Integrated MAC and Physical Layer Design for a High Speed Multi-Antenna OFDM WLAN (1/3)
Date Issued
2005-07-31
Date
2005-07-31
Author(s)
DOI
932219E002015
Abstract
Wireless LAN has enjoyed a growing popularity in recent years due to its unlicensed, low-cost and
high-rate nature compared to cellular networks. This success has triggered a change in the end-user
demand toward high quality, rate-hungry multimedia applications such as wireless high definition digital
videos and games. As a result, the current 2.4 GHz and 5 GHz wireless LAN bands are becoming
inadequate to fulfill the demand, partly due to insufficient bandwidth, but also due to the inefficient
medium access control (MAC) protocols that are being used in the current wireless LAN. In view of this,
it is the purpose of this project to investigate the physical layer (PHY) and MAC issues involved in
deploying wireless LANs with wideband transmission in the 60 GHz band where ample unlicensed
bandwidth is available. In addition to wide transmission bandwidth, multiple-input multiple-output
(MIMO) antenna techniques will be used to further boost the transmission rate. To ease the complexity of
equalization in such a wide bandwidth, orthogonal frequency division multiplexing (OFDM) is adopted.
A coordinated, cellular-like network structure is considered to enhance the MAC throughput and
frequency reuse.
In the first year of the project, some PHY and MAC issues have been considered for the proposed
MIMO OFDM system. Specifically, we considered two multiplexing schemes for multiple users to share
the downlink channel capacity increased by MIMO: grouped transmission, and overlapped transmission.
In the former, for simplicity, the available transmission antennas are divided into a few groups, and each
user is assigned a group of antennas. For such a scheme, each receiver only receives useful data from its
corresponding transmission antennas, therefore it may know only the channel gains from these antennas
through training. In order to achieve a good detection performance, the receiver will need an adaptive
mechanism to measure the statistics of the interference from the other transmission antennas in order to
suppress it. We investigate in this report an adaptive maximum signal-to-noise ratio (SNR) filter bank
receiver for such a grouped transmission. The overlapped transmission is the optimal capacity sharing
scheme in terms of maximizing the total transmission rate. For such a scheme, all users use all the
transmission antennas, thus interfering with one another. We apply the Dirty Paper Coding (DPC) method
to suppress mutual interference among users. The optimal design parameters for Vector (i.e., MIMO)
DPC are derived for a specific superposition coding realization.
To improve the network-wide throughput, a power control and allocation method was adopted for
multi-cell OFDM networks to achieve interference avoidance between cells. Such a method relies on a
simple power allocation mechanism autonomously run by each co-channel cell to achieve multi-cell water
filling. As the mechanism runs, each cell will snatch sub-carriers that are good for it, and gradually
release the sub-carriers with bad channel conditions to reduce interference to other cells. Such a method
has another inherent benefit of lower peak-to-average power ratio as compared to conventional equal
power allocation method. The advantage is also investigated in this report.
Finally, as adaptive modulation and coding (AMC) and Hybrid ARQ (HARQ) are becoming popular in
new high speed wireless network standards for their ability to maximize the end-to-end throughput, we
will also consider them in the proposed MIMO OFDM system. In the first year of the project, we
proposed a simple adaptive method to find the optimal parameters for operating AMC and HARQ. This
method is very general that it can be easily ported to any PHY structure, including MIMO OFDM.
Subjects
WLAN
60 GHz
OFDM
MIMO
diversity
smart antenna
MAC
Hybrid ARQ
adaptive
modulation and coding (AMC)
modulation and coding (AMC)
power control
scheduling
multiple access
dirty paper coding
Publisher
臺北市：國立臺灣大學電信工程學研究所
Type
report
File(s)![Thumbnail Image]()
Loading...
Name
932219E002015.pdf
Size
689.79 KB
Format
Adobe PDF
Checksum
(MD5):54786713c8fb1d659250e808a2b48048