NIETJOURNALoFENGINEERING&TECHNOLOGY

Winter 2011coNcoL.{)I

(j)

NNNZ(J)(J)PLANNING AND OPTIMIZATION

OF UMTS NETWORKRADIO LINK WITH

LIMITED INTERFERENCESatyendra Sharma'

Abstract

Coverage and Capacity are significant issues in RFplanning process for Universal MobileTelecommunication System (UMTS). Radio link isinterference limited and interference limited network iscapacity limited. On the other hand, noise limitednetwork are considered to be coverage limited. Animportant step towards mitigating this problem is toreduce interference by controlling channel power andre-configuring existing Base station(BS),so as toenhance Capacity and Coverage of a network. In thispaper we propose an effective method for optimizingchannel power along with maximizing the number ofserved users and minimizing the number of cell sites.

'Noida Institute of Engineering& Technology Greater Noida.Satyendra Jupa@ indiatimes. com

'Brahm)it Singh. Gautam BuddhaUniversity, Greater Noida.brahmj it.s@gmail.com.

Dr Brahmjit Singh'

Introduction

In cellular networks planning and optimization,parameters such as communication channeltransmit power, level of interference, traffic typeand distribution, soft capacity should beconsidered on priority. Optimizing the channeltransmit power along with maximizing thenumber of several users and minimizing thenumberofcell sites [2].

The coverage problems in a network are definedby considering the signal level in each test pointfrom all Base stations (BS) and our requirementis at least one level above a fixed threshold level.Network may be planned for signal quality ratherthan signal strength [1]. Selection of good Basestation sites and channels will result inacceptable coverage performance at Basestations both in coverage area and in signalquality. Heuristic search method is to solve themaximal coverage location problem oftransmitters. Maximizing coverage andminimizing transmitters may be archived by aGenetic Algorithm [10].

Capacity estimation is an important issue inperformance analysis and call admissioncontrol and it is closely related to power control,strength based power control systems andsignal to interference ratio (SIR) based powercontrol system. The relationship betweenreceived power at a Base station (BS) and total

other cell interference from users who are powercontrolled by other Base stations is derived formaximum capacity [11].

An economic design of network, a trade offbetween the cost of coverage and the benefitresulting from covering area is desired. This typeof coverage model assumes a limited budgetand includes this as a constant on the number .offacilities to be located. Thus, the optimizationtries to place a fixed number of Base stations, sothat the proportion of demand nodes covered bythe cell within the permitted range is maximized.This may be explored using heuristicalgorithm [12].

3G and 4G systems in addition to frequencyplanning involve various other parameters formaking trade-off between coverage/capacityand service quality. These includecommunication power, interference level, trafficand soft capacity. An interference limited networkis usually considered to be capacity limited. Onthe other hand, noise limited networksconsidered to be coverage limited [1, 13, 14].

Capacity and coverage calculation:If there are N users in a cell and the signal isdenoted by 8 then the interference can becalculated as I = (N - 1) 8 + 1"], where I"] is the background thermal noise. Hence the 8NR is given by

SNR = S 1 (1)(N -1)S +Y\ (N -1) +Y\ / S

8uppose the detector for each user can operateagainst the noise at an energy per bit-to noisepower density level is given by EJNo, whosenumerator is obtained by dividing the desiredsignal power by the information bit rate, R, anddividing the noise (or interference) by the totalbandwidth, W. This results in

Eb/N = W/Ro (N-l)+y\/S (2)

Where W/R is generally referred to as the"Processing gain" and the background noisedetermines the cell radius for a given transmitterpower. The above equation can be written as thecapacity in terms of number of users,

N=I+ W/R _~s, /No S

(3)

Winter 2011

That means, the number of users is reduced bythe inverse of the per user signal-to-noise ratio(8NR) in the total system spread bandwidth, W.However, ifthe user is not speaking during part ofthe conversation, the output of the coder islowered to prevent the power from beingtransmitted unnecessarily. This reduces theaverage signal power of all users andconsequently the interference received by eachuser. The capacity is then increased proportionalto this overall rate reduction. In order to attain anincrease in capacity, the interference due toother users should be reduced. This can be doneusing antenna sectorization and monitoring ofvoice activity [13 , 14]. Thus with sectorizationand voice activity monitoring factor a, theaverage EJNo

W/R(4)

Where Ns, the number of users per sector andthe interference to be that received by onesector's antenna. Now, consider interferencefrom the jth user in neighboring cell k, then theratio of other cell interference to the receivedsignal strength at home base station is 1/8 [10].Then equation (4) becomes

E bl W / R (5)/No (Ns-l)a+(J/S)a+ll/S

Where (I/S)=(rm/rJm'(~o-~m) ,rm is the randomdistance to the corresponding home cell basestation, ro is the distance to the neighboring cell,represent the shadowing parameter and m is thepath loss exponent [14]. We can also write theabove equation as

S = ---'ll _

WhEbl/No

(6)

-(Ns -l)a -(J/S)a

Finally, the received power at the base stationfrom the user1 , is given by

8 = 8,-Lp-U (7)

Where 81 is the transmission power of the user,

NIETJOURNALoFENGINEERING&TECHNOLOGY

Lp is propagation path loss at distance d fromthe mobile station to base station and U is theshadow fading losses. From equations (6) and(7) we can build a relation among the receivedpower, number of users and the coverage area.

Results and discussions:A state-of-the-art planning is formulated byproposed solution method. 8ystem modelincorporated here is considering a networkwith an area of m x n (km) which needs to beprovided with coverage. Let us consider anArea, where the number of required sites (St) ismade and the operators have a set of candidatesites; 8= {81, 82, ... , 8t}. The cells need toprovide coverage for a total number of Ut usersin the area, according to the traffic modelsuggested in the dimensioning phase. Each celloperates at a power of P = {P1, P2, ... , Pmax} ..We aim to minimize the cost associated with cellsites, by using the minimum number of cell-sitesand maximize the revenue by increasing thenumber of served users. Hence, the objectivefunction can be defined as:

max F(fl'f2,j3) = WJI + W2f2 + r.r. (1)

8ubjectto:

WI +W2 +W3 =1Where

(2)

(" = Us f = p, - P,t f = St - So (3)JI U' 2 p' 3 S

t t t

Here, Ut is the total number of users, Us is thenumber of served users, Pt represents totalavailable power from all sites, Pu is total usedpower by all sites, 8t and 80 show the totalnumber of sites and total number of "on" sites- incontrast to an "off" site, or a site that does notneed to be deployed. W, to W3 represent theweight associated with the parameter we wouldlike to optimize. While Ut, Pt and 8t are fixednumbers in the system, we intend to find thesub-optimal values for Us>P, and 80 in a way thatthe objective function F(!; ,12' h) is maximized.

Let us consider, for a rural area where each cellsite is divided into three sectors, total areacovered is 10X 10 km. Antenna height is 35 mand user height is 1.5 m and others data are:

Winter 2011

U, = 1000, Us= 764, P, = 3 db, Pu= 2.8 db, 8,=780=4

The simulated results are shown as:

LetW,= Oto 1;

W2= Oto (1-W1);

W3= 1-(W1 +W2);

Then Z = WJl + W2f2 + W3f3Z = WI (1.308)+ ff2 (0.066) + W3 (1- W I - W2)

References1 EdoardoAmaaldi,AntonioCapone,F e d e r i c 0

Malucelli "Radio Planning and coverageoptimization of 3G cellular network ;wirelessNet(2008) 14:435-447.

2 Bahareh Jalli, Mehrdad Dianati Application ofTaboo Search and GeneticAlgorithm in Planningand Optimization of UMTS Radio Net w 0 r k s .IWCMC"1 O. June28-july201 O.

3 Zaher Dawy, Ahmad Husseini, Elias Ya a c 0 u b ,Lina AI-Kanj, "A Wireless CommunicationsLaboratory on Cellular Network Planning". IEEETransactions on Education, vel, 53 No.4 Nov2010.

4 FrankA.Cassara,"Wireless Com m u n i cati 0 n sLaboratory". IEEE Transactions on Education,voI49,No.1 Feb 2006.

5 Teruya Fujii, Yoshichika OHTA & Hideki, OMOTE"Empirical time-spatial propagation model at BSin outdoor NLOS environment" Wireless systemresearch center, Japan.

7 Theodore S. Rappaport, "Wirelesscommunications, Principle & Practice" 2ndEdition (PHI).

8 Ajay R. Misra, "Fundamentals of Cellular NetworkPlanning and Optimization" Willey Publication.

9 S.J. Lee & D.K. Sung" Capacity Evaluation forDS.CDMA Systems with Multiclass ON/OFFTraffic" IEEE Commun. Lett.,pp. 153-155June1998.

10 Chae Y Lee and Hyon G .Kang "Cell planning withcapacity expansion in mobile communication: ATaboo search approach"

11 Duk Kyung Kim "Capacity estimation for amulticode CDMA system with SIR-Based powercontrol" 0018-9545@2001 ieee.

12 K.Tutschku "Demand Based radio networkplanning of cellular mobile communicationsystems" .Report no. 177 July 1997

13 Liang guo , Jie Zhang and Carsten Maple"Coverage&Capacity Calculation for 3G Mob i I eNetwork Planning".University of Luton U.K.

14 T.Rama Rao and M.V.S.N Prasad" C 0 v era g e &Capacity Studies for a CDMA CELL in DifferentRadio Propagation Environments".UniversidadCarlos III DE Madrid, SPAIN ,Radio& AtmosphericSciences Division, N.P.L, New Delhi India.

Winter 2011

15 Duk Kyung Kim & Dan Keun Sung " CapacityEstimation for a Multicode CDMA Sysytem withSIR-Based Power Control". IEEE Transaction onVehicular Technology, Vol-50,No.3. May 2001.

16 Yong Huat Chew, Boon Sain Yeo & Daniel ChienMing Kuan "Single & Multiple ParametersSensitivity Study of Location Management AreaPartitioning for GSM Networks". IEEECommunication Society subject matter expertsfor publication in the ICC 2007 proceedings.

17 S.Z. Ali "Location Management for CellularMobile Radio Networks". 52nd IEEE VTC,pp11 06-111 0, May 2002

18 S.J. Lee & D.K. Sung" Capacity Evaluation forDS.CDMA Systems with Multiclass 0 N / 0 F FTraffic" IEEE Commun. Lett.,pp. 153-155June1998.

ODD