Knowledge based Autopilot with cognitive properties.

Hi all,

Alfatroll AS in Norway has developed a knowledge based autopilot. We intend to cover all aspects of a pilots dutoes, such as....

  • Flight Operation are all normal tasks covered by an autopilot, such as bringing the plane safel3689563612?profile=originaly from waypoint to waypoint along the scheduled trajectory and at the designated time schedule (dubbed 4D control)
  • Mission Progress, including necessary actions to be carried out at pre-defined locations or in certain situations. The task includes contingency management, i.e. reactions to unexpected events that may occur during the flight, such as engine failure, low on fuel, etc.
  • Contact with base, including reports back in the form of data streams and status messages, as well as the ability to react promptly to orders from the ground station.
  • ATC relations. In the beginning, all communication with air traffic controllers (ATC) will most likely be through the ground station controller. When ATC communications becomes message based (Single European Sky specifies this), more and more can be handled directly by the unmanned aircraft itself. Proper responses and reactions are expected by the on-board intelligent controller, of course.
  • System Health. A major part of the responsibility of an autonomous system is to keep track of its own operational status. System health can include such tasks as the checking of and proper reactions to icing or icing conditions, low fuel reserves, sensor information signaling trouble with an engine, slow reactions, etc.
  • Detect Terrain. By following the plane’s movements in relation the terrain below, the control logic shall be able to avoid conflicts with terrain objects, controlled or prohibited airspace, tall objects on the ground, etc. The plane shall also be able to react differently to corrective actions at low level flight from normal high level flights.
  • Sensor info and collection. Sensor information is normally collected on-board or forwarded to the ground control. If the sensors detect items of interest to the current mission (e.g. a hot spot in the sea during search and rescue), this may cause the control logic to launch specific actions (e.g. drop marker, go lower and take picture/video). This task can also be responsible for intersystem cooperation.
  • See and Avoid. This is the focus of many UAV producers at the time: to make systems that can detect other air traffic or other obstacles in the same, non-segregated airspace, and react as if a pilot was on-board. Using ADS-B and other sources, keeping track with traffic nearby is possible. But visual and radar detection will always be less precise.

We use APM Mission Planner as ground control, and some of the features can be seen in this video: See video here

We seek co-operation partners, preferably in Europe, due to distances etc.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • New video here.

  • Thanks. The key issue in this project is that the software is the key. Platforms (planes come with various shapes and properties) are abundant.

  • This sounds great. Keep up the good work!

This reply was deleted.