Skip to Main Content
In this study, the authors investigate the expected complexity of increasing radii algorithm (IRA) in an independent and identified distributed Rayleigh fading multiple-input-multiple-output channel with additive Gaussian noise and then present its upper bound result. IRA employs several radii to yield significant complexity reduction over sphere decoding, whereas performing a near-maximum-likelihood detection. In contrast to the previous expected complexity presented by Gowaikar and Hassibi (2007), where the radius schedule was hypothetically fixed for analytic convenience, a new analytical result is obtained by considering the usage of multiple radius schedules. The authors analysis reflects the effect of the random variation in the radius schedule and thus provides a more reliable complexity estimation. The numerical results support their arguments, and the analytical results show good agreement with the simulation results.