A secure transmission method against intelligent attacks based on reinforcement learning

Abstract

The invention discloses a security transmission method against intelligent attacks based on reinforcement learning, including: S1, through the channel parameter matrix signal of the system, estimate the link parameters of the main channel, interference and spoofing channels, and then respectively maximize the main channel Signal-to-noise ratio, minimize interference and spoof the signal-to-noise ratio of the channel, realize the design of the beamformer at the sending end and the filter at the receiving end, and suppress the attacker with a specific attack mode and a specific number of antennas; S2, in the beamformer and filter On the basis of S3, the zero-sum game is carried out, and based on the power control algorithm of reinforcement learning, intelligent confrontation is launched between the sender and the attacker; S3 dynamically adjusts the number of main channel sending and receiving antennas, and applies it to the beamformer And filter, zero-sum game process, effectively suppressing the attacker with smart attack mode and number of smart antennas; the invention can suppress the attack intention of the smart attacker, reduce the attack probability, and improve the system data transmission rate.

Description

technical field [0001] The invention relates to the field of network communication security, in particular to a method for secure transmission against intelligent attacks based on reinforcement learning. Background technique [0002] As a rapidly developing technology, artificial intelligence has been applied in various fields, such as face recognition, water level monitoring, etc. Recently, the application of artificial intelligence technology to the field of wireless communication has attracted the attention of many researchers. In many cases, an agent is not limited to identifying and classifying objects, but also needs to respond immediately to the state of the environment and take appropriate actions automatically. Reinforcement learning tasks are usually considered as a Markov decision-making process: the agent performs an action at the current moment, and the environment gives a reward to the agent according to the reward function. Continuous trial and error and explo...

AI Technical Summary

Problems solved by technology

The conventional strategies in the past can only deal with the situation that the number of attacking antennas remains unchanged, and the number of antennas of the sender is much larger than that of the attacker. However, in actual situations, the number of antennas of the sender and the attacker is likely to be the same. Solving this problem is the basic motivation for invention

Images