Sam P. Alex

Performance Evaluation of MIMO in IEEE802.16e/WiMAX

The methodology used to evaluate the performance of the various MIMO schemes in orthogonal frequency division multiplexing (OFDM) WiMAXWiMAX stands for Worldwide Interoperability for Microwave Access. systems is described. The per-tone signal-to-noise plus interference ratio (SINR) of a MIMO OFDM system is derived in a multi user, multicell and multisector communication system. The sector average spectral efficiency for a coded MIMO OFDM system is evaluated under a single frequency reuse deployment scenario. It is shown that the second antenna at the subscriber station receiver provides significant gains over single transmit single receive antenna (SISO) systems. The rate 2 spatial multiplexing MIMO scheme on the downlink improves the sector spectral efficiency by 10% over single transmit and two receive antenna (SIMO) system in a single frequency reuse deployment. It is also shown that in a time-division duplex (TDD) WiMAX system, the gain in downlink spectral efficiency from beamforming can be very large if the delay between the uplink channel estimate and the downlink beamforming is small compared to the coherence time of the channel. On the uplink, it is shown that the collaborative spatial multiplexing increases the spectral efficiency by 9% over a SIMO system with the same number of receive antennas at the base station (BS). Significant gains in uplink spectral efficiency are seen with more antennas at the BS and by using uplink beamforming as compared to SISO.

Reduced Complexity Soft-Output MIMO Sphere Detectors—Part I: Algorithmic Optimizations

Optimum soft-output (SO) multiple-input multiple- output (MIMO) tree-search detection algorithms pose significant implementation challenges due to their nondeterministic processing throughput and high computational complexity. In this two-part work, we present extensive algorithmic and architectural optimizations of the sphere-decoding algorithm targeted at achieving practical tradeoffs between desired link performance and affordable computational complexity. The algorithmic optimizations in this part span the tree-search traversal scheme, leaf processing step, internal node-pruning and skipping step, child enumeration based on a state-machine, adaptive radius scaling for LLR clipping, QR-decomposition based on minimum cumulative residuals, and multitree configurations. The optimizations demonstrate that a 64-QAM SO MIMO detector for LTE is capable of attaining almost ML performance with an SNR loss of only 0.85 dB at 1% BLER by visiting at most 200 tree nodes.

Coverage Analysis for IEEE 802.16e/WiMAX Systems

This paper describes a frame work for evaluating the coverage performance of an orthogonal frequency division multiplexing (OFDM) system as defined in the IEEE 802.16e/WiMAX standard. Both the downlink and the uplink coverage are presented under a system with multiple cells and sectors. The spatial area coverage for a single frequency reuse and reuse of 3 are evaluated. Our results show that a WiMAX system can be uplink coverage limited for base station receivers with only two antennas. It is shown that there is a tradeoff between the uplink sector throughput and coverage performance. On the downlink, it is shown that a combination of cyclic delay diversity (CDD) with segmented frequency reuse of 3 significantly improves coverage performance. Also, shown is a comparison of downlink coverage performance between maximum ratio combining (MRC) and minimum mean square (MMSE) receivers at the subscriber station. It is shown that for a single frequency reuse an MMSE receiver with CDD at the transmitter are needed for adequate downlink coverage performance.

Reduced Complexity Soft-Output MIMO Sphere Detectors—Part II: Architectural Optimizations

In Part I, we presented algorithmic optimizations for a soft-output (SO) tree-search MIMO detector targeted at reducing the node count complexity for a given error-rate performance. We present in this part the architecture for a hardware-efficient pipelined detector that incorporates all these optimizations. A detailed gate complexity analysis of the architecture is conducted, including arithmetic optimizations to minimize the use of multipliers for distance computations by exploiting the lattice structure, as well as efficient pointer logic for implementing child enumeration, internal node-pruning and node-skipping steps. For multitone OFDM symbol detection using multiple tree-search detectors, we propose an efficient scheduling scheme that exploits the slack in node count from tone-to-tone detection. By ordering the tones according to increasing order of their channel matrix orthogonality defect, the detector maximizes the slack of early terminating detectors and forwards it to ill-conditioned tones that require more detection time. Case studies and simulation experiments were conducted based on a 4-layer 64-QAM MIMO system for LTE. Simulations and VLSI synthesis results demonstrate that the proposed optimizations require a chip area of only 85 kGE to implement a 64-QAM SO MIMO detector for LTE capable of attaining almost ML performance with an SNR loss of only 0.85 dB.

Soft-Output MIMO Detectors with Channel Estimation Error

New expressions for the soft decision bit log-likelihood ratio (LLR) of a MIMO system using quadrature amplitude modulation (QAM) taking into account channel estimation error (CEE). The bit LLR for the maximum likelihood (ML) and the linear minimum mean-squared error (MMSE) receivers are derived, showing in both receivers explicit scaling of the LLR that is a function of the QAM symbol and the CEE variance. These new expressions for the LLRs are used to show that only modest improvements in the link performance are achieved relative to the LLRs that do not take into account the CEE. This indicates that separating the detector design from channel estimation does not significantly impact the system performance, which leads to simplifications in the overall receiver implementation.

Comments on “Soft Decision Metric Generation for QAM With Channel Estimation Error”

A log-max generalized bit log-likelihood ratio (GLLR) was derived in an earlier work of Wang et al. for soft decision decoding of quadrature amplitude modulation that takes into account channel estimation error. The results in the above paper show that the GLLR metric outperforms the conventional LLR metric that does not account for channel estimation error. In trying to reproduce the results of the above paper, our simulations indicate that their results were obtained without proper scaling of the transmit constellation. Conversely, through simulations, we demonstrate that by properly scaling the constellation to have unity average energy, the GLLR metric does not result in a noticeable advantage in the bit error rate over the conventional LLR metric for practically meaningful values of the channel estimation error.