Performance of a TreeBased Collision Resolution Algorithm in Cellular Systems with Smart Antennas Co

1
Performance of a Tree-Based Collision
ResolutionAlgorithm in Cellular Systems with
Smart Antennas-- Courtesy of Dr. Haipeng Jin
and Dr. Anthony Acampora
Song Mai Electrical Computer Engineering
Department May 3rd, 2007
2
Outline
Introduction MAC Protocol The Collision
Resolution Algorithm Performance
Analysis Numerical Results Conclusion
3
Introduction

• Smart Antenna
• A smart antennas ability to simultaneously
  resolve more than one user on the same channel is
  exploited to help expedite the process of random
  access, especially for the reverse channel.
• Use of a smart antenna requires frequent array
  adaptation, a difficult task in a random access
  environment.
• MAC Protocol
• A Media Access Control (MAC) protocol with smart
  antenna
• uses training fields sent by mobile stations to
  allow the base station to rapidly acquire the
  interference pattern, fully exploiting the space
  division multiple access capability of a smart
  antenna to allow multiple mobiles to transmit or
  receive information simultaneously.
• one reservation slot assigned to each active
  mobile, used to send requests to the base
  station. But to initially obtain a reservation
  slot, each mobile needs to send its requests
  through the random access slots at the beginning
  of the frame.

4
Introduction (cont.)

• Tree-Based Collision Resolution Algorithm
• study the random access performance of the
  proposed protocol when a fast collision
  resolution algorithm is used
• Found are the maximum achievable throughput and
  an upper bound on the expected delay
• The impact of the number of antennas on the
  performance with both flat fading and frequency
  selective fading is studied
• Typical results show significant improvement in
  throughput for systems with smart antennas

5
MAC Protocol

• Intended for bursty data traffic, using Media
  Access Control (MAC) protocol
• seeks to insure an orderly sequencing of packets
  from the various mobile stations onto the shared
  channel,
• with a minimum of time lost to collisions. MAC
  protocol delivers bandwidth on-demand, where a
  mobile having a greater volume of packets to send
  contends more frequently for the channel.
• Applied to cellular radio systems, a MAC protocol
  must also cope with the various impairments
  suffered on the radio link such as multi-path
  fading, shadowing, and co-channel interference
  from other mobiles
• This is especially troublesome since not all
  receivers will hear all transmissions with the
  same intensity, making access cooperation among
  the mobiles more difficult to achieve.
• In addition, the reverse and forward traffic may
  be both highly dynamic and highly asymmetric,
  implying that the full bandwidth channel should
  be shared by the base station and all mobile
  stations within the cell.
• The addition of smart antennas at the base
  station makes the MAC problem even harder. Since
  the specific mobiles seeking to transmit in a
  given slot depends on the dynamic random traffic,
  
• A fast and efficient mechanism must be provided
  to enable proper channel estimation. As a
  solution, we insert training fields in the
  reverse channel radio burst to allow the smart
  antennas to tune slot by slot.

6
MAC Protocol (cont.)

• A reservation based MAC protocol with time
  division duplexing structure for cellular systems
  using smart antennas Shown in Fig. 1

7
MAC Protocol (cont.)

• Reverse traffic transmission period
• mobiles will simultaneously transmit information
  to the base station, with each transmission
  preceded by a training sequence
• enable the base stations smart antenna to
  acquire the proper combining weights
• Forward information transmission period
• each mobile is to receive from the base station
  first sends a training sequence to the base
  station, enabling the base station to acquire
  correct antenna combining weights
• base station will use those weights to send
  information to the intended mobiles
• Reservation and grant period
• used to send requests to the base station and to
  get time slot assignment information from the
  base station
• Random access and broadcast ACK period
• Initially, each mobile that needs to transmit
  will send its random access requests to the base
  station via the random access slots.
• base station broadcasts acknowledgements in the
  corresponding ACK slot, about the contention
  results

8
MAC Protocol (cont.)

• random access requests from the mobiles can still
  take advantage of the smart antenna by employing
  acquisition schemes
• using the reservation slots, unnecessary
  contention for random access slots can be avoided
  and collisions in the random access slots will be
  greatly reduced
• when antenna elements is fixed, the throughput at
  first increases, as more added to each traffic
  slot, and then diminishes, reaching maximum.
  Decreasing the number of mobiles leads to
  under-utilization of the smart antennas
  capability, while allowing more mobiles to
  transmit at the same time causes excessive
  interference and results in decreased throughput
• fast and efficient tree splitting algorithm is
  used to resolve any collisions

9
The Collision Resolution Algorithm

• Collision Resolution Period (CRP)
• The time to resolve one set of collision mobiles
• Operation of the
• collision resolution
• algorithm is illustrated
• in Fig. 3a,b
• Tree splitting
• Collision resolution

10
The Collision Resolution Algorithm (cont.)

• In general, at the beginning of a CRP, a number,
  n, of mobiles transmit requests to BS
  simultaneously in one slot. i requests can be
  captured by BS equipped with smart antennas.
• If i equals n, the CRP stops,
• remaining n-i mobiles divided into two subsets
  containing j and n - i - j mobiles,
• j mobiles start a CRP, while n - i - j mobiles
  wait when first subset completely resolved.
• when a CRP finishes, successful and null
  transmission slots is exactly one more than slots
  containing collisions.
• resolution process forms a complete tree on which
  internal nodes represent slots with collisions
  and the leaf nodes represent successful or unused
  slots.
• mobiles can start to retransmit by counting these
  slots.
• Two counters are maintained at each mobile by
  distributing information from BS to send
  acknowledgement to the involved mobiles.

11
Performance Analysis

• Length of a CRP
• Ideal case
• s of Random access mobile v.s. s of antenna
• Success and fail of transmissions
• Non-ideal case
• Success probability of reception
• Single collision set
• Ideal case
• Ln 1 for all n M
• When n gtM,

12
Performance Analysis (cont.)

• Non-ideal case
• Ln can be approximated by a linear function in n
• Ideal case
• Li ai b, i lt n, a and b are constants and
  setting b -1
• Ln an b an - 1,

13
Performance Analysis (cont.)

• Non-ideal case
• Ln a(n - i) 2b
• Maximum achievable throughput
• Expected length of CRP with rate ?
• throughput for the algorithm to run stably

14
Numerical Results

• Multi-path fading situation
• the larger the number of antenna elements, the
  greater the probability of successful reception

15
Numerical Results (cont.)

• Throughput with smart antenna
• Smart antennas greatly improve the maximum
  achievable throughput.

16
Conclusion

• the capability of a smart antenna to receive
  simultaneous transmissions from multiple mobiles
  can be used to improve the random access
  throughput in a cellular system
• The space division multiple access capability of
  a smart antenna is effectively combined with the
  tree splitting collision resolution algorithm.
• For MAC protocol, random access is limited to a
  small portion of the MAC frame, are applicable to
  a pure random access scheme where both access
  requests and traffic transmissions need to go
  through the random access process.

17
REFERENCES

• 1 H. Jin and A. S. Acampora, MAC protocol
  design and performance study for cellular systems
  using smart antennas Part I. Flat fading, IEEE
  Trans. Wireless Commun., vol. 4, no. 2, pp.
  792801, 2005.
• 2 , Performance of a MAC protocol for
  multicell cellular systems with space time
  processing (invited paper), in Proc. European
  Wireless, Feb 2002, vol. 2, pp. 652662.
• 3 J. H. Winters, Optimum combining in digital
  mobile radio with cochannel interference, IEEE
  J. Select. Areas Commun., vol. 2, no. 4,
  pp.528539, July 1984.
• 4 J. Litva and T. K. Y. Lo, Digital Beamforming
  in Wireless Communications. Boston, MA Artech
  House, 1996.
• 5 J. I. Capetanakis, Tree algorithms for
  packet broadcast channels, IEEE Trans. Inform.
  Theory, vol. 25, pp. 505515, Sep. 1979.
• 6 U. Vornefeld and E. Weltersbach, Fast
  collision resolution with SDMA for the DSA MAC
  protocol for wireless ATM networks, in IEEE 9th
  Symposium Pers. Indoor Mobile Radio Commun., Sep.
  1998, pp. 158162.
• 7 J. Ward and R. T. Compton, High throughput
  slotted ALOHA packet radio networks with adaptive
  arrays, IEEE Trans. Commun., vol. 41, no. 3, pp.
  46070, Mar. 1993.
• 8 C. Sakr and T. D. Todd, Carrier sense
  protocols for packet switched smart antenna
  basestation, in Proc. IEEE Int. Conf. Network
  Protocols, Oct. 1997, pp.4552.