Improved PMADDPG multi-unmanned aerial vehicle task decision-making method based on transfer learning

Abstract

The invention discloses an improved PMADDPG multi-unmanned aerial vehicle task decision-making method based on transfer learning. The method comprises the steps of firstly, performing reasonable mathematical modeling for threats such as air defense missiles and radars in an environment under the background of a multi-unmanned aerial vehicle actual combat environment, then creating a plurality of different two-dimensional combat environment models, designing constraint conditions, and learning and training the plurality of combat environments in sequence, thereby obtaining a final multi-unmanned aerial vehicle task allocation model. According to the method, the defect that in the prior art, task decision making can only be carried out in a known or static combat environment is overcome, decision making can also be carried out efficiently in an unknown combat environment, tasks of unmanned aerial vehicles are achieved, and the viability of an unmanned aerial vehicle group in the unknowncombat environment is greatly guaranteed.

Description

technical field [0001] The invention belongs to the field of flight control, and in particular relates to a method for multi-unmanned aerial vehicle task decision-making. Background technique [0002] For the military of various countries, drones will become one of the indispensable weapons in the future battlefield. UAVs are likely to become the target of attacks and counterattacks by multiple combat platforms, and become the most common and deadly air combat "sword". The cooperative combat method of multiple UAVs will become the mainstream development trend in the future. At present, military academies and scholars at home and abroad pay great attention to the research on multi-UAV mission decision-making problems, and there are many achievements. However, there are still many problems in the research of multi-UAV mission decision-making. For example, in multi-UAV cooperative search, tracking, task assignment, track planning, formation control and other issues, on the on...

AI Technical Summary

Problems solved by technology

Most of these traditional algorithms are proposed based on the optimization problem of a single objective function, and most of them are based on theoretical research, applicable to static combat environments, without considering various dynamic changes and constraints in actual air combat

On the other hand, although some UAV research results have introduced the method of deep reinforcement learning, the existing deep reinforcement learning algorithm takes a long time when dealing with multi-UAV task decision-making related issues, and the demand for real-time performance is still insufficient. needs improvement

And the generalization ability of the algorithm is far from enough, it can only perform well in the environment where it is trained, and the performance in a new environment is far from enough

Images