Hybrid Deep Learning-Based Optimization for Efficient 6G Free-Space Optical Communication

With the introduction of 6G wireless communication, it is necessary to provide ultra-high reliability, high capacity and super-low latency. Free-Space Optical (FSO) solutions are main contributors to this. The changing weather conditions in the air decrease the performance of light signals sent by FSO. We recommend a new system that merges Social Spider Optimisation and Waterwheel Plant Optimisation to ensure that the U-net architecture provides robust signal detection under turbulent conditions in FSO channels. Instead of conventional or mixed designs, ours matches network frameworks to optical channel perturbations, allowing the network to resist changes in environmental conditions. Results from multiple simulation tests show that BER, latency, spectral efficiency and capacity all improve considerably, reaching 25 bps/Hz, 1 ms and 10⁶ linked users. This method supports the IEEE 6G viewpoint and prepares the way for dependable, broadly applied FSO systems for defense, surveillance and remote sensing.