Title: Halim Yanikomeroglu
1Infrastructure-Based Wireless Multihop, Relay,
Mesh Networks
- Halim Yanikomeroglu
- halim_at_sce.carleton.ca
- www.sce.carleton.ca/faculty/yanikomeroglu.html
Broadband Communications Wireless Systems
(BCWS) Centre Department of Systems Computer
Engineering Carleton University Ottawa
2Relaying New Perspective vs Old Perspective
Conventional relaying for coverage Analog/digital
dump (non-selective) repeaters Satellite,
microwave, cellular, deep space,
R
Relay
R
S
R
R
S/R
R
R
Relay/mesh/multihop networks for coverage,
throughput, QoS, , through cooperation
3Relaying
single-hop
multihop
- Relaying paradigm shift in systems level
- wired AWGN ? wireless fading ? wireless
mesh - Impact in all layers of wireless communications
- propagation
- physical layer (PHYITDSP) channel capacity,
cooperative relaying, - multiple access layer (MAC) RRM, scheduling,
CAC, - networking layer load balacing, routing,
handoff, - higher layers and protocols
4Relaying
single-hop
multihop
- Relaying great interest in academia and industry
- great interest does not necessarily mean
successful realization - new network architecture ? network problem
- physical layer just one element
- MAC, networking, protocols important
- time to realization
- TX diversity (Alamouti), turbo codes 4-5 year
- CDMA 15 years
- ad hoc networks 8
- relaying will take some time, but will become
a reality -
5Relaying Comprehensive Investigation Required
- Simple example
- Relaying links (hops) with less path-loss
- higher spectral efficiency in each link
- Relay cannot receive and transmit at the same
channel (half-dublex) - 2-hop link ? 2 channels (time or frequency)
- n-hop link ? 2 to n channels
- Capacity gain or loss?
-
6Infrastructure-based vs Infrastructure-less
Multihop Networks
cellular WiFi WiMax (sensor)
ad hoc
infrastructure-based multihop network BS/AP ?
common source or sink
infrastructure-less multihop network
systems networking layers many differences
physical layer many
similarities
7Outline
- Part I Relaying for cost-effective ubiquitous
high data rate coverage - Capacity limited vs coverage limited networks
- Relaying promising solution for coverage limited
networks - Part II Possibilities
- Analog relaying vs digital relaying
- Fixed relaying vs terminal relaying
- Homogeneous relaying (single air interface) vs
heterogeneous relaying (dual air interface) -
- Part III Further exploitation of the
relay/multihop/mesh architecture - Cooperative relaying
- Novel diversity schemes
- Intelligent routing and scheduling
- Diversity- and AMC (Adaptive Modulation and
Coding)-aware routing in infrastructure-based
TDMA multihop networks - BS-relay coordination
- Dynamic frequency hopping in cellular relay
networks
8Expectations for 4G Wireless Networks
- WWRF (World Wireless Research Forum) mITF
predictions - mobile up to 100 Mbps
- stationary/nomadic up to 1Gbps !!!
But how?
- More bandwidth is needed
- around 3 or 5 GHz band (World Radio Conference,
Nov 2007)
- More bandwidth ? more rates it does not scale
necessarily! - High bandwidth high carrier frequency
- Tremendous stress on link budget
- advanced antenna technologies (MIMO, smart)
- advanced signal processing (modulation, coding,
equalization) - advanced radio resource management techniques
necessary but not sufficient
- A fundamental upgrade in the network architecture
is needed
9Cellular Design Fundamentals
QoS
Eb
Pr
I)
R.Eb
R Data Rate
Pathloss
Pt
Cell Size
Propagation Conditions
II)
10Capacity-Limited Networks
11Coverage-Limited Networks
available capacity / cell gt capacity demand ?
coverage limited
12Capacity-Limited vs Coverage-Limited Networks
1G 2G
Capacity limited ? network grows as needed ?
great success
- Ubiquitous high data rate coverage limited
- ? very high deployment cost from the beginning
- ? great challenge
(3G) 4G
WLAN low deployment cost ? great success
13Coverage Extension through Digital Fixed Relays
- Low cost digital fixed relays located at
strategic locations - No wired internet connection at relays
- Different from conventional fixed relays
(selective relaying) - Different from ad hoc networks (routing is less
of an issue)
What is a relay?
relay
relay
relay
relay
BS/AP
BS/AP
Same high data rate coverage
relay
relay
14Coverage CDFs
H. Hu, H. Yanikomeroglu, D.D. Falconer, S.
Periyalwar Range Extension w/o Capacity Penalty
in Cellular Networks with Fixed Relays,
Globecom 2004
propagation exponent 3
propagation exponent 3
propagation exponent 3.5
15Average spectral efficiency w.r.t. cell size
Outage w.r.t. cell size
propagation exponent 3.5
16Cost-Efficient Range Extension
- Same average spectral efficiency and outage
w.r.t. cell size trends are observed for
different values of - Propagation exponent
- Cluster size
- Shadowing standard deviation
- BS transmit power
- Ex Range extension x2 Cost micro-BS / Cost
relay 10 - Cost microcellular network / Cost relay
network - Cost L micro-BSs / (Cost L/4 micro-BSs
Cost 6L/4 relays) - 40/16 2.5
-
Relay networks significant potential for range
extension
17(IEEE Globecom 2004)
Cost conventional network / Cost relay
network
18Other Cost Effective Network Architecture
Alternatives
Infostations
Distributed Antennas
Microcellular Network
Antenna Remoting / Radio-on-Fiber
Chinese FuTURE Project
(Future Technology for Universal Radio
Environment) www.chinab3g.org/english/futureproje
ct.htm
non-ubiquitous coverage
wiring cost
wiring cost
Base station
Antenna
19Capacity of Cellular Fixed Relay/Mesh Networks
Central Node (CN)
?
Relay
Wireless Link
- Only CN is connected to the backhaul
- No Tx Rx on the same channel for a relay
- Nodes have two kinds of antenna
- Direct link with only the neighbor nodes
- Same BW for each primary link
M of root nodes (trees) N of nodes per
tree MN1 of nodes per cell
20W (Hz) B (Hz) RCCN (bits/sec) RCFRN
(bits/sec) RB M N
Total available bandwidth for a cell
Available bandwidth for a relay
W
B
RCCN RW Capacity of CCN is a function of
W RCFRN (M.N1)RB M.N.RB
21(No Transcript)
22 23 24 25 26 Total Number Channel Groups
27 Previously it is stated that Then, In
general, if each node has child nodes
28 Total number channel groups when all the hop
links use orthogonal channel groups
Then
29Capacity Comparisons
- When every other hop links reuse the same
channel groups
- When all of the links use orthogonal channel
groups
30Bandwidth Allocation to Access and Feeder Systems
Florea, Yanikomeroglu IEEE WCNC06
BT Total BW BA total Access BW BF total Feeder
BW Ba access BW for each relay Bf feeder BW for
each relay ma spectral efficiency for access mf
spectral efficiency for feeder nr number of
relays BT BA BF BA NBa BF nr Bf Bf mf
Ba ma
Access system cluster size N 3
Base Station service area
Each color represents a frequency group
Relay service area
31Bandwidth Allocation to Access and Feeder Systems
Bandwidth allocation for ACCESS
feeder portion
access portion
Bandwidth allocation for FEEDER
if nr ? 8, then BA/BT ? 0 and BF /BT ? 1
Ex nr 25, mf 4, ma 2
Throughput
access, feeder throughput
N 3 19, 81 3.2 BT
N 1 7, 93 3.7 BT
if nr ? 8, then T ? mf BT
32Bandwidth Allocation to Access and Feeder Systems
Why to use relays?
?
100 Mbps only in this small area
67 Mbps in a much larger area
33Relay Network vs Microcellular Network
Potential capacity remains more or less the same
Usable capacity increases (outage decreases)
due to better coverage
Relays distribute the total capacity throughout
the coverage region
Relay Network Microcellular Network
Cost-effective high data rate coverage excellent moderate
Capacity (aggregate throughput) moderate excellent
Rule of Thumb in Design Deploy as many
BSs/APs as needed according to capacity demand
Then distribute the capacity in the coverage
region evenly using as many relays as needed
34Spectral Efficiency of n-hop Link with Orthogonal
Channels
b spectral efficiency of single-hop link ai
spectral efficiency in hop i ac net (overall)
spectral efficiency of n-hop link M message
size B bandwidth T message transfer time
TMH T1 T2 Tn (assuming orthogonal
channels)
TMH M/(Ba1) M/(Ba2) M/(Ban) M/(Bac )
35Multiplexing Loss in Multihop Relaying with
Orthogonal Channels
- When does it make sense to break a single-hop
into multiple hops? - Low SNR case
- Single-hop with 4 dB SNR ? ½-rate QPSK 1 b/s/Hz
- Two hops each with 12 dB SNR ? ¾-rate 16-QAM 3
b/s/Hz - Net spectral efficiency 1.5 b/s/Hz ? use
multihop - High SNR case
- Single-hop with 26 dB SNR ? full-rate 64-QAM 6
b/s/Hz - Two hops each with 34 dB SNR ? 128-QAM 7 b/s/Hz
- Net spectral efficiency 3.5 b/s/Hz ? use
single-hop - Rule of thumb low SNR ? multi-hop
- high SNR ? single-hop
opportunistic relaying
36Multi-Hop Criterion
Florea, Yanikomeroglu IEEE Globecom05
- If single-hop SNR (g) satisfies
- then
- net spectral efficiencyn1 gt
- net spectral efficiencyn
- n - number of hops
- p - path loss exponent
-
- Assumptions
- all links have same path loss exponent p
- All relays are placed uniformly on a in straight
line from source to destination
SNR values under which there exists a (n1)-hop
link with better spectral efficiency compared
with an n-hop link
37Outline
- Part II Possibilities
- Analog relaying relaying vs digital relaying
- Fixed relaying vs terminal relaying
- Homogeneous relaying (single air interface) vs
heterogeneous relaying (dual air interface) -
38Analog Relaying vs Digital Relaying
- Digital relaying (router, bridge)
- Regenerative relaying
- Decode-and-forward (detect-and-forward) relaying
- Adaptive (selective) decode-and-forward
-
- Analog relaying
- Non-regenerative relaying
- Amplify-and-forward relaying
- On channel
- With frequency translation
- Analog relaying ? noise propagation
- Digital relaying ? error propagation
- Analog relaying may be better than digital
relaying in certain scenarios - Hybrid analog/digital relaying another
possibility
DF
CMP
digital relay
39Fixed Relaying vs Terminal/Mobile Relaying
- Terminal relaying rich theoretical area, full of
potentials - But, many technical challenges
- No service guarantee
- Increased energy consumption (fast battery
draining) - Increased transmit power (in CDMA)
- Additional hardware and functionality (higher
terminal cost) - Security issues
- Frequent hand-offs (especially in the presence of
high mobility) - Terminal relaying any incentives?
- Special applications single team
(law-enforcement, military, rescue) - Non-battery powered fixed user terminals
(802.16a) - Cooperative relaying with simultaneous mutual
benefits (symmetric cooperation) - Personal area networks
- Commercial applications business plan needed
(air time offers?) - Ad hoc networks (infrastructureless) no internet
connection!
40Fixed Relaying vs Terminal/Mobile Relaying
- Routing easier in infrastructure-based multihop
networks than infrastructureless ad hoc multihop
networks - Nodes with extra complexity and intelligence
(BS/AP or fixed relays) - Common source or destination
- ? Routing with more demanding goals possible
- Expectations in 4G networks
- first, fixed relays
- then, mobile/terminal relays
- Single-hop (infrastructureless) ad hoc networks
possible - Multihop (infrastructureless) ad hoc networks
commercially difficult!
41Heterogeneous (vs Homogeneous) Relaying
- Decoupling of access and backbone networks
- Access air interface A
- Backbone (feeder) air interface B
- Customized air interfaces
- Easier interference management
- License-exempt bands can be utilized
42Single Air Interface vs Dual Air Interface
Walsh, Yanikomeroglu IEEE CCECE04
New issues emerge Ex power control in the
reverse-link of two-hop CDMA networks
Two-hop PC
Single-hop PC
43Outline
-
- Part III Further exploitation of the
relay/multihop/mesh architecture - Novel diversity schemes
- Advanced cooperative relaying
- Intelligent routing
- Diversity- and AMC (Adaptive Modulation and
Coding)-aware routing in infrastructure-based
TDMA multihop networks - BS-Relay coordination
- Dynamic frequency hopping in cellular relay
networks - WiMax mesh networks
44Theoretical Literature on Relay/Mesh/Multihop
Networks
- Van der Meulen (68,71)
- El Gamal, Cover, Aref (79,80,82) Stanford
- Willems (83,85)
- Sendonaris, Erkip (98-) Rice
- Laneman Notre Dame, Wornell MIT (00-)
- Gupta ASU, Kumar UIUC
- Dohler KCL
- Nosratinia, Hunter UTD
- Tse Berkeley
- Hasna Qatar, Alouini Minnesota
- Giannakis, Cai, Ribeiro Minnesota
- Host-Madsen Hawaii
- Pottie
- Zhang ASU
- Stefano Brooklyn
- Wittneben, Rankov
- Cho, Haas Cornell
- Anghel, Emamian, Kaveh Minnesota
- Bolcskei, Nabar ETH
- Gastpar Berkeley
- Dawy AUB
- Kramer Bell Labs
- Franceschetti UCSD
- Herhold, Zimmermann TUD
- Vetterli EPFL
- Valenti
- Walke Aachen
- Karagiannidis AU Thessaloniki
A few conference papers ? explosion in literature
45Relaying New Perspective vs Old Perspective
coverage extension
R
Relay
R
S
R
R
S/R
R
R
Cooperative relaying ? virtual antenna
array diversity, space-time coding, MIMO,
46Asymmetric vs Symmetric Cooperation
- Asymmetric Cooperation relay terminal between
source and destination - Only one terminal benefits
- Pathloss gain
- Symmetric Cooperation a pair of nearby terminals
cooperate - Immediate benefit for both terminals
- No pathloss gain
47User Cooperative Diversity
Sendonaris, Erkip, Aazhang IEEE T-COM, Nov.
2003 (ISIT 1998)
Channel Model (full-duplex assumption)
48Implementation case example CDMA Resources
Distribution codes
power allocation and
No cooperation
- Using the same number of codes as in no
cooperation scenario
With cooperation
Users exchange cooperative information BS may
hear
Constructed cooperative signals are sent to BS
Users transmit to BS
49Laneman, Tse, Wornell IEEE T-IT,
Dec04 Laneman, Wornell IEEE T-IT, Oct03
50Parallel Relays
relay
first slot broadcast
second slot multiple access orthogonal ?
multiplexing loss beamforming ? CSI (overhead)
relay
diversity order up to R1 provided that error
propagation is prevented
relay
51Multi-Antenna Aspects of Cooperative Fixed Relays
Adinoyi, Yanikomeroglu WCNC06 (IEEE
TWireless07)
use few relays with multi-antennas (even with
selection combining) instead of many relays with
single antennas
Rlay 1
First Hop
Second Hop
52Practical Cooperative Communication Schemes
Through Fixed Relays
Adinoyi, Yanikomeroglu WWRF15, VTCS06
no need for incentives for cooperation no
security risk no complexity incurred in
terminals complexity moved to network
53Practical Cooperative Communication Schemes
Through Fixed Relays
54Practical Cooperative Communication Schemes
Through Fixed Relays
55Multihop Diversity
- Analysis of multihop channels with diversity
- Decoded relaying with diversity intermediate
terminals combine, digitally decode and re-encode
the received signal from all preceding terminals - Amplified relaying with diversity intermediate
terminals combine and amplify the received signal
from all preceding terminals - Diversity in 2-Hop Links
- Mazen Hasna, Mohamed-Slim Alouini
- J. Laneman, Greg Wornell
- Diversity in n-Hop Links (for fully connected
networks) - John Boyer, David D. Falconer, Halim
Yanikomeroglu, Multihop Diversity in Wireless
Relaying Channels, IEEE Trans. on
Communications, Oct. 2004 - Main Observations
- Comparison of relaying and diversity schemes wrt
BER - Multihop Diversity gt Multihop gt Singlehop
- Amplified Relaying gt Decoded Relaying
- DRMD improves when intermediate terminals are
closer to the source terminal - ARMD improves when intermediate terminals are
closer to the destination terminal
56Full-Diversity Relays and Destination
Boyer, Falconer, Yanikomeroglu IEEE T-COM, Oct.
2004
- Full diversity reception at all receivers
- Requires n channels for n relaying hops
- Complex relay behavior (diversity combining)
57Aggregate SNR of Amplified Relaying Channels
- John Boyer, David D. Falconer, Halim
Yanikomeroglu, On the Aggregate SNR of - Amplified Relaying Channels, IEEE Globecom 2004.
- Aggregate Signal to Noise Ratio
- Aggregate inclusion of propagated noise terms
in the SNR formulation. - Propagated noise terms are generated as amplified
relaying terminals amplify both the information
and noise portions of received signals
indiscriminately. - Motivated by findings indicating that the
performance of amplified relaying can approach
and in some cases exceed that of decoded
relaying. - Aggregate SNR expressions developed for amplified
relaying channels with given source, destination,
and relaying terminals, link connectivity, link
attenuation, transmit power, and receiver noise. - Aggregate SNR expression developed for following
network connectivity - Serial Amplified Relaying Channels (serial node
connectivity) - Parallel Amplified Relaying Channels (parallel
node connectivity) - General Amplified Relaying Channels (general node
connectivity)
58Amplified Relaying
Source terminal
Destination terminal
59ith Amplifying Relay
From Terminal k
60Definitions
- Aggregate SNR at for Link SNR at
for - where
- is the transmitted power
- is the complex amplitude of the information
symbol - is propagated noise
- is distance-dependant attenuation,
shadowing, and fading - is the variance of the zero-mean Gaussian
random variable - This model normalizes the transmit signal such
that
61Some Equations
- Serial Amplified Relaying Channels (recursive
form) - Serial Amplified Relaying Channels (sum of
products form) - Parallel Amplified Relaying Channels
- General Amplified Relaying Channels
62Example Channels
Fig 1. Serial Connected
Fig 2. Parallel Connected (multi-route diversity)
Fig 3. General Connected
Fig 4. General Connected 2
Fig 5. Fully Connected
63Aggregate SNR vs. Link SNR
Fig. 1 Fig. 2 Fig. 3 Fig. 4 Fig. 5
64Comparison with a Linear Relation (high order
multiplicative termsin formula at bottom of
slide 62 removed)
Fig. 1 Fig. 2 Fig. 5
65Observations
- Serial Amplified Relaying Channels Allocate
relatively more power to weaker links - Parallel Amplified Relaying Channels Allocate
relatively less power to weaker links - General Amplified Relaying Channels Allocate
relatively more power to weak links that are not
parallel to strong links and relatively less
power to weak links that are parallel to strong
links - Maximizing Links in parallel (diversity gains)
more important than maximizing links in serial
(attenuation gains) for the example simulations. - For strong links in serial the performance is
approximately linear with respect to the
component link SNRs - For weak links in serial the performance is less
than linear with respect to the component link
SNRs - Amplified relaying may not be appropriate for
very low SNR systems as propagated noise becomes
limiting factor.
66System Resource Constraints
- John Boyer, David D. Falconer, and Halim
Yanikomeroglu, Impact of System Resource
Constraints on the Connectivity of Wireless
Relaying Networks, ICC 2005 (under review in
TWireless) - System resource constraints have a direct impact
on the link connectivity of wireless relaying
networks and therefore the possible distributed
spatial diversity techniques - Constraints considered
- Number of Channels Available (NCA) N or 2
available channels - Relay Combination (RC) Diversity combining or
simple relay - Destination Combination (DC) Diversity combining
or simple relay - Multiple Channel Reception (MCR) Multiple
channels or single combined by receiver - Multiple Channel Transmission (MCT) Multiple
channels or single transmitted - Possible constraint combinations are analyzed and
reduced - Relaying types
- Amplify-and-forward (analog) relaying
- Decode-and-forward (digital) relaying with error
propagation - Decode-and-forward (digital) relaying without
error propagation
67Full-Diversity Relays and Destination
Boyer, Falconer, Yanikomeroglu IEEE T-COM, Oct.
2004
- Full diversity reception at all receivers
- Requires n channels for n relaying hops
- Complex relay behavior (diversity combining)
682-Channel Diversity Relays and Destination
- Successive transmitters along multihop path
transmit/receive on alternating channels. - Requries 2 channels for n relaying hops.
- Complex relay behavior (diversity combining).
692-Channel Diversity Relays w Full Diversity
Destination
- Alternating channels with full diversity at
destination - Requires 2 channels for n relaying hops
- Complex relay behavior (diversity combining)
70Non-Diversity Relays w Full Diversity Destination
- Alternating channels with full diversity at dest.
- Requries 2 channels for n relaying hops.
- Simple relay behavior (no combining).
71System Connectivity Models
- Models defined by the connectivity of source,
relay, destination - 1X Terminal class X connected to one transmitter
- 2X Terminal class X connected to the subset of
transmitters on one channel - FX Terminal class X connected to all
transmitters - Inter-model transitions based on minimum cost
constraint sets
S Source R Relay D Destination
72System Connectivity Models (Samples)
1
1
0
1
0
1
0
1
1
0
0
1
1
0
0
0
1R1D
2R2D
0
1
1
1
0
1
0
0
1
0
0
1
1
1
1
0
0
0
1
0
0
0
1
1R2D
2RFD
0
1
1
0
0
0
1
1
0
0
1
0
1RFD
FRFD
73Minimum Cost Constraint Sets
- Mark of X indicates that the constraint is
lifted ? the capability described by the
corresponding system resource is available
Model NCA N Channels Available RC Relay Combination DC Destination Combination MCR Multiple Channel Reception MCT Multiple Channel Transmission
1R1D
1R2D X
1R2D2S X X
1RFD X X
2R1D X
2R1D2S X X
2R2D X X
2R2DFS X X X
2RFD X X X
2RFDFS X X X X
FR1D X X X
FRFD X X X X
74Amplified Relaying
1R1D
Singlehop
1R1D
1R2D
FR1D
2R2D
2R2D
1RFD
1RFD
FRFD
FRFD
2RFD
2RFD
75Decoded Relaying with Error Propagation
1R1D
1R2D
1RFD
Singlehop
2R2D
2RFD
FRFD
1R1D
FRFD
76Decoded Relaying without Error Propagation
1R1D
Singlehop
1R2D
1R1D
1RFD
2R2D
2R2D
FRFD
FRFD
2RFD
77Observation I
- Simulation of connectivity models for amplified
relaying, decoded relaying with error
propagation, and decoded relaying without error
propagation isolates the impact of each
constraint - Impact summary of system resource constraints
- Relaying Method NCA RC DC MCR MCT
- Amplified Small Small Large Large Small
- Decoded w Prop Small Medium Medium Small Medium
- Decoded w/o Prop Small Medium Large Large Small
- Different impact for different relaying methods.
- Connectivity priorities for optimal performance
- Amplified ? Maximize destination connectivity
- Decoded w Prop ? Equalize relay destination
connectivity - Decoded w/o Prop ? Maximize destination
connectivity disjoint network paths
78Observations II
- Diversity order of the system is dependent on the
connectivity - Amplified relaying ? Dependent only on
destination - Decoded relaying w Prop ? Dependent only on
minimally connected relay - Decoded relaying w/o Prop ? Dependent on
destination and disjoint paths - Guidance for the order in which the constraints
should be lifted - Amplified ? DC, MCR, RC, MCT
- Decoded w Prop ? RC, DC, MCT, MCR
- Decoded w/o Prop ? DC, MCR, RC, MCT
- Impact of lifting NCA contraint when all other
constraints are lifted is small - Likely will not be implemented in practice since
the cost of N orthogonal channels per information
signal transmission is very high with respect to
spectral efficiency.
79Ongoing Work
- Incorporation of Interhop Interference
Cancellation (IC) as a constraint - Explicit separation of Common Channel Combination
and Orthogonal Channel Combination as constraints - Generalization of analysis for arbitrary number
of channels available (KltN) - Formalization of method for deriving system
connectivity models - Mapping of models to cooperative diversity
techniques in the literature - Interrelationships between system resource
constriants - Relationship of system resource constraints to
relay node placement - Grid placement
- Colinear equidistant between source and
destination placement - Bunched nodes placement between source and
destination - Random placement with varying uniformity of relay
density - Practical issues and cost of system resource
constraints - Common channel combination techniques
- Cooperative coding
- Relay block size
- Spatial reuse of channels
- Wireless fading models
- Spectral efficiency
80Coordination among BSs/APs and Relays
- Coordination among BSs
- Scheduling
- Interference management
- Radio resource management
- Admission control
-
- Rich literature
- Limited usage in practice in conventional
cellular networks - May be used in cellular relay networks
81CLASSICAL DYNAMIC FREQUENCY HOPPING WITH NETWORK
ASSISTED RESOURCE ALLOCATION (DFH with NARA)
ATT Bell Labs
i.
ii.
Each terminal measures path losses to the
neighboring bases and transmits this information
to its serving base on a regular basis.
Each base communicates to several tiers of its
neighbors the information about its own resource
utilization (i.e. time slots, frequency hopping
patterns and current power levels).
iii.
The serving base station calculates the
interference level at each available resource,
determines the least-interfered time slot and FH
pattern pair, and assigns this to the terminal.
82BLOCK DIAGRAM OF A CELLULAR SYSTEM THAT SUPPORTS
DFH WITH NARA FOR DOWNLINK
MOBILE STATION 1
.
Measure Pathloss on BCCH for BS 1
Measure Pathloss on BCCH for BS K
.
Read and Use Specified FH Pattern
Average
Average
.
Send to BS1
Send to BS1
WIRELESS MEDIUM
To Other Mobiles
From Other Mobiles
BASE STATION 1
Send Orders To MSs with Next FH Patterns
Collect Measurements From All MS in this BS
Coverage Area
Local Copy of Measurements For All MS from All BS
Local Copy of FH Patterns For All BS
Manage Frequency Hop Patterns For This BS
From Other Base Stations
From Other Base Stations
To Other Base Stations
To Other Base Stations
LANDLINE NETWORK
83(No Transcript)
84DFH with LIMITED INFORMATION (Time Slot 2)
Mubarek, Yanikomeroglu, Periyalwar
- R1 3 in-cell interferers (R2,R6,BS) and 3
out-of-cell interferers (R3,R4,R5) - UE pathloss info R1?BS
- BS already has resource utilization information
of the in-cell interferers of R1 - BS decide on DFH pattern based on limited info
- BS?R1 DFH pattern
85FIGURE 9
INTERFERING SUB-CELLS
First Tier Interferers
Second Tier Interferers
Base Station
Relay Station
User Equipment
Cell Border
Relay or Base Neighborhood
86UE
Measure Pathloss for BS
Measure Pathloss for Interfering Relay1
Measure Pathloss for Interfering Relay2
Get the new FH Pattern
UE in RS-Neighborhoods
Average out Rayleigh
Average out Rayleigh
Average out Rayleigh
BASE STATION
Notify RSs with the new FH Patterns for UEs in RS
Neighborhoods
UEs IN RS NEIGHBORHOODS
Collect Pathloss Reports from UEs in RS
Neighborhood
Create new Random FH Patterns for UEs in RS
Neighborhoods from the pool of available and
unblocked Resources
Calculate which resources result in a SIR less
than the threshold SIR, SIRTh, and block them
NO COORDINATION
Notify UEs in RS Neighborhoods with their new FH
Patterns
IN-CELL RELAYS
LANDLINE NETWORK
Get the new FH Pattern
OTHER BASE STATIONS
OUT-OF-CELL RELAYS
87DFH with LIMITED INFORMATION (Time Slot 2)
UE in BS service region DFH with full information
88UE
Measure Pathloss for RS1
Measure Pathloss for RS6
BS-Neighborhood, BS relays as interferers
Get the new FH Pattern
Average out Rayleigh
Average out Rayleigh
BASE STATION
Notify UEs in BS Neighborhood with their new FH
Patterns
Collect Pathloss Reports from UEs in BS
Neighborhood
UEs IN BS NEIGHBORHOOD
Create new FH Patterns for UEs in BS Neighborhood
NO COORDINATION
LANDLINE NETWORK
IN-CELL RELAYS
OTHER BASE STATIONS
OUT-OF-CELL RELAYS
89Cooperative Induced Multi-user Diversity Relaying
(CIMDR)
Navaie, Yanikomeroglu VTCS06
- First hop
- Multi-user diversity exploits through
transmission with maximum bit-rate - Second-hop
- Multi-user diversity exploits through
transmission on a good channel - Two phase protocol Feeding and delivery phase
90Performance Improvements through the Mesh
Architecture in TDMA based Broadband Fixed
Cellular Network
Syed, Ahmed, Yanikomeroglu, Mahmoud WCNC04
91Cellular Mesh Network with Global Resource
Allocation
92Algorithm for Constructing Routing Table
Step 1 Reject all node-node node-BS links with
PLgtPLmax
Step 2 List all 2-hop 3-hop routes between
source node and BS(s)
Step 3 Arrange in ascending order,the routes
found in Step 2 using criterion min max (PLi )
where i1,2 or i1,2,3
Step 4 If tie in max(PLi), then min ? PLi
route on the top
93Route Selection
- Policy
- Minimum Number of Hops Route First
- Conditions
- Free slot(s) available on all hops
- SINRr?SINRth on the free slot(s)
- Salient Features
- Less spectral resources used (time slots)
- Viable SINR links used (SINRr?SINRth)
- Simple call admission policy
- Simple algorithm
94Adaptive Modulation Coding
SINR (dB) Code Mod. Info. bps/hz
lt4.65 - 0
4.65-7.45 ½ QPSK 1
7.45-10.93 ¾ QPSK 1.5
10.93-12.0 ½ 16-QAM 2.0
12.0-14.02 2/3 16-QAM 2.67
14.02-15.0 ¾ 16-QAM 3
15.0-17.7 7/8 16-QAM 3.5
17.70-19.0 2/3 64-QAM 4
19.0-21.94 ¾ 16-QAM 4.5
21.94-26.0 7/8 64-QAM 5.25
gt26.0 64-QAM 6
95Main Simulation Parameters
System Parameter Simulation Value
Network Cellular, Noise limited, 200 Nodes, 4-Square Cells, Cell Size 3x3 km2
Multiple Access TDMA / FDD, 10 slots/channel
Propagation Channel nn-BS 3.8, nn-n 4, ?n-BS 6dB, ?n-n 4dB
Carriers Bandwidth Single Carrier _at_ 2.5GHz, BW 5 MHzs No. of Channels 2 to 6
Antenna Type 30o, switched beam, Gml 7 dB, Gsl,bl 0 dB, Rooftop (node end)
Transmit Power Fixed, Pt 2 watts
Noise AWGN, Noise Power - 130 dBW
96Simulation Parameters Cont.
System Parameter Simulation Value
Network Traffic Traffic Arrival Poisson, ?400-8000 burst/sec Burst size Exponential, ?15 kbits
PLmax for Routing 126 dB
SINRth 4.65 dB
Frame specification Tf10ms, 10 slots/frame
Slot allocated per hop 1
Upper limit on consecutive frame drop before retransmission 3
97Simulation Assumptions
- Snap shot processing at frame level
- All transmissions are slot synchronized
- Independent fixed shadowing on all links
- Doppler shift negligible
- Multipath fading handled by micro diversity
- Infinite buffer size on the node
- All user nodes are active
- Separate control channels are available
- Continuous ARQ Protocol
98Outage Probability
99Outage Analysis of SH MH Network, 2-Channels
100Outage Analysis of SH MH Network, 4-Channels
101Outage Analysis of SH MH Network, 6-Channels
102Connectivity Analysis MH Network, 2-Channels
103Connectivity Analysis MH Network, 4-Channels
104Connectivity Analysis MH Network, 6-Channels
105Net Node Throughput
106Diversity- and AMC (Adaptive Modulation and
Coding)-Aware Routing inInfrastructure-based
TDMA Multihop Networks
Hares, Yanikomeroglu, Hashem VTCF03 Globecom03
4
1
3
2
6
0 AP
5
Time Domain MAC Frame
Connection to 3
Connection to 4
Connection to 6
Hop 0 to 4
Hop 0 to 1
Hop 1 to 2
Hop 2 to 3
Hop 0 to 5
Hop 5 to 6
Extra channels are not used. Connections and hops
are orthogonal in the time domain.
107Routing
- Routing objective
- Select relay nodes and hop modulation/coding (MC)
to maximize throughput - Throughput (Information Rate bits/sec)(1 -
Probability of error) - Increase Info. Rate or decrease end-to-end error
rate to increase throughput. How?
108Frame Allocation for Relaying
- Adaptive modulation and coding (AMC)
- Different MC used on hops
- Amount of data entering and exiting relaying
nodes are equal
QAM4
Hop0
Hop0
Hop0
2
0
QAM64
QAM64
Hop0
Hop1
Hop0
Hop1
Hop0
Hop1
0
1
2
109Multihop Diversity
Time Domain
QAM64
QAM16
QAM64
Hop1
Hop2
Hop0
Hop2
Hop1
Hop1
Hop0
Hop0
0
1
2
3
Node 2 Receiver Operation Equivalent
Symbols
Yes
Bits
Hop0
Decoder
CRC OK?
No
Symbols
Yes
Bits
Hop1
Decoder
CRC OK?
Node 3 Receiver Operation Equivalent
Hop0
MRC
Symbols
Yes
Bits
Decoder
CRC OK?
Hop2
No
Symbols
Yes
Bits
Hop1
Decoder
CRC OK?
110Adaptive Modulation Coding Maximization (AMCM)
- Originally, hop modes were selected to maximize
data rates on hops. - AMCM adapts hop modes to maximize the connection
throughput for systems using MRC diversity. - Performed after route has been selected.
- Possible modes a hop can assume is limited to the
set of modes used in the connection (i.e. QAM16,
QAM64, QPSK1/2). - For each iteration, examine all possible modes
for all hops. - Change a mode for a single hop that generates the
maximum metric for the connection. - Use the new set of modes for the subsequent
iteration. - Stop when a mode change does not increase the
metric for the connection.
Hop1
Hop2
Hop3
QAM64
QAM16
QPSK1/2
From all possible changes, changing the mode of
hop 2 from QAM64 to QAM16 generates the max.
metric.
Hop1
Hop3
QPSK1/2
QAM16
Use the new set of modes to continue to maximize
the metric.
Stop when mode changes do not increase the
connection metric.
111Routing Types
- Route Selection Strategies
- Single Hop (SH)
- Multihop (MH)
- Multihop Selection Combining Diversity (MHSC)
- Routing metric factors selection combining
diversity - Multihop MRC Diversity (MHMRC)
- Routing metric factors MRC diversity
- Multihop Adaptive Modulation MRC Diversity
(MHAMMRC) - Routing metric factors MRC diversity
- Uses AMCM
- Hybrid Digital and Analog Relaying (HDAR)
- Nodes relay incorrectly decoded signals as analog
signals
112Routing Example (1)
Iteration
Next 1
Current 0
- Initially, routes only contain the AP and
destination node. - Examine all next routes.
- Next routes (black) built off current routes
(red). - Next routes generating max. metrics are in
purple. - If metric of next route gt metric of current
route, on next iteration, current route next
route. - Next iteration routes
- B AP-A-B
- C AP-E-C
- E AP-A-E
AP
A
A
B
B
C
C
D
D
E
E
113Routing Example (2)
Iteration
Next 2
Current 1
- Only check routes to nodes not included in
current route. - Next iteration routes
- C AP-A-E-C
- D AP-E-C-D
- Need to only examine next routes built from
routes which changed on previous iteration.
AP
A
A
B
B
C
C
D
D
E
E
E
114Routing Example (3)
Iteration
Next 3
Current 2
- Stop searching when next generation of routes do
not yield higher metrics. - Final connections
- A AP-A
- B AP-A-B
- C AP-A-E-C
- D AP-E-C-D
- E AP-A-E
AP
A
A
B
B
C
C
D
D
E
E
115Simulation - Routing Example
116Simulation - Routing Example
117Simulation - Parameters
- NLOS office environment
- ETSI-A channel model
- Rayleigh fading 50ns RMS delay spread
- Noise power, PN0 -90dBm
- Propagation exponent, a 3.4
- Carrier frequency, fc 5.3GHz
- Shadowing, s 5.1dB
- Omni-directional antennas
- Fixed transmit power, Ptx 23dBm
- Adaptive modulation
- Constant interference
- Hexagonal radius, R 128m
- Cluster size, N 12
- No mobility
118Simulation Throughput, 128m Cell
119Simulation - Number of hops, 128m Cell
120Simulation Throughput, 256m Cell
121Simulation - Number of hops, 256m Cell
122Simulation Results
Routing Type System Diversity Avg. Network Throughput Mbps Avg. Network Throughput Mbps Avg. Hops in Route Avg. Hops in Route
Routing Type System Diversity 128m Cell 256m Cell 128m Cell 256m Cell
SH None 7.75 2.07 1 1
MH None 12.77 4.17 2.21 2.93
MHSC SC 13.17 4.70 2.64 4.17
MHMRC MRC 13.19 4.70 2.62 4.14
MHAMMRC MRC 13.26 4.85 2.62 4.24
MHAMMRC-HDAR MRC 14.32 5.62 2.56 3.70
- Routing Type routing/metric type
- System Diversity form of diversity used at
nodes - SH singlehop, MH multihop (routing algorithm)
- MHAM multihop adaptive modulation (routing
algorithm)
123Observations
- Multihop routes ? Optimal 2-hop routes
- Diversity techniques very attractive -- they do
not use additional radio resources (power or
bandwidth) - Routing incorporates diversity benefits
- HDAR increases diversity benefits
- Average aggregate throughput increases 2-3X
- Outage reduces very significantly (range
extension remarkable) - Strategically placed fixed relayers may be very
attractive - Can be used in any TDMA network, if
- PER models are known channel updates supported
- For more information
- Shoaev Hares, Halim Yanikomeroglu, and Bassam
Hashem, "Diversity- and AMC (Adaptive Modulation
and Coding)-Aware Routing in TDMA Multihop
Networks", IEEE GLOBECOM 2003 - Shoaev Hares, Halim Yanikomeroglu, and Bassam
Hashem, "A relaying algorithm for multihop TDMA
TDD networks using diversity", IEEE VTC Fall 2003
124Previous Standardization Efforts
- Opportunity-driven multiple access (ODMA) 1999,
3GPP - HiperLAN/2 (non-contention based multiple access)
125Current Interest in Relay/Mesh/Multihop Networks
(1)
- IEEE 802.11s WLAN (Wireless Local Area Network)
ESS Mesh Networking - Auto-configuring multihop paths between APs in a
wireless distribution system. - Targeted to be approved by 2008.
- IEEE 802.15.5 WPAN (Wireless Personal Area
Network) Mesh Networking - Aims at determining the necessary mechanisms that
must be present in the PHY and MAC layers of
WPANs to enable mesh networking. - Targeted to be approved by 2007.
- IEEE 802.16 WMAN (Wireless Metropolitan Area
Network) - 802.16-2004 standard Air Interface for Fixed
Broadband Wireless Access Systems approved in
July 2004. MAC layer supports an optional mesh
topology. - 802.16e amends 802.16 to support mobility for the
devices operating in the 2-6 GHz licensed bands. - An optional mesh mode is being considered based
on 802.16e-2005 OFDMA - MMR-SG Mobile Multihop Relay Study Group ?
802.16j - (Taipei, Sep05 Vancouver, Nov05 New Delhi,
Jan06 Denver, Mar06) - http//grouper.ieee.org/groups/802/16/sg/mmr/
- http//ieee802.org/16/sg/mmr/
-
- IEEE 802.20 MBWA (Mobile Broadband Wireless
Access) - Aims at developing the specification of PHY and
MAC layers of an air interface for interoperable
mobile broadband wireless access systems,
operating in licensed bands below 3.5 GHz,
optimized for IP-data transport, with peak data
rates per user in excess of 1 Mbps.
http//grouper.ieee.org/groups/802/11/index.html h
ttp//www.802wirelessworld.com
126Current Interest in Relay/Mesh/Multihop Networks
(2)
- Cellular 4G Networks (WINNER Project)
- Propriety solutions by industry
- BelAir, Firetide, Strix, Tropos, RoamAD, Mesh
Networks, - Nortel, Nokia, IBM,
- Booming literature
-
127WINNER Wireless World Initiative New Radio
- Integrated Project funded by European Union under
the 6th Framework Program (FP6) - Objective to develop a ubiquitous radio system
concept based on global requirements for mobile
communication systems beyond 3G. The project
covers a full scope from short-range to wide-area
scenarios and will provide significant
improvement to current systems in terms of
performance, efficiency, coverage and
flexibility. - 01 Jan 2004 31 Dec 2009 (three 24-month phases)
- 50 partners (all European, except 2 Chinese 1
Canadian Carleton) - Manufacturers, network operators, academic
institutions and research centres - Including Siemens, Alcatel, DoCoMo Europe,
Ericsson, Nokia, France Telecom, Fujitsu Europe,
IBM Europe, Philips, Samsung Europe, Vodafone,
Qualcom Europe, Nortel Europe - Relaying integral part of WINNER network
deployment concept - Check https//www.ist-winner.org ? Public
Deliverables ? D3.x
128Concluding Remarks
- Infrastructure-based multihop networks
cost-effective ubiquitous high data rate coverage
in future wireless networks - Fixed relay stations with add-on terminal relays
- Impact in all layers of wireless communications
- Propagation, PHY, MAC, networking, higher layers
and protocols - Relay networks will soon become a reality
- Goal to develop advanced cooperation protocols
and algorithms among relays and APs to obtain
further performance gains at - physical layer (cooperative diversity, virtual
antenna arrays, ) - systems layer (interference avoidance and
management) - networking layer (smart scheduling and routing,
load balancing, ) - by relying on other advanced technologies, such
as OFDM(A) and MIMO, as much as possible.
129Tutorial/Overview/Perspective Papers
- H. Yanikomeroglu "Fixed and mobile relaying
technologies for cellular networks", Second
Workshop on Applications and Services in Wireless
Networks (ASWN'02), pp. 75-81, 3-5 July 2002,
Paris, France. - H. Yanikomeroglu, "Cellular multihop
communications infrastructure-based relay
network architecture for 4G wireless systems",
the 22nd Queen's Biennial Symposium on
Communications (QBSC'04), 1-3 June 2004, Queen's
University, Kingston, Ontario, Canada invited
paper. - R. Pabst, B. H. Walke, D. C. Schultz, P. Herhold,
H. Yanikomeroglu, S. Mukherjee, H. Viswanathan,
M. Lott, W. Zirwas, M. Dohler, H. Aghvami, D. D.
Falconer, and G. P. Fettweis, Relay-based
deployment concepts for wireless and mobile
broadband radio, IEEE Communications Magazine,
vol. 42, no. 9, pp. 80-89, September 2004. - R. Bruno, M. Conti, and E. Gregori, Mesh
networks commodity multihop ad hoc networks,
IEEE Communications Magazine, vol. 43, vol. 3,
pp. 123-131, March 2005.
130Relay/multihop/mesh networks research at Carleton
University
- Halim Yanikomeroglu
- 2 PDF
- 7 Ph.D. students
- 2 M.A.Sc. students
- 6 B.Eng. students
- visiting professors and researchers
- halim_at_sce.carleton.ca
- www.sce.carleton.ca/faculty/yanikomeroglu.html