Multi-Dimensional Multiple Access-Based Hybrid Semantic-Bit Communications for Individualized QoS Provisioning

Future wireless networks are expected to support massive heterogeneous devices with distinct capabilities in terms of semantic communications (SemCom) and bit-based communications (BitCom). However, conventional system designs face two fundamental limitations in enabling the coexistence of SemCom and BitCom users. First, existing schemes neglect the limited interference cancellation capability of BitCom users due to hardware constraints and reliance on single-domain resource utilization, leading to ineffective interference management. Second, conventional capacity-based metrics fail to capture user-specific service requirements and device heterogeneity, causing inefficient and non-adaptive resource utilization. To overcome these challenges, we propose a multi-dimensional multiple access-based hybrid semantic-bit communication (MDMA-HSBC) framework with two major contributions: (i) a dynamic switching mechanism at the base station that supports adaptive operation between SemCom and BitCom modes for dual-capable devices and enables the selection of the optimal access scheme for each user across multiple resource domains to ensure efficient interference management; and (ii) a cost-aware performance metric that jointly optimizes access schemes and communication modes to incorporate user-specific service demands and outperform capacity-driven designs. The resulting nonconvex problem is efficiently addressed through a heuristic matching algorithm, achieving significant performance gains over conventional baselines.