Performance of Smart Antennas and PCF

1
Performance of Smart Antennas and PCF

• Ari Alastalo, Steven Gray and Venkatesh Vadde
• Nokia Research Center

2
Introduction

• IEEE802.11 will be an important method for
  providing high rate low mobility data services
• While existing capacity may seem high,
  particularly for IEEE802.11a, the enterprise
  environment may experience capacity limits
• Antenna technologies offer a means to boast
  IEEE802.11 capacity without changing the existing
  PHYs

3
What is a SMART Antenna?
Antenna that adjusts its beam pattern based upon
the channel and interference between AP and STA
4
Approach to Model Performance

• Measure channel using a channel sounder to
  determine
• Multipath power profile as a function of time
• Signal-to-interference ratio
• Simulate PHY to obtain PER information as a
  function of channel measurements and the number
  of packets delivered as a function of time
• Use PER, number of packets delivered by the PHY
  as a function of time and models of real-time
  traffic to examine delay and throughput using a
  PHY with and without smart antennas
• PCF is used for delivery of audio and video
  packets

5
PHY Simulation Parameters
6
PHY Channel Sounder Approach

• 127 chip pn sequence is transmitted at 5.3 GHz
  with a bandpass bandwidth of 30 MHz
• A 32 element array with 0.5 wavelength space is
  used to downconvert the transmitted pn sequence
• Snap shots of the delay spread are written to
  memory and stored on a hard disk

For further information see Jarmo Kivinen, Timo
O. Korhonen, Pauli Aikio, Ralf Gruber, Pertti
Vainikainen, and Sven-Gustav Häggman, IEEE
Transactions on Instrumentation and
Measurement, vol. 48, no. 1, pp. 39-44, 1999.
7
Smart Antenna Operation (5GHz)
Uplink Operation
Downlink Operation
8
Channel Correlation

• The figure to the right shows the correlation
  coefficient for different measurement
  environments
• For smart antenna operation, the AP needs the
  ability to probe the channel between any STA and
  AP

Note Ruholahti and Heikkiläntie are Nokia office
buildings
9
MAC Simulation Parameters
10
Traffic Model

• All traffic measured in data-units/slot-time
• 1 data-unit takes 1 slot-time for transmission
• Max traffic in network 1.0
• Audio and video traffic originates from calls
  made by the user
• Calls are Poisson distributed once placed, each
  call generates periodic packet traffic
• Mean inter-call-arrival-time controls load on the
  network

11
Traffic Model (cont.)
12
Throughput Latency Curves Audio Packets
13
Throughput Latency Curves Audio Packets
14
Throughput Latency Curves Video Packets
15
Throughput Latency Curves Video Packets
16
New Control Frame Subtypes for Smart Antennas

• AP to STA
• A message requesting the STA to transmit a
  preamble (null frame) for channel estimation
• Channel Id Request
• STA to AP
• A response to the above request
• Channel Id Response

17
Remove Polling Dependence

• The existing standards reads,
• "During each CFP, the PC shall issue polls to a
  subset of the STAs on the polling list in order
  by ascending AID value".

18
Conclusions

• Particularly in a large enterprise environment,
  smart antennas can help boast capacity
• Wireless office replacement for "wired" Ethernet
• Public service networks such as airports
• Changes to the existing MAC are minor to enable
  antenna technologies in IEEE802.11a networks
• SDMA is not for all WLANS
• Multiple antennas cost additional money for the
  AP that may not be required in homes and small
  businesses