List Sequential MIMO Detection: Noise Bias Term and Partial Path Augmentation

Improved Length Term Calculation and MMSE Extension for LISS MIMO Detection

“…The performance of an MAlgorithm/ITS based detector [13] with the same maximum number of node extensions is plotted as reference (stack size M = 11, e IT S = 1 + (M T − 1) · M = 34). 6 6.5 For the ZF-based LISS, the best performance is achieved when using the noise bias as length term. For the MMSE case, the LISS using a Babai bias again outperforms the LISS employing the noise bias.…”

“…Again, the noise bias provides the best performance-complexity trade-off in the ZF case while in the MMSE case the Babai bias should be used. 6 6.5 The lower performance of the ZF-based LISS using the Babai bias is due to the fact that the search tree is centered on the ZF-Babai solution in this case, while when using the noise bias it is still centered on the ML solution (only an estimate for the metric accumulated by the ML path is subtracted from the metrics). The relevance of centering the tree search on the ML solution has been stressed in [14] in the context of sphere decoding, and we observe the same behavior here in the context of list sequential detection.…”

“…Furthermore, we extended the ZF-based LISS which has so far been studied [5], [6] to the MMSE case and showed that this approach enables a far better performance-complexity trade-off. Particularly, using MMSE preprocessing enables the LISS with bounded complexity to achieve better performance than M-Algorithm based MIMO detection having the same maximum allowed complexity.…”

“…The a priori knowledge is taken into account, but the effort required for calculation of the bias term is still substantial. In [6] is was therefore suggested to approximate the bias term by the average accumulated noise energy over the receive antennas. Due to the normalization of the Euclidean path increment with N 0 , this noise bias term requires only an increment of 1 per detected antenna.…”

Low-Complexity Decoding via Reduced Dimension Maximum-Likelihood Search

“…By using an appropriate value for (such as through empirical simulations), we can achieve substantial reduction in back-tracking operations with negligible performance loss. A similar approach has been applied to the SD algorithm in [17] and to SA in [36]. Our approach differs from that of [36] in that rather than using the expected noise power to obtain a bias term, we employ a probabilistic condition (34) to choose .…”