Cell Breathing and Coverage Prediction in WCDMA

Cell Breathing and Coverage Prediction in WCDMA

Mohd Dani Baba, Muhammad@Yusoff Ibrahim, Farah Jida Abd Ghafar Faculty of Electrical Engineering

Universiti Teknologi MARA 40450 Shah Alam, Selangor, Malaysia

[email protected]

Abstract—Cell breathing technique is commonly used in cellular network such as Wide Code Division Multiple Access (WCDMA) to overcome load congestion by expanding or shrinking the cell coverage. This paper presents the study of the cell breathing boundary of urban environment for WCDMA. The base station cell coverage may expand or shrink depending on the number of mobile station being served. The COST231 HATA model is used in this study to predict the cell coverage for urban environment. The actual received signal strength indicator (RSSI) obtained by conducting drive test measurement are compared to the received signal code power (RSCP) of the COST231 HATA propagation model. The error calculated in the measurements enables to determine the minimum discrepancy between RSSI and RSCP data. Thus the COST231 HATA model can be used to predict correctly the cell coverage for urban environment. The result also shows that during light load the cell coverage radius is 0.18km and then decreases to 0.15km so as to limit soft handover and hence reduces load congestion.

Keywords- cell breadthing; cellular network; soft handover; propagation model;

I. INTRODUCTION WCDMA is the commonly used air interface for

Universal Mobile Terrestrial System (UMTS). The air interface is developed by the third generation partnership project (3GPP), which is the joint standardization body from Europe, Japan, Korea, USA and China. In 3GPP, WCDMA is called Universal Terrestrial Radio Access (UTRA) which covers the Frequency Division Duplex (FDD) and Time Division Duplex (TDD) [1]. 3G systems enable people to communicate with each other with high quality images and videos besides easy downloading of data and video [2]. WCDMA follows the CDMA spreading codes to spread the user information bit over a wide bandwidth by multiplying the user data with quasi-random bits or sometimes called as chips. The chip rate for 3G system is 3.84 Mcps with bandwidth approximately 5 MHz.

II. LITERATURE REVIEW Several researchers have adopted the concept of cell

breathing in Wireless LANs. For example [3] have applied cell breathing technique to balance the load between clients and access points (APs) by controlling the transmitted power of an AP. Cell breathing allows the available AP to

reconfigure the cell boundaries by changing the transmitted power. Moreover, by increasing or decreasing the transmitted power will affect the AP coverage size.

Currently the publication on cell breathing for UMTS is very limited. Nevertheless [4] have studied the cell breathing concept used in WCDMA radio network. The pilot energy per-chip to total wideband interference density ratio or Ec/Io is the main parameter involved in the study. They believed the change in value of common pilot power can improve the operability of the network and it is implemented using software for load and coverage balancing.

Another research concerning cell breathing concept was conducted by [5] and they focus on Call Admission Control for Wideband CDMA. According to this paper, a current user in a call admission is affected when a new call is accepted due to cell breathing. The call admission decision is made by considering the CDMA uplink power from mobile station (MS) to base station (BS) and discovered that due to frequency reuse, cell size is continuously reduced. Thus increases the number of handoffs of MS from one cell to another [6].

III. METHODOLOGY An urban area located in Kuala Lumpur (KL) has been

selected for this study. Drive tests were conducted in the morning and afternoon. This is to analyze the cell breathing margin of the base station. In real network, the probability of MS to be served by the same base station at different times is rather small.

A. COST231 HATA Propagation Model A model that is widely used for predicting path loss in

urban environment is the COST231 HATA radio propagation model [7]. The simplicity and availability of correction factors make it favorable for path loss prediction. The equation for path loss model in dBm is as shown in (1).

( ) ( ) GhmaRBAPL +−+= 10log (1)

where

( ) ( )txc hfA 1010 log82.13log9.333.46 −+=

( )txhB 10log55.69.44 −=

( )hma is the correction factor

2009 International Conference on Computer Technology and Development

978-0-7695-3892-1/09 $26.00 © 2009 IEEE

DOI 10.1109/ICCTD.2009.24

287

2009 International Conference on Computer Technology and Development

978-0-7695-3892-1/09 $26.00 © 2009 IEEE

DOI 10.1109/ICCTD.2009.24

287

( ) ( )( )( )( )8.0log56.1

7.0log1.1

10

10

−−−=

c

c

fhmfhma

The COST231 HATA model is restricted to the following range of parameters as shown in Table 1.

TABLE I Parameters COST231 HATA

Carrier frequency, fc 150 MHz to 1500 MHz

Antenna height of base station,

htx (m)

30m to 200m

Antenna height of mobile

station, hrx (m)

1m to 10m

Distance between base station

and mobile station, R (km)

1km to 20km

Table 1: Parameters for COST231 HATA model

B. Signal Strength, Base Station and Mobile Station Data The received signal strength indicator (RSSI) is recorded

from the drive test measurement while the information on BS and MS were provided by the local network operator. In this study, the urban area located in Kuala Lumpur is selected. All required information was determined such as position of BS and MS (longitude and latitude), 3G site identification, site name and scrambling code. The sample of RSSI data from the drive test is as shown in Fig. 1.

Figure 1: RSSI data from drive test

IV. RESULTS AND DISCUSSION

A. Cell Breathing Margin The cell size can shrink or expend depending on the

number of MS attempts in order to maintain the admission power control of MSs. This technique is called cell breathing. In this study, a cell with scrambling code (SC) of 129 is chosen as an example to determine the changes in the cell size.

Figure 2: Normal pdf for SC 129 at two different time (morning and afternoon)

Fig. 2 above shows the normal probability distribution function (pdf) for a cell with SC 129 at two different times, namely in the morning and afternoon. The data for both graphs were obtained during drive test by using the software tool. The drive tests were conducted in the morning (09.00 to 10.00) and afternoon (12.00 to 13.00) at the same selected route in Kuala Lumpur. Fig. 3 and Fig. 4 represent the graphs of Common Pilot Chanel (CPICH) Ec/No power (dBm) against distance (km) at BS with SC 129.

Figure 3: Ec/No against distance at SC 129 on first day

288288

The first graph in Fig. 3 shows that the BS is served by the MS at 0.2km away with the value of Ec/No of about -13 dBm before increases to -9 dBm at 0.21km and maintain until it reaches 0.24km. Then Ec/No drops to -16 dBm at 0.26 km and increases drastically at 0.27 km. The drop in value shows that a soft handover has occurred between SC 129 and SC 278 without changing the serving cell. Hence, a MS is still being served by SC 129. The value of Ec/No becomes worst when it reaches a distance about 0.38km. Ec/No is reduced constantly until it reaches approximately -16 dBm. Hence, at this stage the SC 129 cell is out of its capability to serve the MS at distance above 0.38 km. Then, it may assume that the radius of coverage served by SC 129 in the morning is about 0.18km.

Figure 4: Ec/No against distance at SC 129 on second day The second graph in Fig. 4 is tabulated from the data

collected in the afternoon for SC 129. The graph shows the SC 129 begins serving the MS at 0.22km. Then after two seconds, the Ec/No power decreased to -12 dBm at 0.25km. This drop in value occurs because of BS handover. Then after 0.26km, the power increases to -8 dBm at 0.27km. Then it drops gradually at 0.34km because a soft handover occurs between SC 129 and SC 278. However, the MS is still served by SC 129 because the Ec/No have not reach the threshold value of -16 dBm. At 0.35km, the Ec/No decreases rapidly and reaches -16 dBm at 0.37km. From the discussion of Figure 3 and Figure 4 above, the coverage radius served by SC 129 in the morning is about 0.18km and 0.15km in the afternoon. Thus, the coverage radius served by SC 129 becomes smaller in the afternoon which is 0.15km. Hence the coverage radius served by SC 129 is decreased by about 16.67 %.

B. RSCP of Kuala Lumpur is compared with RSCP of Putrajaya and Cyberjaya

When the Received Signal Code Power (RSCP) of Kuala Lumpur is compared to Putrajaya and Cyberjaya, their pattern of the RSCP signal were rather quite similar to each other.

Figure 5: RSCP signal at Kuala Lumpur

Figure 6: RSCP at Putrajaya and Cyberjaya

Fig. 5 shows that the range of RSCP is from -100 dBm to -64 dBm. The peak value of RSCP is about -64 dBm when the MS reached 96km. In WCDMA, usually when RSCP value is below -95 dBm the current serving BS will handover the process of serving that MS to another BS that has good RSCP value. This is because when the RSCP value is less than -95 dBm it is not good enough to support the coverage, in addition the presence of interference from other MS.

While Fig. 6 shows the ranges of RSCP signal from -110 dBm to -55 dBm for the Putrajaya and Cyberjaya locations. According to the local network operator, the values for RSCP can be classified into three. The first class is strong RSCP where its value is greater than -85 dBm. The second class is poor RSCP for value from -95 dBm until it reaches -100 dBm. The third class is weak RSCP for value below -100 dBm to -120 dBm. In urban area, the value of RSCP is usually observed in the range -50 dBm to -120

289289

dBm. The network operator has decided the first six levels; from -30 dBm to -100 dBm are the good RSCP values.

Fig. 7 and Fig. 8 show the minimum error in signal strength for both selected locations. From Fig. 7 the mean error for Kuala Lumpur is around 0 dB while for Putrajaya and Cyberjaya is about 9 dB. From Fig. 8, the probability of RSSI or RSCP at Kuala Lumpur location to reach 0 dB is 0.39 while for Putrajaya and Cyberjaya is 0.25.

Figure 7: Normal pdf for both locations

Figure 8: Empirical cdf for both locations

C. Effect of Cell Breathing Table 2 shows the total number of soft handover

increases when the number of Radio Access Bearer (RAB) attempted increases. The MS shared the CPICH power among themselves and hence the BS cell coverage is reduced. For non- peak hour (09:00 – 10:00 am), the total number of soft handover is less about 7.7 % from the peak hours (12:00 – 01:00 pm).

TABLE II

Time Duration

Average no. of RAB

attempted

SHO: No

SHO: Yes

09:00 – 10:00 am

97 (44.29 %)

19 (73.08 %)

7 (26.92 %)

12:00 – 01:00 pm

122 (55.71 %)

17 (65.38 %)

9 (34.62 %)

Table 2: Impact of cell breathing on soft handover and average number

of RAB attempts on SC 129

V. CONCLUSION This study shows that the cell breathing in cellular

network is affected by drop calls, number of soft handover and cell coverage. The result obtained shows that when the number of MS increases, the cell coverage area will generally decrease.

The COST231 HATA propagation model can be used to predict correctly the cell coverage of cellular network for urban areas. The RSSI data from the drive test and the calculated RSCP data indicate minimum discrepancy. Hence, it can be concluded that the RSSI or RSCP for the selected urban areas are quite similarly and fulfill the RSCP requirement.

REFERENCES

[1] Harri Holma and Antti Toskala from Nokia Network Finland; “WCDMA for UMTS, Radio Access for Third Generation Mobile Communications”; Third Edition, John & Wiley Sons Ltd., 2004.

[2] Kay Leong Thng1,2, Boon Sain Yeo1 and Yong Huat Chew1;”Performance Study on the Effects of Cell-Breathing in WCDMA”, 1Institute for Infocoom Research, 21 Heng Mui Keng Terrace, Singapore 119613, 2Electrical and ComputerEngineering Dept., National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260.

[3] Olivia Brickley, Susan Rea, Dirk Pesch,; “Enhancing QoS in IEEE802.11e WLANs Using Cell Breathing”, Centre for Adaptive Wireless Systems, Department of Electronic Engineering, Cork Institute of Technology, Cork, Ireland.

[4] Kimmo Valkealahti, Albert Höglund, Jyrki Parkkinen, and Ari Hämäläinen; “WCDMA Common Pilot Power Control for Load and Coverage Balancing”, Nokia Research Center, FIN-00045 Nokia Group.

[5] J Jyoti Laxmi Mishra, Keshav. P. Dahal and M. A. Hossain; “Call Admission Control using Cell Breathing Concept for Wideband CDMA”, School of Informatics, University of Bradford BD7 1DP, UK.

[6] Xiaoming Bo and Zujue Chen; ”On Call Admission and Performance Evaaluation for Multiservice CDMA Networks”, Mobile Computing and Communicatin Review, Vol.8, No.1.

[7] Blaustein Matha, Radio Propagation in Cellular Networks: COST 231 Hata Model, Artech House, Boston-London, Interna Report,1999, http://www.ee.bilkent.edu.tr/~microwave/ programs/wireless/prop/CostHata.htm.

290290