|(Left: File photo of a Chinese drone of CH series. Right: Snapshot of the match between AlphaGo and Go champion Lee Sedol)|
China should set its sights on designing an unmanned automatic aerial vehicle battle system to prepare for the threat of air battle between drones, according to a professor with the Air Force Engineering University (AFEU).
Based on a number of flight parameters, the battle system, coded as “AlphaEagle,” should be able to automatically assess battle situations and calculate threats before responding with offensive or defensive measures--theoretically similar to Google DeepMind's AlphaGo, said the professor, Huang Changqiang, at a forum in Beijing on Wednesday.
The forum is a sub-forum of the 4th China Conference on Command and Control, which is a three-day conference held by the Chinese Institute of Command and Control.
AFEU has been accumulating parameters from air battle training and live ammunition fight tests for three generations of fighter jets since 1998. All this data could be of use in the development of an unmanned combat aerial vehicle (UCAV), according to Huang. Huang added that the university is already testing a model UCAV.
“Drones are currently like kites flying under the control of human beings. If we can learn from AlphaGo and create the artificially intelligence ‘AlphaEagle,’ our air force will hover high in the sky,” he promised.
Calling the system a “virtual pilot,” Huang said it should be able to learn from the data of human pilots, and in doing so figure out how to calculate its offensive-defensive role so as to maintain an advantage during battle.
He also pointed out that drone battles are a serious imminent threat, as unmanned aerial vehicles (UAV) can easily overcome human pilots. This is because their only limitation is the weapon itself or the aircraft’s performance. Huang added that major powers like the U.S. have already stepped up their efforts toward UCAV development.
“A human pilot will spend 50 percent of his or her attention on winning the battle, and the other 50 percent on self-protection. The virtual pilot can keep the ratio around 90 to 10,” Huang explained.
Huang also mentioned that drone pilots can suffer mental trauma during and after battle, while an automatic system would not have that potential problem.