Robotic choppers learn new stunts through observation
Flying-instruction manuals may now become a thing of the past, for scientists at Stanford University have developed an artificial intelligence system that would allow robotic helicopters to learn to perform complex stunts by imitating other helicopters. The helicopter performed stunts on its own and the reason it could is the underlying software. Helicopters are unstable when compared to other aircraft, and without constant monitoring, they are bound to crash. The Stanford scientists have tried to model the technology on an expert radio control pilot's abilities. The learning algorithms helped helicopters to learn from the radio control pilot, Garett Oku. In this new technology, the instruments would be installed on the ground and the rest on the helicopter. These mechanisms monitor the position of the helicopter and also its direction, orientation, velocity, acceleration and spin in many dimensions. The chopper would carry accelerometers, gyroscopes and magnetometers. The ground control would have the GPS receiver and the cameras. The helicopters could be used in detecting land mines and also during war situations. However, the technology really has to improve if one has to trust helicopters to fly on their own and also learn from other helicopters.
Via: The Raw Feed
