image generator provided by Dassault Aviation.

The activity concerning the 'database' was led by ETHZ and had the following results:
- Work has been done, between ETHZ and Dassault Aviation, to reorganise the planning and the related activities in order to answer the alert raised during the first year.
- Walphot SA, helped by Dassault Aviation, completed the work done upon the state of the art of the existing data sources to take into account the emerging satellite systems to come in the following years. The obtained data has been used but is an achievement on itself and could be provided for other purposes.
- ETHZ provided to ELPHO the fusion algorithm to be implemented in a commonly agreed geographical tool (ARCGIS). This algorithm has been thoroughly tested and developed on real data.
- ETHZ worked on the quality assessment of database introducing local measures such as roughness. First insights now exist that could lead to proposal towards certification world through EUROCAE working group.
- ETHZ elaborated a methodology to be further developed in order to automatically extract edges, leading to automatic buildings identification.
- Dassault Aviation promoted a new way to approximate terrain data using Brownian motion. This approach, though at its research step, could lead to more accurate way to assess quality and manage accuracy of interpolated terrain data. This is an ongoing work.
- The first assessment framework has been defined between the partners concerned during this period (ETHZ as leader, Dassault Aviation and Walphot). It should be used throughout the last period.

The activity concerning 'visual tracking and servoing' was led by INRIA and had the following results:
- Runway/helipad detection: image processing algorithms have been developed by CNIT for the detection of a runway or a helipad in the images acquired by the on board camera.
- 2D tracking: an image processing algorithm has been developed by CNRS for the tracking of a runway in the image sequence.
- 3D localisation: an image processing algorithm has been developed by EPFL for the 3D localisation of the on board camera using a visual memory composed of a set of points of interest.
- 3D tracking: complementary image processing algorithms have been developed by INRIA. EPFL and IST for the 3D tracking and trajectory estimation of the on board camera with respect to the environment observed during the landing phase. These algorithms are based on various image features and a geographical database provided by WP5. PBVS: JSI, IST and INRIA have developed position-based visual servoing methods for the automatic landing of fixed-wing aircraft.
- IBVS: Similarly, CNRS and INRIA have developed image-based visual servoing methods for the automatic landing of fixed-wing aircraft. These position-based and image-based control schemes address the different levels of the AFCS provided by Alenia Aeronautica in the PEGASE simulator.
- Obstacle detection: this activity has been defined last year following questions from the commission. It is assumed entirely by CNIT that proposed a first version of algorithms for the detection of potential fixed or moving obstacles lying on the runway or the helipad.

CNRS has developed an autonomous waypoint navigation method. Finally, Eurocopter Deutschland has worked on the definition of the human-machine interface to be used in the piloted simulation software.

Related documents