Reverse Link Power Control in Mobile Data Networks

1
Reverse Link Power Control in Mobile Data Networks

• Wiklom Teerapabkajorndet
• Telecommunication Programs
• University of Pittsburgh

2
Outline

• Introduction
• Power control game
• Pilot study
• Game design
• Performance analysis
• Conclusion

3
I. Introduction

• Research challenges in mobile data networks
• Related work
• Problem statement
• Scope of work

4
Research challenges in mobile data networks

• Is the fixed target PER (or SIR) of the existing
  power control (for voice) efficient in the case
  of mobile data networks?
• A different PER (or SIR) level for each MS might
  be more efficient.
• What approach should be used to determine the
  suitable transmitted power for each MS?
• The actions of other MSs in power control will
  affect the actions and strategies of a given MS
  making this problem suitable for a game theoretic
  analysis.

5
Related work

• C. U. Saraydar, N. B. Mandayam and D. J. Goodman,
  Efficient Power Control via Pricing in Wireless
  Data Networks, IEEE Trans. Commun., vol. 50,
  Feb. 2002 , p.p. 291 -303.
• C. U. Saraydar, N. B. Mandayam, and D. J.
  Goodman, Pricing and Power Control in a
  Multicell Wireless Data Network, IEEE J. Select
  Area Commun., vol. 19, No. 10, Oct. 2001.
• X. Mingbo, N. B. Shroff and E. K. P. Chong,
  Utility-based Power Control in Cellular Wireless
  Systems, Proc. IEEE Infocom 2001 Conference,
  vol. 1, p.p. 412-421.
• A. B. MacKenzie and S. B. Wicker, Game Theory in
  Communications Motivation, Explanation, and
  Application to Power Control, Proc. IEEE
  GLOBECOM, vol. 2, 2001, p.p. 821 -826.
• N. Feng, N. Mandayam, D. Goodman, "Joint Power
  and Rate Optimization for Wireless Data Services
  Based on Utility Functions," Proc. of CISS, 1999.

6
Drawback in related work

• Utility function
• Bits per joule
• Sigmoid function of SIR
• Performance metric
• Utility and transmitted power
• Missing piece
• How are these power control games better than the
  existing scheme?

7
Problem statement

• How to assign the transmitted power for each MS
  that accounts for unequal target SIRs such that
  each of them obtains the equilibrium BER that is
  at least more efficient than the existing power
  control scheme (fixed target SIR)?

8
Scope of this paper

• The new idea of unequal target SIR is proposed
  for reverse link power control for non-real time
  data MSs in a single cell
• A new utility function is invented as a function
  of an effective rate which can represent the
  satisfaction of non-real time data-MS in the
  sense of throughput.
• Pricing that is also included in the utility
  function is used in designing the power control
  game to obtain a low BER at least less than the
  existing fixed target PER.
• The BER performance of our proposed power control
  is compared with the existing scheme (fixed PER)
  when mobility is considered.

9
II. Power control game

• Problem formulation
• Nash equilibrium
• Equilibrium existence and uniqueness
• Numerical method solution approach
• Computational implication

10
Problem formulation

SIR
Modulation scheme
11
Global maximum of the utility function
12
Existence of the Nash equilibrium
13
Existence and uniqueness of the Nash equilibrium
14
Numerical solution approach (1)

• for all
• subject to

Newtons numerical method
Computational implication of Newtons numerical
method
15
Numerical solution approach (2)
Iterative step and convergence check are in the
paper
16
III. Pilot study of a two player power control
game
BER of a moving MS
BER of a stationary MS
17
IV. Game design cell partitioning

• Each range has different level of c as specified
  below.
• A (0-250m) 287470 bps/W
• B (250-500m) 17970 bps/W
• C (500-750m) 3550 bps/W
• D (750-1000m) 1120 bps/W
• Each c is determined approximately at the
  boundary of each range that can achieve 1 packet
  error rate.

18
V. Performance analysis

• Experimental design
• Numerical results

19
Experimental design scenario

• Compare with the fixed target PER 1 or SIR
  13.4905 and fixed pricing coefficient game
• N MSs in the cell consist of one moving tagged MS
  and N-1 stationary MSs.
• The stationary MSs locate between the BS and the
  cell boundary with equidistance to each other.
• The cell radius in this study is 1000 m.
• The tagged MS is moving at constant speed at 50
  km/h.
• The update interval of power control is 0.0017 s.
• Maximum power is limited at 1 W.
• Assume equal and constant transmission rate for
  all MSs at 10 kbps.

20
Experimental design parameters

• Response variable BER (bit error rate)
• Factor and levels
• Number of MSs in the cell
• 2 levels 4 and 7 MSs

Cell capacity for PER 1 (calculated by a
simple capacity estimation)
21
Numerical results
22
4 user case
Figure 10 BER of MS2 (the MS that locates at
the middle between the BS and the cell boundary)
Figure 9 BER of MS1 (the closest MS to the BS)
23
4 user case
Figure 11 BER of MS3 (the MS locates at the cell
boundary)
Figure 12 BER of MS4 (the moving MS)
24
7 user case
Figure 14 BER of MS2 (the MS at the middle
between the BS and the cell boundary)
Figure 13 BER of MS1 (the closest MS to the BS)
25
7 user case
Figure 15 BER of MS3 (the MS at the cell
boundary)
Figure 16 BER of MS4 (the moving MS)
26
VI. Conclusion

• The new idea of unequal target SIR for power
  control is proposed for mobile data networks.
• BER, a new metric in power control game, is
  introduced here.
• Mobility is firstly included in the study of
  power control game.
• A new utility function is invented as a function
  of an effective rate which can represent the
  satisfaction of non-real time data-MS in the
  sense of throughput.
• Pricing, included in the utility function, is
  used in designing the power control game to
  obtain low BER at least less than the existing
  fixed target PER.
• Cell partitioning for adaptive pricing power
  control game is a result of our game design.
• Our initially designed game shows some potential
  results that can be further developed to achieve
  our objective of power control game design.