Researchers at the Massachusetts Institute of Technology (MIT) say they have developed a new technique using which they successfully conducted an experiment on the lines of the mission shown in the Tom Cruise-starrer summer 2022 blockbuster ‘Top Gun: Maverick.’
In the film, Cruise’s US Navy veteran pilot, Captain Pete ‘Maverick’ Mitchell, leads a team of young pilots on an extremely dangerous mission, which the group accomplishes. In the experiment, however, the MIT team showed that in real-life, the operation is possible with an autonomous aircraft as well.
First, what was the mission in ‘Top Gun: Maverick’?
Captain Mitchell and his team are tasked with destroying an unsanctioned uranium enrichment plant. Their target, however, is located deep inside a rocky canyon. The pilots, therefore, must fly their jets deep inside the canyon, staying so low to the ground that they cannot be detected by radar, hit the plant, and then rapidly climb out of the canyon, taking care to avoid the rock walls.
All this is in addition to the high-tech weaponry at the enemy’s disposal to protect the plant.
What did MIT team do in their experiment?
The researchers effectively piloted a simulated ‘jet’ through a narrow corridor without the machine crashing into the ground.
How was this achieved?
For an autonomous aircraft, if sent on such a mission, the most straightforward path towards the target goes against what the machine must do to avoid colliding with the canyon walls, or to stay undetected. This conflict, which many existing artificial intelligence (AI) methods have not been able to resolve, is called ‘stabilise-avoid’ problem.
To solve this, the researchers broke down the issue into 2 parts: they reframed the stabilise-avoid problem as a constraint-optimisation problem, and then, this constraint-optimisation problem was solved using a deep-reinforcement learning algorithm.
The experiment
The team conducted several control experiments under different initial conditions. Finally, they flew their ‘jet’ in a scenario simulating the one portrayed in the Oscar-nominated movie. Their controller was able to prevent the jet from crashing or stalling.
The ‘jet’
The model was open-sourced in 2018. It was designed by flight control experts as a testing challenge.
“