Method for applying deep learning algorithm in multi-fighter coordinated airspace exploration

Abstract

The invention relates to a method for applying a deep learning algorithm in multi-fighter collaborative airspace exploration, which comprises the following steps of: acquiring environmental data from a sensor by adopting a six-layer CNN (Convolutional Neural Network) to extract features, enabling a plurality of fighters to share the same PPO network for learning, and then selecting and implementing corresponding flight actions. The method is applied to a scene of collaborative exploration of multiple fighters; the time required for exploration is reduced, and the income of finding the enemy plane and the risk of being found by the enemy plane are properly balanced; and through a deep reinforcement learning algorithm, the efficiency of collaborative exploration of the multiple fighter planes is maximized.

Description

technical field [0001] The invention relates to a method for applying a deep learning algorithm in multi-fighter cooperative airspace exploration. Background technique [0002] In the important air combat mode of fighter-plane coordinated operations, how to effectively carry out information fusion of different fighters and multiple sensors has a great effect on integrating comprehensive information on the battlefield and improving systemic combat capabilities. The essence of this information fusion is the algorithm problem of information exchange and maintenance between fighters (Agents). [0003] Existing schemes are based on Brian Yamauchi's (1998) algorithm for multi-fighter exploration via boundaries first proposed. This study upgrades the traditional exploration mode of a single fighter in an unknown area with obstacles to a boundary-based multi-fighter cooperative algorithm. According to this scheme, each fighter maintains a part of the global map and makes explorati...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a method for the construction of a deep learning algorithm applied in multi-fighter cooperative airspace exploration, and an algorithm model for mutual cooperation between multiple fighters. Through effective information sharing, the fighters can maximize the exploration efficiency, so that in the Gain the first advantage in application scenarios such as enemy reconnaissance to solve many technical problems in the existing technology

Images