Low Complexity MIMO-OFDM Receivers for Achieving Near Optimal Performance
No Thumbnail Available
Date
2011-07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Since the useable frequency spectrum is a scarce resource, future wireless systems should focus on
new technologies that provide high spectral efficiency and improved link reliability. The
combination of Multiple Input Multiple Output (MIMO) wireless communication with
Orthogonal Frequency Division Multiplexing (OFDM) is a very promising technology to meet
these requirements.
Although MIMO OFDM systems have demonstrated very high spectral efficiency, the available
capacities of these systems are not fully exploited due to lack of low complex and, at the same
time, optimal signal detection scheme. To this end, after reviewing the available detection
techniques, this thesis proposes a low complex MIMO signal detection technique which has near
optimal performance.
It is found that among the MIMO signal detection techniques the Maximum Likelihood Decoding
(MLD) has optimal performance; however, the corresponding computational complexity grows
exponentially with the number of antennas and/or constellation size. The conventional detection
schemes such as ZF, MMSE, ZF with SIC and MMSE with SIC have relatively manageable
complexity; however, in general their BER performance is poor especially for large MIMO sizes.
The sphere decoding algorithms (SDAs), specifically the depth first sphere decoding (DF SD),
achieves similar performance to MLD at relatively reasonable complexity, but it has very large
variation of computational throughput. In order to reduce the complexity of DF SD, initial radius
setting strategy is introduced in this thesis and, with this, enormous computational complexity is
saved. Fixed throughput SDAs, namely k best SDAs, are also analyzed and as a result, it is
obtained that these schemes have performance degradation compared to the MLD. So, a hybrid
sphere decoding which combines the desirable features of the two SDAs is proposed in this thesis
and simulation results show that the proposed scheme can achieve performance very close to the
MLD at very high computational throughput which is more than twice of that of k best SDAs.
Description
Keywords
MIMO-OFDM Signal Detection, Sphere Decoding Algorithms, Hybrid Sphere Decoding