A Proximal Policy Optimization Approach for Shortwave Detection and Localization

The detection and localization of shortwave signals is of great importance for long-distance communication. The complexity and variability of the channel result in the detection and localization of signals often relying on human effort. We provide an approach that uses the Proximal Policy Optimization (PPO) model to create an intelligent signal detection and localization method based on the PPO algorithm, thereby reducing the workload associated with signal detection and localization. We have created an agent window that searches the time-frequency spectrum for signals before obtaining the localization result using a localization technique. Experiments demonstrate that the PPO algorithm may converge swiftly and steadily to accomplish our objective when compared to another reinforcement learning algorithm, Deep Deterministic Policy Gradient (DDPG).