Intelligent Anti-jamming Algorithm Based on Time-frequency Domain Joint

In view of the poor ability of terahertz communication network to resist intelligent jamming, based on the model of multi-arm gambling machine (MAB) and reinforcement learning theory, this paper studies the problem of channel selection and transmission time scheduling. Firstly, based on the advantage actor-critic algorithm, the marine terahertz communication ship is regarded as an agent, the perceived spectrum state is used as a priori information to complete the secure channel decision; Then, according to the return parameters of the MAB game model, the length of time to access the channel is selected. Simulation results show that, compared with the fixed transmission duration algorithm based on single reinforcement learning mode, the proposed algorithm has obvious improvement in accuracy, stability and practicability.