Jianxuan Du's research while affiliated with Georgia Institute of Technology and other places

We propose a modified layered structure for multiple-input multiple-output systems, where the layer detection order is fixed and the data rate for each layer is allocated based on the detection order and channel statistics. Using a Gaussian approximation of the layer capacities, we derive an asymptotic optimum data-rate-allocation approach. For optimum data-rate allocation, the amount of backoff from the mean layer capacity is proportional to the standard deviation of the layer capacity. With statistical data-rate allocation, only limited channel feedback is needed to update channel statistics at the transmitter. Simulation results show significant performance improvement with the proposed algorithm. We also find that the performance gap between the layered structure and the channel capacity diminishes with increasing ergodicity within each codeword. Numerical results show a singal-to-noise ratio improvement of 6.3 and 3.6 dB for TGn channels "B" and "D," respectively, for 1% outage probability and 9 b/s/Hz spectral efficiency

Multiple-input multiple-output (MIMO) techniques promise huge capacity increase in fading channels. Groupwise space-time coding is a tradeoff between complexity and performance. In this letter, the structure of particular component space-time code trellises is exploited, using partial information from a Viterbi decoder of the simultaneously decoded interfering component codes. We have shown that by calculating the a posteriori probabilities of different states of the encoder at each time, and predicting the conditional mean and covariance of interference for the next time index, the system performance can be improved by soft interference cancellation. The word-error rate performance improvement for a system with four transmit and four receive antennas is 1.7 and 1.5 dB for two space-time codes with 8 and 16 states, respectively. In addition, the parallel structure does not have a decoding delay, as in successive interference cancellation, and is highly modular for very-large-scale integration implementation

In this letter, we investigate the issue of selecting the number of antennas at the base station and at the mobile to optimize ergodic capacity of a multiple-input/multiple-output system, when the costs of implementing antennas at the base station and at the mobile are unequal. Total system capacity, defined as a linear combination of the uplink and downlink ergodic capacity, is used as the objective function to be maximized. The asymptotic expression for the ergodic capacity is used as an approximation. The limiting case gives some insight on how the ratio of the number of antennas at the base station to the number of antennas at the mobile may change with signal-to-noise ratio and cost ratio when the total system capacity is maximized.

We propose a modified layered structure for multiple-input multiple-output (MIMO) systems, where the layer detection order is fixed and the data rate for each layer is allocated based on the detection order and channel statistics. With Gaussian approximation of layer capacities, we derive the optimum data rate allocation and the amount of backoff from mean layer capacity is proportional to the standard deviation of the layer capacity. The minimum overall outage probability of a layered system is uniquely determined by the normalized capacity margin. We then investigate how to select the total information rate to maximize effective throughput. Simulation results show significant performance improvement with the proposed algorithm, and the performance gap between layered structure and the channel capacity diminishes with increasing ergodicity within each codeword.

In this paper, we investigate the issue of selecting the number of antennas at the base station and at the mobile to optimize asymptotic ergodic capacity of a multiple-input-multiple-output (MIMO) system when the costs of implementing antennas at the base station and at the mobile are unequal. Total system capacity, defined as a linear combination of the uplink and downlink ergodic capacity, is used as the objective function to be maximized. The limiting case gives some insight on how the ratio of the number of antennas at the base station to the number of antennas at the mobile may change with signal-to-noise ratio (SNR) and cost ratio when the total system capacity is maximized.

Latency of available channel state information (CSI) at the transmitter in time-varying channels greatly affects the performance of multiple-input multiple-output (MIMO) systems. We have derived a simple algorithm to calculate an approximation of the expected performance loss based on the most recent channel feedback. The proposed algorithm can be used to determine the maximum tolerable channel feedback delay for each particular channel realization.

Multiple-input-multiple-output (MlMO) techniques promise huge capacity increase in fading channels. Group-wise space-time coding is a tradeoff between complexity and performance. In this paper, structure of particular component space-time code trellises is exploited using partial information from Viterbi decoder of the simultaneously decoded interfering component code. We have shown that calculating the a posteriori probabilities of different states of the encoder at each time, and predicting the conditional mean and covariance of interference for the next time index, the system performance is improved by soft interference cancellation. The word-error-rate (WER) performance improvement for a system with 4 transmit 4 receive antennas is 1.7 dB and 1.5 dB for 2-space-time codes with 8 and 16 states, respectively. In addition, the parallel structure does not have decoding delay as in successive interference cancellation and is highly modular for VLSI implementation.

Multiple-input and multiple-output (MIMO) systems formed by multiple transmit and receive antennas can improve performance and increase capacity of wireless communication systems. Diagonal Bell Laboratories Layered Space-Time (D-BLAST) structure offers a low-complexity solution for realizing the attractive capacity of MIMO systems. However, for broadband wireless communications, channel is frequency-selective and orthogonal frequency division multiplexing (OFDM) has to be used with MIMO techniques to reduce system complexity. In this paper, we investigate D-BLAST for MIMO-OFDM systems. We develop a layerwise channel estimation algorithm which is robust to channel variation by exploiting the characteristic of the D-BLAST structure. Further improvement is made by subspace tracking to considerably reduce the error floor. Simulation results show that the layerwise estimators require 1 dB less signal-to-noise ratio (SNR) than the traditional blockwise estimator for a word error rate (WER) of 10-2 when Doppler frequency is 40 Hz. Among the layerwise estimators, the subspace-tracking estimator provides a 0.8 dB gain for 10-2 WER with 200 Hz Doppler frequency compared with the DFT-based estimator.

In this paper, we propose a channel estimation algorithm for multiple-input and multiple-output orthogonal frequency for division multiplexing (MIMO-OFDM) systems, which has considerably less leakage than DFT-based channel estimators. This algorithm uses the optimum low-rank channel approximation obtained by tracking the frequency autocorrelation matrix of the channel response. The coefficients corresponding to dominant eigenfactors of the autocorrelation matrix are estimated every OFDM block while the eigenfactors are only updated using the training block that is transmitted every M blocks due to the slowly-varying feature of the frequency autocorrelation. Simulation results show that the proposed algorithm can effectively reduce channel estimation error and thus improve system performance.

The Diagonal Bell Laboratories Layered Space-Time (D-BLAST) structure offers a low complexity solution to realize the attractive capacity of multiple-input and multiple-output (MIMO) systems. We apply D-BLAST in orthogonal frequency division multiplexing (OFDM) systems and address the issue of channel estimation. Different from other MIMO-OFDM, where symbols at all tones are always available for decision-directed channel estimation, in D-BLAST OFDM, we update estimated channel parameters each time a layer is detected with a least square (LS) approach, using a pieced combination of received signals at previous and current OFDM blocks. The initial estimate is further refined by a robust estimator to exploit the time correlation of channel parameters among OFDM blocks. Computer simulation results show the performance improvement over block-wise channel estimation. It is also shown that D-BLAST with proposed channel estimation is robust to fast fading of channel parameters.