A Novel Architecture of Sphere Decoder for Low Complexity and High Throughput

Since finding the nearest point in a lattice for multi-input multi-output (MIMO) channels is NP-hard, simplified algorithms such as a sphere decoder (SD) have been proposed. In this paper, we propose a low area and high throughput sphere decoder based on the well-known one-node-per-cycle architecture and present the implementation result. Three key contributions are a new direct Schnorr-Euchener (SE) enumeration scheme for QAM modulation, an efficient storage manipulation and the calculation interleaving to reduce the critical path. Compared with current state-of-the-art sphere decoder architecture, the proposed sphere decoder can save 48% of hardware complexity and enhance 54% of the maximum operation clock frequency.