Complex QR Decomposition Using Fast Plane Rotations for MIMO Applications

QR decomposition (QRD) is widely used in various engineering applications and its implementation has a significant impact on the system performance and complexity. This paper develops a low-complexity QRD algorithm based on fast plane rotations, which does not require square-root operations for decomposing a complex-valued matrix. Furthermore, an update-based implementation is presented where computations are performed incrementally as the data arrives sequentially in time to drastically reduce the overall latency and hardware resources. Practical results for QRD-based spatial correlation estimator are provided to demonstrate the effectiveness of our solution for multiple-input multiple-output (MIMO) systems with complex-valued signals.