We consider space division multiplexing in a MIMO-OFDM system for high data rate transmission. Channel estimation is very important for suppressing interference and demultiplexing signals. In a wireless LAN system such as IEEE 802.11a, only a few training symbols are inserted in each subcarrier. First, we propose a channel estimation method for a MIMO-OFDM system with two training symbols per subcarrier. The basic idea is to estimate the time-domain channel responses between the transmit and receive antennas. The array response vectors for each subcarrier are calculated by applying a fast Fourier transform to them. We then can obtain the adaptive weights to cancel the interference. We show that employing training symbols having a lower condition number of the matrix used for the channel estimation improves the estimation accuracy. Furthermore, we show the bit error rate for several signal detection schemes using the above estimated channel. It is shown that an ordered successive detection based on an MMSE criterion has excellent performance, that is, it can achieve higher-speed transmissions with a lower transmit power.

This paper presents a simplified maximum likelihood detection (MLD) scheme for multiple-input and multiple-output spatial division multiplexing (MIMO-SDM) systems. In the scheme, ordered successive detection (OSD) is applied to multiple symbol candidates retained in the preceding stage to limit the number of symbol vector candidates. Accordingly, the subsequent MLD searches for the most likely signal vector among the limited symbol-vector candidates. Simulation results demonstrated that the proposed scheme provides the bit error rate performance close to that achieved by MLD while reducing the computational complexity.

Multiple-input multiple-output (MIMO) systems can improve the spectral efficiency of a wireless link, by transmitting several data streams simultaneously from different transmit antennas. However, at the receiver, multi-stream detection is needed for extracting the transmitted data streams from the received signals. This letter considers ordered successive detection (OSD) for multi-stream detection. OSD consists of several stages, and at each stage only one data stream is chosen to be detected among the remaining streams according to a specified ordering metric. OSD has been formulated using both the zero forcing (ZF) and minimum mean square error (MMSE) criteria. This letter clarifies the reason behind the superiority of OSD using the MMSE criterion to OSD using the ZF criterion through the investigation of the relation between their ordering metrics. For uncorrelated MIMO channels, we show that both ordering metrics yield the same performance for OSD using either ZF or MMSE criterion. Accordingly, the superiority of OSD using the MMSE criterion to OSD using the ZF criterion is clarified to be a direct result of the inherent superiority of MMSE nulling to ZF nulling, and to be independent of the ordering operation. Performance comparisons of OSD and maximum likelihood detection are also given for modulation schemes of different sizes.

To increase the spectral utilization efficiency of a wireless link, multiple-input multiple-output (MIMO) systems can be employed to transmit several data streams in parallel at the same time and on the same frequency but from different transmit antennas. However, at the receiver side multi-stream detection is needed. In this paper, ordered successive MMSE detection (OSD) is considered as a low-complexity detection scheme. OSD's main computational cost lies in computing the nulling weights that correspond to each stage of successive detection. In this paper, we develop an efficient semi-adaptive approach to generate MMSE weights. This semi-adaptive approach efficiently combines two approaches: channel estimates-based direct matrix inversion weights generation (direct approach) and Recursive Least Squares (RLS) algorithm-based weights generation (adaptive approach). Although the direct approach alone performs better than the adaptive approach, it is more complex for updating weights within the tracking mode. On the other hand, the adaptive approach alone is less complex in updating weights within the tracking mode, but converges slowly within the training mode. Our combined semi-adaptive approach effectively offsets these disadvantages. We demonstrate, through computer simulations, that the semi-adaptive approach can achieve the BER of the direct approach in slow time-varying MIMO channels, while its computational complexity is less than or comparable to that of the adaptive approach.

Ordered successive MMSE detection (OSD) can achieve a good tradeoff between performance and complexity for multiple-input multiple-output (MIMO) wireless systems where different data streams are transmitted and received simultaneously over several equal number of transmit and receive antennas. However, over time-varying MIMO channels, the performance of OSD is degraded with realistic channel estimation. In this paper, we introduce two different approaches in order to improve the performance of OSD. First, we propose to improve the accuracy of the generated nulling weights by using a modified noise variance that takes into consideration the weights update lag error besides to the noise and channel estimation errors. Second, we introduce a new iteration approach in order to improve the decisions made by OSD on transmitted streams. The proposed iterative scheme, named as backward iterative detection, exploits the tentative decisions on lately detected streams in order to mainly improve the decision on the firstly detected stream that limits the overall performance. Simulation results of combining both approaches show significant performance improvements of OSD.