RDN Assisted Signal Detection for Adaptively Biased Optical OTFS System

The development of free-space optical (FSO) communication systems requires advanced techniques to mitigate effects induced by atmospheric turbulence. This paper explores an adaptively biased optical orthogonal time frequency space (ABO-OTFS) modulation scheme to address atmospheric turbulence challenges in FSO communication systems. Through adaptive bias optimization, ABO-OTFS demonstrates superior bit error rate (BER) performance compared to direct current biased optical OTFS systems. Theoretical analysis derives closed-form average BER expressions for ABO-OTFS under turbulence Málaga channels. Building upon the theoretical foundation, this paper proposes a residual dense network (RDN) assisted signal detection designed for the ABO-OTFS scheme. Through comparing RDN with maximum likelihood detection under Málaga turbulence conditions without channel state information, the results demonstrate that RDN achieves superior BER performance at high signal-to-noise ratios. Notably, the RDN model trained offline shows feasibility and generalization ability when deployed in realistic scenarios.