03-06-2013, 03:24 PM
On Scheduling and Power Control in Multi-Cell Coordinated Clusters
On Scheduling and Power.pdf (Size: 2.93 MB / Downloads: 20)
Abstract
Recently, tight network coordination in cellular
systems has been demonstrated to improve the spectrum efficiency
by means of signal processing methods. However, the
performance of signal processing based multi-cell coordination
is sensitive to backhaul delays, channel estimation errors and
imperfections in fast link control. In this paper we consider tight
network coordination for fast radio resource management (RRM)
including packet scheduling, power control and modulation and
coding scheme selection. We use a system level simulator to
analyze the uplink performance of a multi-cell coordinated
system that is built around a fast backhaul transport infrastructure
for the purpose of enabling coordinated RRM rather
than coordinated signal processing.
INTRODUCTION
Network coordination as a means to provide high spectrum
efficiency for the downlink of cellular networks is on its
way from theoretical studies to practical implementations, see
for instance [1], [2], and [3]. For the uplink, Venkatesan
showed that the spectral efficiency can be doubled or even
quadrupled when the network has several coordination clusters
[6], [7]. These promising results have triggered the interest of
standardization bodies and industry players [4], [11] to investigate
the system design aspects of such tightly coordinated
systems. Coherent multi-site transmit antenna coordination
requires a high capacity backhaul infrastructure that allows
for the exchange of large amount of data corresponding to
“raw” receive/transmit (rx/tx) signals on the time scale of the
packet scheduler. That is, for the coordination to work, there
is a need for a high data rate multi-site communication on the
time scale of milliseconds or even less [1], [5].
MULTI-CELL RADIO RESOURCE ASSIGNMENT
System Considerations
We consider a cellular system in which multiple cells are
assigned a common central node that is responsible for the
coordination of the radio resource allocation in each of the
connected cells. The set of cells that are under the coordination
of the same central entity is called a coordination cluster. There
can be multiple number of transmission and reception antennas
at the different cell sites, connected to the central node via high
capacity transport links. Such a network layout is presented in
Figure 1, where three cells are connected to the central entity
Central Scheduling
The main idea behind the central scheduler is to control
the interference caused and suffered by UEs scheduled on
the same resource in different cells by properly selecting the
UEs that can be scheduled concurrently. Utilizing the channel
information available in the central node for all UEs and
antenna site pairs, the scheduler can estimate the interference
that would be caused to neighbor sites if a given UE were
scheduled. Then, it is possible to recalculate the optimal
transmit power setting corresponding to the new source of
interference and obtain the achievable SINR, after which the
link adaptation (LA) can choose a corresponding modulation
and coding scheme (MCS), which gives the number of bits that
can be transferred on the given connection. Obviously, adding
a new concurrent UE transmission to an existing schedule
decreases the amount of bits that can be carried by the already
scheduled connections but it increases the number of parallel
transmissions, which is a tradeoff that needs to be evaluated
by the scheduler.
SIMULATION RESULTS
Simulation Environment
For our simulations we use a system level simulator [13],
which implements detailed channel propagation models as
well as higher layer link protocols and functions, such as
HARQ, ARQ, link adaptation and scheduling. Network layer
protocols such as TCP/IP are also implemented. The channel
propagation models are according to the ones defined
in [11], from which we use the typical urban channel for
our simulations. This simulator also supports all the features
of the investigated multi-cell coordinated system, including
central scheduler, multi-cell power control algorithm and link
adaptation.
Numerical Results
We summarize the simulation parameters in Table I. The
configuration that we investigate comprises one coordination
cluster with 7 antenna sites, in which the users are distributed
uniformly. In the simulations we compare four scenarios. The
Single cell - open loop PC scenario is used as a reference
case, in which each antenna site has its own scheduler without
any multi-cell channel knowledge. In this solution there is
neither CSI nor scheduling information from neighbor cells
available, and therefore it is not possible to accurately estimate
the interference that would hit a scheduled transmission. The
interference needs to be estimated using a moving average
filtering of the interference measured in past transmission time
intervals (TTIs). Then the power control algorithm adjusts
the transmission power based on the estimated interference
and on the large scale path gain of the channel (i.e., without
considering multipath effects), which scheme is here referred
to as open loop power control.
CONCLUSIONS
In our analysis we have shown that multi-cell coordinated
RRM functions can provide considerable performance benefits
without imposing strong requirements on the backhaul
infrastructure. The combined multi-cell power control, link
adaptation and scheduling algorithms, proposed in this paper
require only the exchange of channel state information as
additional signaling data on the backhaul links, while user
plane data needs to be transferred only from one site at a
time, meaning that parallel data transfer with more significant
transport network implications can be avoided. Moreover,
coordinated fast RRM schemes have demonstrated low sensitivity
to backhaul delays during the simulations.