Reconfigurable MIMO Transceiver Design using the
Tunable Channel Decomposition
Jing Wang and Gerald E. Sobelman
Department of Electrical and Computer Engineering
University of Minnesota
Minneapolis, MN 55455 USA
Abstract—Available spectrum resources are becoming scarcer,
while additional services are being integrated into wireless
devices. Cognitive radio and multiple-antenna (MIMO)
techniques can be combined to provide a solution to this dilemma.
Given the currently available frequency bands and a set of bit-
rate constrained services to be transmitted, the antennas can be
partitioned among frequency bands and the recently proposed
tunable channel decomposition (TCD) may be applied to obtain
subchannels having prescribed gains. We present a
reconfigurable closed-loop transceiver design which can adapt to
various partitioning scenarios. As a result, our system has the
potential to deliver the required services with reduced
transmitting power.
Keywords – Reconfigurable, multiple-input multiple-output,
MIMO, transceiver, cognitive radio, quality-of-service, QoS,
tunable channel decomposition, TCD.
I. INTRODUCTION
With the ever-increasing demand for more versatile and
multi-functional mobile devices, new types of services are
being integrated into modern wireless systems, which will
require additional bandwidth and power consumption.
However, the available spectrum resources are scarce and very
expensive, thus placing considerable constraints on the services
that can be added onto a system. What’s more, these services
usually have different quality-of-service (QoS) constraints
requiring different transmission bit-rates. How to satisfy these
disparate QoS constraints in a power efficient way is a
significant issue, given that power consumption is a critical
parameter in modern mobile devices. To address these
challenges, cognitive radio and multiple-antenna (MIMO)
techniques may be applied.
Cognitive radio may utilize a software defined radio (SDR)
in which some or all of the physical layer functions are
determined with software or other reconfigurable techniques [1]
[2]. A cognitive radio is adaptable in the sense that it is aware
of its environment and can adjust its operation according to
dynamic changes in the local spectral conditions [3]. In the
system described here, we assume that it can determine the
currently available frequency bands and adapt itself
accordingly. In other words, our system can make use of
certain frequency bands found through spectrum sensing during
a specific time period, i.e. when they are not occupied in the
local area by the licensed or primary users [4].
Given one or more available frequency bands, the next step
is to accommodate the required services in such a way that the
system consumes the least amount of power. Considering
several services sharing the same channel, one possible
solution is MIMO technology which has already been widely
adopted [5] [6]. Assuming that channel state information (CSI)
is available at both the receiver (CSIR) and at the transmitter
(CSIT), channel decomposition methods can be applied to
divide the channel into multiple independent parallel
subchannels, each of which can then be used for a specific
service. It is well-known that the singular value decomposition
(SVD) and the water-filling technique can be combined
together to achieve the channel capacity [7] [8]. However, the
gains of the subchannels obtained by the SVD are determined
by the singular values of the channel matrix and cannot be
adjusted by the user. This aspect makes it less suitable for
transmitting the QoS-constrained services [9]. Instead, we use
another method called the Tunable Channel Decomposition
(TCD) which was proposed in [10] and which can decompose a
MIMO channel into an arbitrary number of subchannels having
prescribed capacities. The TCD scheme can ensure minimal
power consumption for transmitting a group of QoS-
constrained services in a single MIMO system [10].
Given that the TCD scheme can offer the minimal power
solution in a single MIMO system, what we need to do in our
system is to find the best way to: 1) group antennas together
onto different frequency bands to form subsystems, if
necessary; 2) partition the required services onto the
subsystems. Then, applying the TCD scheme to each
subsystem should result in a power efficient solution. Prior
work has established the algorithmic basis for this approach
[11]. In this paper, we will present the hardware architecture
design for such a system. Specifically, we will present a
reconfigurable MIMO transceiver which can operate as either a
single MIMO system or as two MIMO subsystems, all of them
using the TCD. Thus, this transceiver can provide the flexible
hardware that is necessary for achieving minimum power
operation.
The remainder of this paper is organized as follows: Section
II briefly reviews the TCD scheme that achieves minimal
power in a single MIMO system. The reconfigurable MIMO
transceiver design is presented in Section III. Section IV gives
the simulation and synthesis results, and Section V presents our
conclusions.
This work was supported in part by the U.S. National Science Foundation under
grant number ECS-0621879.