We have all seen movies where the hero who doesn’t know how to fly a plane successfully lands thanks to the autopilot mode. But did you know that these technologies have some artificial intelligence (AI) basis?
Aviation is one of the most heavily regulated industries in the world. Because of the need for security in order to reduce the number of accidents (even if commercial aviation is already the safest transport system – after lifts), the aeronautics industry was among the first to use AI technologies with autopilots or load-shedding electrical systems, which use computer power to make smart decisions.
Currently, they are mostly used as passive system to control the airplane after entering specific parameters. But the future is much more interesting! AI will enable airplanes to become proactive and make decisions, thanks to machine learning and neural networks.
Training and assisting pilots
To enhance safety, the first thing AI is going to improve is the pilot’s training. By using AI in simulators, pilots will have more realistic simulations with dynamic scenario generation and adaptation. During the training, the computer also records flight data and pilot behavior to teach autopilots how to become more efficient in real flights. These simulations are also used to train future air traffic controllers.
The second thing is to assist pilots better. This comes by enhancing autopilots with the real power of AI: being predictive instead of reactive. Just as you have a notification on your phone when you need to leave for a meeting, an autopilot could alert you about your fuel level, systems status, weather or even remind you to do a checklist before you have to ask or think about it. Beyond the alerts, the autopilot could also advise on new flight plans dynamically generated with weather data, fuel consumption rate and other parameters it needs to take into account.
We are at the point where autopilots are more than just software autopilots: DARPA has presented a project this year called ALIAS (Aircrew Labor In-cockpit Automation System), which is a physical autopilot that manipulates the stock controls of an airplane. It is primarily designed for the military. This way, no need to buy or design a new plane with next-gen AI capabilities, militaries can just use this drop-in option on any airplane.
Flight safety and planning
An everyday hassle for the pilot is to plan his flight well and ensure its safety. As a matter of fact, bad planning is responsible for 60% of accidents in leisure aviation. To help them plan their flights more effectively, more and more sophisticated autopilot programs have emerged.
The Garmin’s Electronic Stability and Protection system is one of these. This passive safety system continuously monitors the airplane’s attitude and uses the autopilot to put the airplane back to safety if it is either too low or at a wrong angle.
Garmin is also trying to introduce voice commands with its latest product, the Garmin Telligence Voice System Command. It uses voice recognition, NLP and machine learning to provide an interactive experience with your plane. That’s why bots are not just chatbots, but can also be integrated in complicated systems like those. You can ask Garmin Telligence how far you are from your destination, to show the page of the nearest airport, to save data such as the temperature or to switch panels from map to flight plans, and this is only a small part of the more than 300 commands it can carry out. To see it in action you can watch the video below:
Systems monitoring and incident handling
Even if pilots are well trained and well assisted, incidents still happen. AI can also help with monitoring systems and handling incidents.
In 2003, NASA’s Dryden Flight Research Center created a software which could land a damaged aircraft by compensating the damaged areas with the undamaged ones. NASA also uses the Integrated Vehicle Health Management System, which placed on an aircraft can determine its structural integrity and implement protocols in case of damage. Another interesting innovation was developed by researchers at University College London. They created the autopilot Cirrus SR-22 which can automatically and dynamically adapt to the environment and flight situation by using machine learning. The software builds a list of scenarios by storing hundreds of hours of detailed data from real flights. Here, AI is used to process a large amount of data, which can also be applied to the maintenance of the airplane.
As prevention is better than cure, day to day maintenance of the aircraft itself is crucial. AI can automatically and periodically run maintenance scenarios and check systems statuses. By using huge amounts of flight data processed thanks to machine learning, it can detect failure or inconsistent patterns and trigger repair at the first signs of weakness before something is broken. By learning from hours of flight data, maintenance systems will also be able to optimize the parameters of the systems during a flight to avoid weakening as well as providing insights to engineers designing new planes.
So, what’s stopping these technologies from spreading even further in aviation? Just one thing: regulation.
Historically, it moves slower than technology, because ensuring safety requires lots of tests and certifications. However aviation is ahead of the automotive sector in many areas, thanks to the dynamism of this industry.
There is something that is beginning to accelerate autopilot technology, and that’s the applications used in drones. Because drones don’t have a pilot, they need to be autonomous in the decision-making process, avoiding obstacles, handling traffic, weather, self-diagnosing mechanical problems and many more. Drones are a great and mostly safe playground to test all these technologies before putting them into mainstream aviation.
Until now, the main thing that enhanced safety has been machine automation: the more we let machines control a system, the more effective and safe it is. Even if this feels uncomfortable, the moment we will allow machines to be entirely autonomous in flight, we’ll have the safer, faster, and more efficient flights we dream of.
Julien Blancher – Recast.AI