Multi-unmanned aerial vehicle cooperative confrontation method and system, and storage medium

Abstract

The invention discloses a multi-unmanned aerial vehicle cooperative confrontation method and system and a storage medium, and the method comprises the steps: building a Markov game model under the condition of our multi-unmanned aerial vehicle cooperative confrontation based on unmanned aerial vehicle information in the multi-unmanned aerial vehicle cooperative confrontation system, and carrying out the training of the Markov game model based on MADDPG, an improved MADDPG is used for setting a group of actuator network and evaluator network for each our unmanned aerial vehicle, a centralized training and decentralized execution strategy is utilized, the evaluator network for acquiring global state information is given to guide the our unmanned aerial vehicle to train during training learning, an actuator for outputting specific maneuvering actions only acquires local observation state information, and the optimal maneuvering action is realized. Under the guidance of an evaluator network, proper network parameters are gradually converged, so that the autonomous decision-making ability of the unmanned aerial vehicle in confrontation is improved in an unknown and dynamic confrontation environment, and the method is high in convergence speed and short in decision-making time.

Description

technical field [0001] The embodiments of the present application relate to the technical field of unmanned aerial vehicles, and in particular, to a method, a system and a storage medium for cooperative confrontation of multiple unmanned aerial vehicles. Background technique [0002] During the development of air combat, people realized that the coordinated operation of multiple fighter jets can improve combat efficiency. In the early stage, single-aircraft air combat has gradually developed into two-aircraft formation, multi-aircraft formation and other modes, and the combat effectiveness has been continuously enhanced. Theoretically, the increase in the number of fighters provides more firepower output points and a wider observation field, and has a stronger advantage in both offensive and defensive missions. [0003] If the number of fighters performing the same task in the same area is too large, some defects will inevitably occur. For example, the larger the fighter gro...

AI Technical Summary

Problems solved by technology

[0005] For our drones that cooperate against enemy targets, each of our drones detects and communicates with its own onboard sensors to receive information. If our drones that cooperate against targets can maintain stable information communication, The optimal decision can be made by concentrating global information, but this method has two main drawbacks. One is that it assumes that the battlefield electromagnetic environment meets the requirements of stable and reliable communication, which is too ideal and unrealistic. Now that electronic warfare has entered the actual battlefield, electronic interference Technologies such as suppression, deception, etc. emerge in endlessly. The UAV platform expects to grasp the overall information at all times, but this does not conform to the actual scene

Second, as the number of UAVs participating in the confrontation increases, the amount of information to be processed increases rapidly, which puts high demands on the complexity of the decision-making algorithm, the computing power of the onboard computer, and the carrying capacity of communication transmission. A large amount of calculation may lead to long decision-making time or algorithm non-convergence

Images