IIT Kharagpur

1
IIT Kharagpur
A Simulation Testbed for Performance Evaluation
of Open Loop Power Control Algorithms for
Cellular CDMA
Joydeep Acharya Sumit Kundu Umesh Jalan Saswat
Chakrabarti
2
Introduction to Power Control
Reception Problems

• Near/Far Problem
• Multipath Fading
• Multiple Access Interference (MAI)
• Other Cell Interference

Solution Power Control
This ensures that the received signal strengths
for all users are same at the base station
3
Channel Model
In a mobile radio environment the channel power
gain consists of short term Rayleigh fading, a
propagation gain associated with slow lognormal
fading and path loss
The propagation gain can be modeled as
A constant that depends on the antenna gains,
signal wavelengths etc. m path loss
exponent x zero mean Gaussian random variable
with variance ?2 and r normalized distance
between the base station and mobile.
4
Classifications of Power Control Algorithms
Broad Classification

• Open Loop Algorithms
• Closed Loop Algorithms

Types of Open Loop Algorithms

• Power Truncation
• Power Limitation

5
Power Truncation Algorithm

• If Lk is the channel power gain at time t, then
  the transmitted power
• from a mobile is
• S(Lk) SR/ Lk, Lk ? ?o
• 0 , Lk lt ?o

• The fraction of time for which the mobile
  transmits, p(?o), is given by

• The average transmit power is given by

6
Power Truncation Algorithm (Continued)
ST has been assumed to be a constant irrespective
of ?0. So the target received power SR is
actually a function of ?0 given by
If Eb is the received energy per bit the power
gain Gp over perfect power control is
7
Power Limitation Algorithm
Transmission of a mobile is limited to a fixed
level (instead of suspending the transmission)
when the mobile is in deep fade.
If the time limit, for which a call is limited is
chosen as unity
S(Lk) SR/ Lk Lk ? ?o , t tj


SR/?, Lk lt ?o, t tj
Lk ? ?o, t tj-1
0, Lklt ?o, t tj
Lklt ?o, t tj-1
8
Work Done

• Development of a simulation testbed for
  testing open loop
• power control algorithms.
• Using this testbed for studying the
  performance of the
• power truncation and power limitation
  algorithms and
• comparison of their results.

9
Simulation Testbed Assumptions

• A single cell CDMA environment.
• L(t), the channel gain is estimated perfectly The
  effect due to
• Rayleigh fading, R(t) is to be taken care of by
  a RAKE receiver at
• the receiver end.
• Call arrival is Poisson distributed with mean (?)
  3. Initial traffic
• at the start of simulation has been assumed to be
  10.
• The mobiles location is uniformly distributed
  within a cell.
• This leads to a probability density function
  (pdf) of r given by

10
Simulation Testbed Procedure

• 1000 simulation cycles of duration 1 sec each

• A mobile unit moves an incremental distance in a
  direction ?
• (uniformly distributed between 0,2?) in each
  cycle

• The average call duration has been assumed to be
  100 sec.

• The ?0 threshold is chosen in such a way that
  p(?0) or the
• percentage of time the mobile is active has a
  value of between
• 0.5 and 0.9.

• mC denotes the monitor count i.e. the no. of
  cycles the call is
• limited

11
Observations
Parameters chosen m 4, ? 8 dB ?0 5, mC 3
Values obtained
32
95
12
Observations
Parameters chosen m 4, ? 8 dB ?0 0.5, mC
3
Values obtained
95
211
13
Observations
Parameters chosen m 4, ? 8 dB ?0 0.5, mC
3
Values obtained
?S 5/49
14
Observations
Parameters chosen m 4, ? 8 dB ?0 0.5, mC
3
Values obtained
?L varies from 0.11 to 0.35 with mean of 0.22
15
Observations
Parameters chosen m 4, ? 8 dB mC 3
Parameters defined
?S Survival Ratio of a Call
Average Survival ratio for all calls
p(?0)
Fig 5. Average survival ratio for all calls vs
p(?0)
16
Observations
Parameters chosen m 4, ? 8 dB
17
Observations
Values obtained
? 10 dB

• (1 - p(?0) ) 10-4 and ? 6 dB
• (GP)analytical 0.0128 dB
• (GP)simulation 0.0130 dB
• (error 1.56)
• ? 8 dB, (GP)analytical
• 0.091 dB and (GP)simulation
• 0.0915 dB (error 5.49),
• ? 10 dB (GP)analytical
• 0.491dB and (GP)simulation
• 0.4916dB (error 1.22)

? 8 dB
GP (dB)
? 6 dB
1 p(?0)
Fig 7 Analytical values of power gain Gp vs
truncation probability 1 p(?0) for several
values of ?, for m 4
18
Conclusion
We have simulated the performance of power
truncation and power limitation algorithms in
single cell CDMA systems with lognormal fading
and path loss. We showed that power limitation
leads to more number of active users in the
system and thus leads to an increase in average
traffic. This increase depends on the threshold
value of ?0 and also the duration of monitoring
period. We also verified that power truncation
scheme exhibits a power gain of 1.3-1.4 dB
relative to conventional power control for
truncation probabilities between 10-2 and 10-1.
19
Thank You