Variational Bayesian Detection for Sparse AFDM-IM in High-Doppler Wireless Channels

In this paper, we propose a Variational Bayesian (VB) based sparse signal detection framework for Affine Frequency Division Multiplexing with Index Modulation (AFDMIM), tailored for high-Doppler wireless communication scenarios. AFDM, a chirp-based multicarrier waveform, offers inherent resilience to rapidly time-varying channels, and when combined with index modulation, further enhances spectral and energy efficiency. The sparse nature of the AFDM-IM signal is exploited using a hierarchical Bayesian model, where the VB inference method is used to iteratively estimate the transmitted symbols. An energy detection (ED) based post-processing technique is then employed for subcarrier index recovery. Simulation results demonstrate that the proposed VB-based detector significantly outperforms conventional MMSE detection methods in terms of bit error rate (BER) under severe Doppler conditions, and it maintains reliable performance across various modulation schemes and sparsity levels. These findings highlight the potential of VB learning for robust and efficient AFDM-IM signal recovery in future 6 G wireless systems.