10-04-2012, 05:16 PM
Multiuser Wireless Communications
Multiuser Wireless Communications.pdf (Size: 606.22 KB / Downloads: 92)
Abstract
In this project, a multiuser diversity scheme for MIMO-OFDM systems has been
proposed to improve system performance. In the space-time MIMO-OFDM system with
multiuser, it needs a scheduler at the base station to do the spatial multiplexing. The
scheduler distributes sub-channels to different users based on the feedback channel state
information (CSI). The goal is that each user determines a limited feedback CSI under his
data rate requirement and the scheduler dispatches sub-channels to each user according to
feedback CSI. The simulation results demonstrate that the proposed scheduler can
significantly improve the system performance with limited feedback CSI.
1. Introduction
In the past years, the research has proved that performance can be improved by
scheduling, power allocation, or adaptive modulation for CDMA, OFDM, and MIMO
systems. OFDM can support high data rate transmission over wireless channels by
multiplexing technique. In [1], several techniques have been discussed briefly to improve
system performance and spectral efficiency for multiuser OFDM. These include adaptive
modulation and adaptive user allocation. An adaptive modulation algorithm with bit and
power allocation for multiuser OFDM has been studied in [2]. The objective is to
minimize the total transmit power for given transmission rate of the users and given QoS
requirements. Another approach [3] for dynamic subchannel allocation was to increase
the capacity of multiuser OFDM system by maximizing the smallest capacity of all users.
Recent research also shows that the rich-scattering wireless channel is capable of
enormous theoretical capacities if the multipath is properly exploited [4-5]. The MIMO
wireless communication architecture, knows as V-BLAST [5], which deploys multiple
antennas at both transmitter and receiver had been demonstrated spectral efficiencies of
20 – 40 bps/Hz.
Due to high spectral efficiencies of OFDM, MIMO, and MIMO-OFDM systems, a
multiuser diversity scheme can be applied to these systems. In [6], the concept of
multiuser diversity has been discussed briefly which includes a fair scheduling for IS-856
system. To improve system performance, the basic idea of diversity is creating several
different and independent transmitting paths from the transmitter to the receiver. The
conventional diversity can be obtained over frequency, time, and space. Due to the
characteristic of wireless network with multiple users, there are several users can receive
the signal from the transmitter. Therefore, the transmitter establishes different signal
paths to different users. This called multiuser diversity, the transmitter selects an
appropriate user to improve performance. This condition can be maximizing channel
capacity or maximizing average throughput over users. If system transmits to users with
large SNR channel at all times, the overall spectral efficiency of the system can be
reached significantly higher than that of a nonfaded channel. . An application of
multiuser diversity for MIMO system has been studied in [7]. This paper proposed a
number of different scheduling schemes for MIMO spatial multiplexing systems with
perfect channel knowledge. In [8], the practical implementation and consideration of
multiuser diversity for CDMA/HDR was discussed.
When OFDM operated with MIMO techniques, there are multi-dimensional
parameters can be adapted for radio resource allocation, such as code, tone, time, and
space. In [9], the authors proposed an iterative bit and power allocation schemes for VBLAST
based MIMO-OFDM systems. The algorithm is based on the detection method
of V-BLAST to allocate the sub-channels to single user according to his bit error rate
requirement. The same research teams also present a simplified bit allocation for VBLAST
MIMO-OFDM system [10]. The new algorithm requires one bit for feedback
CSI. It’s more efficiency to allocation resource to single user. However, these two
algorithms only consider the single user case. Hence, we can apply the multiuser
diversity technique to such systems.
In this project we describe a multisuer diversity scheme for MIMO-OFDM systems
that uses V-BLAST as a detection algorithm. This scheme includes two parts, first is
determination of feedback CSI which is based on the V-BLAST algorithm. The other one
is a scheduler which allocates sub-channels to different users according to their CSI. We
can apply the optimization problem, integer linear programming, to the scheduler for
channel allocation.
This paper is organized as follows. In Section 2 we introduce the system model of
MIMO-OFDM system and in Section 3 the V-BLAST detection algorithm is described.
In Section 4, we propose the multiuser diversity scheme for MIMO-OFDM systems.
Then we compare the system performance with different feedback CSIs in Section 5.
Finally, a conclusion is made in Section 6.
2. System Model
A simple model for the downlink of a V-BLAST based MIMO-OFDM system is
illustrated in Figure 1. The base transceiver station (BTS) has Mt transmit antennas while
each antenna transmits Nc sub-carrier signals. In total, there are Mt × Nc sub-channels. A
queue of packets is stored at the BTS for each of the K users. Based on feedback CSIs
from users, the BTS allocate several sub-channels to user k to send the data packets of
length Rk
b. Therefore,
1 1
ΣΣ ,
= =
⋅ ≤ ∀
Mt c
i
k
b
N
c
k
i c B x R k (1)
where k
i c x , is the sub-channel allocated for user k, B is the number of bits per sub-channel.
Here, we assume that an ordinary M-QAM signal constellation is being user in the
modulation.
At the receiver side, there are Mr receive antennas for each user. The received signal
from all receive antennas will pass through a Fast Fourier Transform. We assume that the
path delays for all the spatial channels are the same. Furthermore, perfect symbol timing
synchronization and perfect channel estimation is assumed. There signals will be
processed by using the V-BLAST detection algorithm to retrieve the original signal.