Cooperative Spectrum Sensing for the Offshore Cognitive Ship Ad Hoc Network

Cooperative spectrum sensing by maritime ad hoc networks suffers from low accuracy owing to signal attenuation and variations in node distances depending on the sea state, self-organizing characteristics, and ship movement patterns. To address this problem, we develop a physical model for an offshore cognitive ship ad hoc network (OCSAN) that incorporates factors such as ship movement, theoretical communication ranges, and maritime signal-tointerference ratios (SIRs). Based on the model, a cooperative communication strategy for OCSAN is designed along with a hedge algorithm for online learning combined with a soft loss function (Hed-SLC) for cooperative spectrum sensing in scenarios involving ship movement and complex sea states. Hed-SLC utilizes the historical energy detection statistics of nodes to measure the quantized increments of each node and dynamically adjust the weight values online. Soft information fusion statistics are then combined so that a final sensing decision can be made. Simulations are conducted to evaluate the effectiveness of Hed-SLC in different scenarios, and the results show its superior detection performance, adaptability, and robustness to different sea states and ship movements compared to those of traditional algorithms.