Dynamic Topology Control for Multi-hop Relaying in a Cellular TDD-OFDMA System Hye J. Kang, Hyun S. Ryu, and Chung G. Kang School of Electrical Engineering,
Post on 18-Jan-2016
214 Views
Preview:
Transcript
Dynamic Topology Control for Multi-hop Relaying in a Cellular
TDD-OFDMA System
Hye J. Kang, Hyun S. Ryu, and Chung G. KangSchool of Electrical Engineering, Korea University
IEEE WCNC 2009
Outline Introduction Motivation & Goal Proposed Algorithm Simulation Conclusion
PreambleDL/UL MAP
MS
Introduction Recently, Multi-hop relay systems are considered as a
useful means for enhancing coverage and throughput. The 802.16j Relay Task group was formed to standardize
a WiMAX multi-hop relay (MMR) system.
MR-BS
RS
PreambleDL/UL MAP
MS
RS
Non-transparent RS
Introduction In an MMR system, MSs are allowed to route through inte
rmediate RSs to reach the BS, which differs from the single-hop WiMAX topology.
MR-BS
RS
PreambleDL/UL MAP
MS
MR-BS
MS
Non-transparent RS
Motivation & Goal Multi-hop relay frame structure in this paper
TDD-OFDMA
RS
MS
MS
Motivation Multi-hop relay frame structure: N=4 hops
the relay zone can be time-divided into the multiple subzones Each sub-zone is reserved for a subsequent relay link
RS 1
RS 3
RS 5
RS 2
RS 4
RS 6
RS 7
Motivation In fact, if N is too large, the N-th subzone becomes almost a null.
RS 1
RS 3
RS 5
RS 2
RS 4
RS 6
RS 7
RS 7
Motivation In fact, if N is too large, the N-th subzone becomes almost a null.
RS 1
RS 3
RS 5
RS 2
RS 4
RS 6
RS 7
RS 7
Motivation In general, we can divide the relay zone into K sub-zones a
nd then, reuse them among the different layers. In this paper we consider k = 3, because it will be the most efficient form of frequen
cy reuse.
Motivation Co-channel interference in the same branch is one serious
problem under the divide-by-K reuse strategy.
Goal Our objective is to configure a feasible tree topology subje
ct to the divideby-K reuse strategy for N-hop maximizing the bandwidth efficiency of resource available for rela
y links.
Proposed algorithm RS Layering Algorithm
The objective of RS layering is to minimize the average inter-hop interference.
RS Clustering Algorithm The objective of the RS clustering algorithm is to determine a
super-ordinate RS for each RS in the next upper layer so as to minimize the effective delay.
Assumption Given data rate between two RSs. Given the received signal strength between each RS and B
S. Given the received signal strength between two RSs. Each RS must have its own unique super-ordinate RS in th
e tree structure.
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS1 RS2 RS3 RS4 RS5 RS6 RS7 RS8
MCS 64 64 16 16 QPSK QPSK QPSK QPSK
RSS 20W 20W 15W 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
RS RS1 RS2 RS3 RS4 RS5 RS6 RS7 RS8
MCS 64 64 16 16 QPSK QPSK QPSK QPSK
RSS 20W 20W 15W 14W 10W 2W 3W 11W
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
1bit/5Mbps1bit/10Mbps
1bit/9Mbps
RS 5
RS 6
RS 7RS 8
RS RS1 RS2 RS3 RS4 RS5 RS6 RS7 RS8
MCS 64 64 16 16 QPSK QPSK QPSK QPSK
RSS 20W 20W 15W 14W 10W 2W 3W 11W
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
RS RS1 RS2 RS3 RS4 RS5 RS6 RS7 RS8
MCS 64 64 16 16 QPSK QPSK QPSK QPSK
RSS 20W 20W 15W 14W 10W 2W 3W 11W
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
WRSS 25,7
WRSS 105,2
WRSS 15,6
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
10,min5,6
5,2
5,7
5,2
thRSS
RSS
RSS
RSS
WRSS 25,7
WRSS 105,2
WRSS 15,6
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
10,min5,6
5,2
5,7
5,2
thRSS
RSS
RSS
RSS
WRSS 104,1 WRSS 5.04,8
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
10min4,8
4,1
thRSS
RSS
WRSS 104,1 WRSS 5.04,8
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
1bit/5Mbps1bit/10Mbps
1bit/9Mbps
RS Layering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS RS4 RS5 RS6 RS7 RS8
MCS 64 64 QPSK 16 16
RSS 14W 10W 2W 3W 11W
RS 5
RS 6
RS 7RS 8
10min4,8
4,1
thRSS
RSS
Proposed algorithm RS Clustering Algorithm
The objective of the RS clustering algorithm is to determine a super-ordinate RS for each RS in the next upper layer so as to minimize the effective delay.
RS Clustering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
RS Clustering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
ms12
ms20
RS Clustering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
ms12
RS Clustering Algorithm
RS 1 RS 2 RS 3
RS 4
BS
RS 5
RS 6
RS 7RS 8
Simulation We consider a simulation scenario in which relay stations and mobile s
tations are uniformly in a single cell of 5km radius. The transmit BS power and RS power are given by 20W and 10W, res
pectively.
Simulation Among 27 downlink OFDM symbols, 14 symbols are assi
gned to access zone while the rest of them are assigned to the relay zone.
The loading factor threshold η th for the proposed algorithm is set to 10 in the current simulation.
Simulation the proposed scheme (labeled by “Layering + Clustering”) One which employs layering only, without resort to cluster
ing (labeled by “Layering”). a layering process can be replaced by the best rate selectio
n in the proposed scheme (labeled by “Best rate + Clustering”).
Simulation Average end-to-end throughput: Divide-by-3
Simulation Average MS outage probability: Divide-by-3
Conclusion In this paper, we have proposed a dynamic topology contro
l algorithm that deals with the routing and resource allocation for the relay stations in the cellular.
top related