Design of pre-processing algorithms for efficient MIMO–OFDM receiver architectures

Modern wireless communication applications are characterized by the need for advanced signal processing techniques such as Multiple-Input Multiple-Output (MIMO) technology for achieving high throughput and diversity and Orthogonal Frequency Division Multiplexing (OFDM) for achieving robustness to multipath fading. The implementation of such techniques at the transceiver level typically involves the design of algorithms with high processing complexity.This paper considers the efficient design of MIMO–OFDM receivers in preamble-based systems and addresses the problem of large processing delays associated with pre-computations and symbol detection. The existence of large processing delays has a huge impact on the performance and resource requirements (vector processing, increased clock rates and increased power consumption) of modern receivers. More specifically, we address the performance and complexity bottleneck introduced by the pre-computations involved for MIMO–OFDM channel decomposition. We propose a redesign of channel decomposition algorithms which achieves a better matching of the processing rate of MIMO–OFDM receivers to the real-time processing deadlines imposed by the structure of the incoming data packets. It is demonstrated that for a specific MIMO-OFDM channel training frame structure (alternating antenna preamble), simple modifications to typical channel decomposition algorithms can achieve significant processing performance and complexity gains compared to typical receiver designs.