Airbus used machine learning technology to conclude the fully autonomous flight
After a comprehensive testing programme that has taken place over a period of two years, Airbus has announced that it has completed its ATTOL (Autonomous Taxi, Take-Off and Landing) project.
The project has made use of machine learning and artificial intelligence to enable a commercial aeroplane to perform taxiing, take-off and landing autonomously.
This world-first has been achieved over a series of automatic vision-based flight tests using innovative image recognition technology.
Airbus’ “Wayfinder” software gathers information using computers, cameras, radar and LiDAR (a detection system that uses light from a laser) to help an aircraft assess its environment and navigate its way through.
Over the course of the programme, 500 test flights were carried out. The purpose of around 450 of these was to collect video data and calibrate algorithms. Six test flights involved five take-offs and the same number of landings per flight to assess the aircraft’s ability to fly autonomously.
The core focus of Airbus’ ATTOL project was to understand more about how technology powered by machine learning and artificial intelligence, including algorithms and automated tools to enable data labelling, processing and the generation of models, could allow pilots to divert their focus from operational to strategic and decision-making matters.
Referring to the project, Airbus stated that it believes autonomous flight to be a joint venture between humans and machines. The new technology allows the collection and presentation of data, while the pilot’s role is to analyse this data and base decisions upon it.
Video: Airbus autonomous take-off
The company also said that it would use these learning opportunities to create new ways of developing, building, flying, powering and servicing aircraft.
It believes that new technologies such as these can help enhance air traffic management systems, sustainability and advance aircraft safety even further.