Joint PHY-MAC Designs and Smart Antennas for Wireless Ad-Hoc Networks

1
Joint PHY-MAC Designs and Smart Antennas for
Wireless Ad-Hoc Networks

• CS 838 - Mobile and Wireless Networking
• (Fall 2006)

2
Review of IEEE 802.11a/b/g PHY/MAC

• PHY
• Modulation Orthogonal Frequency Division
  Multiplexing (OFDM) (11a/11g) or Direct Sequence
  Spread Spectrum (DS-SS) (11b/11g)
• Antenna Technology Single omni-directional
  antenna
• 2 antenna Access Points (APs) ?
• APs with directional antennas ?

• MAC
• Physical Carrier Sensing Carrier Sense Multiple
  Access with Collision Avoidance (CSMA/CA)
• Virtual Carrier Sensing Request-to-Send/Clear-to-
  Send (RTS/CTS) handshake (Hidden node avoidance)

3
Review of Hidden Node Problem

• Pure CSMA/CA

Carrier Sense? Clear
4
802.11 Solution of Hidden Node Problem

• CSMA/CA with RTS/CTS

Virtual Carrier Sense? Busy
RTS
CTS
5
Some Limitations of 802.11 PHY/MAC

• PHY
• Throughput (bits/sec) Antenna technology limits
  spatial re-use

Physical Carrier Sensing
Virtual Carrier Sensing
6
Some Limitations of 802.11 PHY/MAC (contd.)

• MAC
• Throughput RTS/CTS handshake further limits the
  spatial re-use in the network
• Fairness (and Throughput) RTS/CTS fails to
  completely take care of the hidden node problem,
  resulting in dropped packets for one transmission
  more than the other
• Interference range is typically more than the
  successful reception range of CTS
• Fairness 802.11 MAC can unfairly favor one
  transmission over the other as a function of the
  distance between the nodes

X
CTS
7
802.11 Simulated Performance TPHN04
Linear Topology
2
3
1
0
200 m
200 m
D
Throughput reduction (and unfairness) due to
spatial proximity
Hidden node effect
8
TPHN04 Proposed Solution

• PHY
• Single transmit/multiple receive antennas with
  OFDM modulation

• MAC
• Mitigating Interference using Multiple Antennas
  MAC (MIMA-MAC)
• Built on top of 802.11 MAC with antenna awareness
• N nodes in spatial proximity would be allowed to
  transmit simultaneously in a network of nodes
  with N receive antennas
• PHY expected to cancel the interference of (N-1)
  unintended flows using advanced signal processing
  techniques

9
TPHN04 Simulation Results
Topology

• Simulation Technique
• PHY Simulation MATLAB (with channel bandwidth of
  2 MHz and data rate of 1 Mbps)
• MAC Simulation ns-2 (fed with look-up tables
  mapping channel realizations to corresponding
  BERs obtained from MATLAB)

• Other Parameters
• Input SNR 12dB Path-loss exponent 4 Packet
  reception threshold BER 10-5 Carrier sensing
  threshold BER 10-1

10
TPHN04 Simulation Results (contd.)
Throughput Performance for Multiple Receive
Antennas MIMA-MAC vs Conventional 802.11 MAC
11
TPHN04 Simulation Results (contd.)
Fairness Performance for Multiple Receive
Antennas MIMA-MAC vs Conventional 802.11 MAC
12
Smart Antennas for Wireless Ad-Hoc Networks

• Switched Beam Antennas
• Pre-determined set of weights applied to
  different antenna elements to form a fixed number
  of high-directionality beams
• A K element array can form up to K beams

• The directionality gain of each beam at the
  transmitter and the receiver is given by
  (assuming LOS/low angular spread)
• Assuming that the transmitter and the receiver
  know each others direction, the total
  transmission gain (SNR gain) is bounded by

13
Smart Antennas for Wireless Ad-Hoc Networks

• Fully Adaptive Arrays
• Fully adaptive set of weights applied to
  different antenna elements to adaptively change
  the radiation pattern
• A K element array has K degrees-of-freedom
  (DOFs), and can adaptively null (K-1)
  uncorrelated interferers

• Even in the presence of significant multipath
  scattering, the total transmission gain (SNR
  gain) of an adaptive array can be given by
• Very high multipath scattering and low signal
  correlation can some- times limit the gain to

14
Smart Antennas for Wireless Ad-Hoc Networks

• MIMO Links
• Digital adaptive arrays capable of operating in
  two modes Spatial Multiplexing and Diversity
• A rich set of multipath scattering between the
  transmitter and the receiver transforms a K
  element MIMO link into K independent links

• In multiplexing mode, this can result in K fold
  increase in the data rate of the MIMO link
• In diversity mode, this can result in a reduction
  in the variance of the received SNR. At high SNR,
  this results in

15
Smart Antennas can be leveraged for

• 1) Higher Data Rate
• For a given modulation scheme, the bit-error-rate
  (BER) on a link is determined by the link SNR
• Switched Beam/Adaptive Array Gain in SNR (G) ?
  Perform adaptive modulation to increase bits
  transmitted per symbol and keep BER the same
• MIMO Link Operate the link in the spatial
  multiplexing mode

16
Smart Antennas can be leveraged for

• 2) Increased Transmission Range
• The transmission range of a link is related to
  the link SNR by
• Switched Beam/Adaptive Array Gain in SNR (G) ?
  Obtain a range extension factor given by
• MIMO Link Operate the link in the diversity
  mode. Not a straight forward relationship between
  the diversity order and the range extension, so
  resort to MATLAB simulations (diversity mode only
  reduces SNR variance)

17
Smart Antennas can be leveraged for

• 3) Increased Link Reliability
• For a fixed data rate (modulation scheme), the
  bit-error-rate (BER) on a link is determined by
  the link SNR
• Switched Beam/Adaptive Array Gain in SNR (G) ?
  For the same data rata, obtain a reduction in the
  BER by a factor of
• MIMO Link Operate the link in the diversity
  mode. For the same data rate, this can result in
  a reduction in the BER by a factor of

18
Smart Antennas can be leveraged for

• 4) Reduced Transmit Power
• Switched Beam/Adaptive Array Gain in SNR (G) ?
  For the same BER, obtain a reduction in the
  transmit power by a factor of
• MIMO Link Operate the link in the diversity
  mode. For the same BER, this can result in a
  reduction in the transmit power given by

19
SLS06 Simulation Model

• Antenna Model
• Switched Beam Array Pre-determined, fixed beam
  pattern
• Adaptive Array/MIMO Link Dynamically tunable
  beam pattern

• Channel Model
• PHY BER obtained from MATLAB simulations by
  assuming a fast Rayleigh fading collision channel
  model (per location, antenna technology and
  strategy), with data rate of 2 Mbps, transmit
  power of 20 dBm, SINR of 10 dB and fade margin of
  0-10 dB
• Link Packet loss probability obtained from ns-2
  (fed with look-up tables of PHY simulations),
  with packet size of 1000 bytes

20
SLS06 Simulation Model (contd.)

• Network and Traffic Model
• 100 nodes over a rectangular grid of 400x400 m to
  1000x1000 m
• Number of simultaneous flows in the network
  varied from 1 to 50
• Multipath scattering varied from LOS to 180
  degrees (rich scatter)
• Number of antenna elements per node varied from 1
  to 12
• Initial transmission range of each node set to
  100 m

• Metrics
• Throughput (T) Bits per second, normalized by
  the number of flows
• Throughput/Energy (TE) Bits per unit of Joule
  consumed (consisting of communication circuit
  power Pc, transmit power Pt and computational
  power)

21
SLS06 Simulation Model (contd.)

• Protocols and Algorithm
• Goal Obtain fundamental tradeoffs in the
  operation of different antenna technologies
• Requires Suppressing the inefficiencies of other
  factors
• Solution Centralized algorithm for finding
  routes, scheduling slotted transmissions,
  ensuring fairness, taking care of interferences
  etc.
• Routing Strategy Djikstras algorithm

• Caveat
• Simulation results are not indicative of how
  things might perform in a distributed setting

22
SLS06 Strategy Comparison T Metric
Switched Beam
Adaptive Array
MIMO Links

• Setup
• High density network, load of 50 flows, fading
  loss of 5, scattering angle of 90 degrees

23
SLS06 Strategy Comparison T Metric (contd.)

• Exceptions
• Under low node density and small number of flows,
  range works better (better connectivity)

4 Antenna elements per node
24
SLS06 Strategy Comparison TE Metric
Switched Beam
Adaptive Array
MIMO Links

• Setup
• High density network, load of 50 flows, fading
  loss of 5, scattering angle of 90 degrees and Pt
  gtgt Pc

25
SLS06 Strategy Comparison Inferences
Moderate-High Network Densities
Low Network Densities
26
SLS06 Antenna Technology Comparison

• Parameters of Interest
• Network node density
• Number of antenna elements
• Number of network flows
• Scattering angle
• Fading loss

• Components Impacting T and TE Metrics
• Number of independent contention regions ?
  density
• Number of active links/contention region ? flows
  and density
• Number of resources/contention region ? elements
  and scattering

27
SLS06 Technology Comparison T
Metric(Scattering and Elements under Rate
Strategy)
Rich Scattering
Low Scattering
Rich scattering does not degrade MIMO links rate
performance
28
SLS06 Technology Comparison T
Metric(Scattering and Fading under Rate Strategy)
Fading impacts all rate strategies alike!
29
SLS06 Technology Comparison T Metric(Flows
and Elements under Rate Strategy)
Low Load
High Load
No logarithmic effect for MIMO at low load
30
SLS06 Technology Comparison T Metric(Flows
and Density under Rate/Range Strategy)
Low Load High Density
Low Load Moderate Density
Low Load Low Density
31
SLS06 Technology Comparison TE Metric(Pt gtgt
Pc Power/Range Strategy)
Other Network Conditions
Low Scattering Large Elements
32
SLS06 Technology Comparison TE Metric(Pt lt
Pc Rate/Range Strategy)
Majority of Network Conditions