Configurable Quasi-Optimal Sphere Decoding for Scalable MIMO Communications
Published in IEEE Transactions on Circuits and Systems I: Regular Papers, Vol. 68, Issue 6, 2021
This paper addresses the challenge of balancing detection accuracy and hardware cost in Sphere Decoders (SD) for Multiple-Input Multiple-Output (MIMO) communication systems, particularly in constrained environments like IoT. The authors introduce a Robust Bounded Spanning with Fast Enumeration (R-BSFE) method that enhances channel matrix pre-processing and symbol enumeration to maintain quasi-maximum-likelihood (ML) accuracy while reducing computational complexity by up to 74%.
Key contributions include:
Robust Bounded Spanning with Fast Enumeration (R-BSFE): Novel strategies for pre-processing and symbol enumeration that reduce complexity and maintain quasi-ML detection accuracy.
High-Performance FPGA Implementation: Design of accelerators for 802.11n on Xilinx FPGA achieving lower cost and higher throughput.
Cost-Accuracy Trade-off: Provides configurable SD designs to balance detection performance against accelerator cost, suitable for power-constrained MIMO systems.
State-of-the-Art Results: Demonstrated as the highest performance, lowest cost quasi-ML SD accelerators reported in literature.
For further details, refer to the full paper: https://ieeexplore.ieee.org/document/9375712
Recommended citation: Wu, Y., & McAllister, J. (2021). Configurable Quasi-Optimal Sphere Decoding for Scalable MIMO Communications. *IEEE Transactions on Circuits and Systems I: Regular Papers*, 68(6), 2021–2032. https://doi.org/10.1109/TCSI.2021.9375712
Download Paper | Download Slides | Download Bibtex