Author : Suneeta V. Budihal and R. M. Banakar

Maximum likelihood (ML) decoding is an optimal detector with high-performance for multiple-input-multiple-output (MIMO) communication systems. While it is attractive due to its superior performance in terms of BER, its complexity using an exhaustive search which grows exponentially with the number of antennas and order of the modulation. It becomes infeasible to apply to practical systems as it searches through all lattice points in the constellation. Sphere decoding (SD) is a promising method to reduce the average decoding complexity without compromising performance. It provides optimal performance with reduced complexity as it searches the points within the specified radius of sphere. The complexity of the sphere decoder depends on the initial radius selection of the sphere, to begin search process. Attention is drawn to initial radius selection strategy, since an inappropriate initial radius can result in either a large number of lattice points to be searched, or a large number of restart actions. The simulations are performed for constellation size of 4-QAM, 8-QAM and 16-QAM for antenna size of 2X2 MIMO. It is observed that the performance of Probabilistic Tree Pruning (PTP)-SD converges with ML by taking less time and maintaining the same performance. It is proposed that a Look Up Table (LUT) for initial radius Using Radius Choice Algorithm is generated. The complexity reduces by 60% as the number of FLOPS required reduces.

Keywords: Multiple Input Multiple Output, Maximum-likelihood decoding, Sphere Decoding, Sphere Decoding, Radius Choice algorithm.

Article published in the Proceedings of National Conference on ‘Women in Science & Engineering’ (NCWSE 2013), SDMCET Dharwad