导航

ACTA AERONAUTICAET ASTRONAUTICA SINICA ›› 2023, Vol. 44 ›› Issue (10): 128094-128094.doi: 10.7527/S1000-6893.2023.28094

• Fluid Mechanics and Flight Mechanics • Previous Articles     Next Articles

Air combat maneuver decision-making test based on deep reinforcement learning

Sheng ZHANG1, Pan ZHOU1, Yang HE1, Jiangtao HUANG1(), Gang LIU2, Jigang TANG1, Huaizhi JIA3, Xin DU1   

  1. 1.Aerospace Technology Institute,China Aerodynamics Research and Development Center,Mianyang 621000,China
    2.China Aerodynamics Research and Development Center,Mianyang 621000,China
    3.School of Aeronautics,Northwestern Polytechnical University,Xi’an 710000,China
  • Received:2022-10-08 Revised:2023-01-05 Accepted:2023-02-15 Online:2023-05-25 Published:2023-02-24
  • Contact: Jiangtao HUANG E-mail:hjtcyf@163.com
  • Supported by:
    National Natural Science Foundation of China(11902332)

Abstract:

The air combat intelligent decision-making will greatly change the form of wars. Deep reinforcement learning decision-making machine, as an important technical paradigm to realize the intelligent decision-making in air combat, can explore the potential of unmanned aircraft. However, reports on its engineering implementation are rare. Aimed at the practical implementation of the maneuver intelligent decision-making based on deep reinforcement learning in the one-to-one fighters’ close-range air combat, an online deep neural network maneuver decision-making model suitable for application is developed. The maneuver control scheme that the trajectory guidance decision-making commands are tracked with the flight control law is proposed. The corresponding software and hardware architectures are realized and the human-machine combat flight test is carried out, which achieves the transfer from virtual simulation to real flight in intelligent air combat. The research results show that, based on the close-range air combat maneuver decision-making and control method developed in this paper, the intelligent unmanned aircraft can make logical maneuver decisions quickly in favor of its own side and thus is soon in the advantageous situation by maneuver when combatting with human “pilots”. The flight test results demonstrate the potential application value of the deep neural network intelligent decision-making machine in air combat decision-making.

Key words: close-range air combat, intelligent decision-making, deep reinforcement learning, human-machine combat, flight test

CLC Number: