A Transformer Inspired AI-based MIMO receiver
This addresses MIMO detection for 5G communication systems, offering a hybrid approach with interpretability and flexibility, though it appears incremental as it adapts an existing Transformer paradigm to a specific domain.
The authors tackled MIMO detection in 5G systems by proposing AttDet, a Transformer-inspired method that treats transmit layers as tokens and uses self-attention to learn inter-stream interference. The result showed that AttDet approaches near-optimal BER/BLER performance under realistic 5G channel models while maintaining polynomial complexity.
We present AttDet, a Transformer-inspired MIMO (Multiple Input Multiple Output) detection method that treats each transmit layer as a token and learns inter-stream interference via a lightweight self-attention mechanism. Queries and keys are derived directly from the estimated channel matrix, so attention scores quantify channel correlation. Values are initialized by matched-filter outputs and iteratively refined. The AttDet design combines model-based interpretability with data-driven flexibility. We demonstrate through link-level simulations under realistic 5G channel models and high-order, mixed QAM modulation and coding schemes, that AttDet can approach near-optimal BER/BLER (Bit Error Rate/Block Error Rate) performance while maintaining predictable, polynomial complexity.