Design Of A Cdma Based Mobile Communications System Engineering Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The demand for mobile communication services has been on a continuous rise all over the world climbing from 500 million subscribers a decade ago to 4.6 billion in 2009[1] This rise accompanied with competition among network providers and users demand for higher data rate at lower cost lead to continuous evolution in telecommunication technologies. The drive to maximize the available bandwidth has seen the technologies evolve from first generation (1G) and second generation (2G) networks that is totally circuit switched to 2.5G and 3G networks that are both circuit and packet switched and moving to 4G that is totally packet switched. In parallel with this is also the shift from FDMA and TDMA techniques to CDMA techniques which provides spectrum efficiency. In using codes to define user channels, this multiplexing technique provides solution for the ever increasing number of subscribers and holds much prospect for the future of telecommunication which is the 4G. The CDMA technique as employed in 2G and 3G is used in this project for a design of 10kbps voice only network and multiple services network of 10kbps voice and 150kbps non-real time data and are hereby discussed.


The CDMA network designs proposed in this project exploits the advantages of performance enhancement techniques such as antenna sectoring, soft handoff and voice silence detection. The flexibility in handling data rates and soft blocking capabilities makes the designs robust and efficient.

Option one is a 2G network which uses a nominal bandwidth of 2MHz and chip rate of 1.664Mcps to provide quality voice only network with blocking probability of 1%. This requires about 276 cell sites to cover Orange Island with 204 base stations located in the City centre an area of 100km2. The suburban and sparsely populated areas assumed to be 400km2 and 1000km2 required 64 and 8 base stations respectively. OQPSK is the modulation technique uplink where the interference is worse to ensure constant amplitude with limited phase variation while QPSK is used downlink, maintaining the same transmission rate and enhancing spectral efficiency. Cyclic coding with 24 bits CRC is used for error detection to ensure maximum detection in the case of worst fading (burst errors) and a convolution encoder of rate 1/3 k =7 is used for error correction.

Option two requires a nominal bandwidth of 4.6MHz with chip rate of 3.744Mcps in a single carrier three sector cell system and will effectively provide voice and data services to about 145,000 users in all the regions of the Island. The number of base station required is 111, 45 and 6 for the city centre, suburban and sparsely populated areas respectively. QPSK with coherence demodulation is the modulation techniques for uplink and downlink, maintaining the same transmission rate on both links and providing spectral efficiency with enhanced signal detection.

A rate 1/3 turbo code and cyclic code of 12 bits CRC is employed to provide high error detection and correction. The robustness and efficiency of turbo code is about the highest. [2]


The practical application of spread spectrum technology can be traced back to military operations since World War II. [3] Until Dixon's publications in 1976, most literature on spread spectrum technology were not detailed as application information were classified. [4] The authorisation for civil use of the technology according to [4] was obtained in 1985 with Qualcomm as one of the pioneering company. The 2G standard IS-95 was published in 1993 and revised in 1995 with a second revision standardizing IS-95B referred to as 2.5G. [5] The 3GPP and 3GPP2 established in 1998 defined the UMTS for WCDMA (3G network) in Release 99 and standardized CDMA2000 respectively while the LTE (3GPP long term evolution) and LTE advanced are mapping CDMA into the fourth generation networks. [6]


To design a CDMA based mobile communication system for a major city Voda in Orange Island covering an area of 1500km2 with a densely populated centre of about 100km2. The density of users in the city centre, the suburban and the sparsely populated areas are 1000/km2, 100/km2 and 5/km2 respectively. Available frequency bands are 1920 - 1980MHz for uplink and 2110 - 2170MHz for the downlink. The design is to cover two options: a voice only network with user data rate of 10kbps and a 10kbps voice with 150kbps non-realtime data network. Additional assumptions:

The suburban and sparsely populated areas are 400km2 and 1000km2 respectively.

The physical terrain of the areas is considered to be averagely plain with no additional path loss or network obstruction by high lands.


In 2001, GSM licensing was done in Nigeria, at a license price of $285 million [7] [8] (approximately N 39.9 billion @ $1 = N 140) for 40MHz (2x5MHz in the 900MHz band and 2x15MHz in the 1800MHz band [9]). The cost per Hertz = N 997.5. The cost of building a cell site is about $230,000 [7] excluding the cost of lease for the land which is about $21,500. Total site cost is about $251,500 (N 35.21 million). This means that the cost of 35.3kHz is equivalent to the cost of 1 cell site and 1MHz comparable to 28 cell sites.

The 150th round of bidding ended the UK licensing in 2000 [10] with average cost per Hertz being £161.8 as shown in table 1.

The average cost of building a new site ranges from £220,000 - £240,000 [12] or €266,000 [13].

Therefore the cost of 1.422kHz is equivalent to the cost of 1 cell site and 1MHz comparable to 703 cell sites!

Table 1: Cost of 3G bandwidth (Sources: [11] & [12])


The analysis above present the extremes of the cost of bandwidth relative to cell site. The task at hand is to deliver a quality network services to the people of Voda, ensuring adequate coverage using the best optimization method and at a considerably low cost. Bandwidths are expensive but in trying to save in bandwidth, more base stations are required which is a relatively low cost. So trade-offs has to be made in the most economical way to ensure conservative use of the bandwidth. Based on these considerations, the following decisions are made:

The chip rates and bandwidths are considerably chosen to reduce cost in hardware, cell site acquisition and bandwidth.

The base stations can be shared with other providers to recover some of the cost.

Extra cost of bandwidth in option 2 can be recovered from data users.

The coding schemes are maximized to enhance system capacity and service quality, and to avoid long delay due to ARQ for data services.

Blocking probability is set at 1% (considered highly satisfactory from my interview and interaction with users).


Figure 1 (matlab code located in appendix 1) shows that as user data rate increases, the number of users that can be accommodated at a particular bandwidth decreases.

Fig.1: Number of users as a function of bandwidth Pole Capacity

To enable us determine the effective bandwidth for the system, we start from the required user data rate and calculate what the data rate at the base station should be. Figure 2 is a block diagram of uplink network for voice only service. A coded voice signal to be transmitted in a 10ms frame structure at 10kbps contains 100 bits. The frame is coded for error detection adding 24 check bits and forwarded for error correction coding. The encoder is a rate 1/3 convolution encoder which is described below in section 2.1. The tail bits appended to the data from the encoder is 6 resulting to total user data rate of 13kbps. The output encoded sequence of 390 bits (130 x 3) is interleaved to reduce the effect of burst errors. The interleaved output of 39kbps is used to generate Walsh code (every 6 bits of the 390 is used to generate a 64 bit sequence); the resulting signal is 4160 bits at a data rate of 416kbps. The output is fed into a data burst randomizer which regulates the transmitter power (turns the transmitter off or to a power save mode when the total user data rate is less than 13kbps). With spreading factor of 128 for a total user data rate of 13kbps, the resulting chip rate is 1.664Mcps.

BT= Baud rate (1+r)…….. (1)

For a roll-off factor of 0.20 using a root raised cosine, BT = 1.9968MHz. The nominal bandwidth is set to 2MHz including guard bands.

Fig. 2 : Uplink network (Mobile to Base station) Adapted from [14]

The downlink transmission shown in figure 3 is almost the same as in the uplink except that a rate ½ encoder is used since there are less interference downlink and 4 power control bits (400bps, using a decimator of 65 on the 26kbps data rate) are multiplexed with the data to be transmitted which is at the rate of 26kbps. The 64 Walsh codes generated at the chip rate of 1.664Mcps define the downlink channels and are used to spread the 26kbps signal resulting to a spreading gain of 64. A total of 55 traffic channels are available to the base station, 1 pilot channel, 1 synchronisation channel and 7 paging channels. Also the modulation technique in the downlink is different without a delay in the quadrature signal.

Fig. 3 : Downlink network (Base station to Mobile) Adapted from [14]


Cyclic coding with 24 check bits reduces the probability of undetected error to 10-24 [14] and a k = 7 rate 1/3 convolution coder provides error correction. The efficiency of the convolution encoder increases with decreasing rate and increasing constraint length.[14] With a large blocking factor (number of digits in a block), the efficiency of the coding tends to maximum. These choices ensure that the bit error rate is not more than 10-3 and any frame with uncorrected error is dropped for worst case fading.


The modulation technique for the uplink is OQPSK. CDMA is an interference limited system and this is worse on the uplink. OQPSK will provide enhanced signal detection at the base station since the signals are of constant amplitude and phase variation is less than. This reduces the bit error rate and increases the quality of the service. In the downlink where interference is less, modulation is QPSK. It is important to note that coherent detection is not required with these modulation techniques and the baseband signal for the uplink or downlink is not more than 1.1MHz ( half the transmission bandwidth for a BPSK) leading to enhanced spectral efficiency.


The link budget for the uplink and the downlink are located in appendix 2 and 3 respectively. The final path loss allowed in the uplink is 140.5dB and 157.5dB for the downlink. Since the allowed path loss for the uplink is less, the system is said to be uplink limited

The Okumura - Hata propagation model as used in [15] for a base station antenna height of 30m, mobile station antenna height of 1.5m relates the allowed path loss in dB (L) and the cell range in km (R) by:

L1 = 137.4 + 35.2 ……… (2)

If we provide a correction factor of 8dB for the suburban region [15] and 27dB for the sparsely populated area [16], for the suburban equation (1) becomes

L2 = 129.4 + 35.2 ……… (3)

and for the sparsely populated area

L3 = 110.4 + 35.2 ……… (4)

for L = 140.5dB, the range R1, R2 and R3 for the densely populated, suburban and sparsely populated areas respectively corresponds to 1.225km, 2.067km and 7.163km. These values give bounds in terms of path loss (coverage). However, we need to check limitation due to capacity.

The number of users per cell (single carrier one sector) in the uplink is given by:

NUL = + 1………. (5)

where W= chip rate; R = user bit rate; = bit energy to Interference spectral density; = Interference from other cells and = activity factor of the user.

For voice service with: W=1.664Mcps, R = 10kbps, = 7dB (5) for a vehicular user at 120km/h, β=50% (0.5) and =65% (0.65),

NUL ≈35 user channels.

In the downlink, the capacity of a single carrier one sector cell is given as:

NDL = + 1………. (6)

where 𝜶 (orthogonality factor due to imperfect power control) = 0.4.

NDL ≈ 48 user channels.

The figures calculated above are referred to as pole capacity corresponding to 100% loading of the cell. This however is not practical and 6dB noise rise is provided for as interference margin in the link budget corresponds to 75% load factor (ηL). [15]

Actual capacity in the uplink is given by:

NUL x ηL ......... (7)

From equation (7), NUL ≈ 26 user channels & NDL ≈ 38 user channels.

From the foregoing, the system is totally uplink limited in capacity and coverage. Therefore the design will focus on the uplink.

Fig. 4 : Cell sectoring

For a single carrier three sector cell as in figure 4, the uplink capacity is 26 x 3 = 79 user channels.


For the soft capacity of the system, the Erlang table will not give an optimum result.[15]This is because if interference is less in the surrounding cells, the referent cell's capacity increases. Using the procedure summarised in [15] and located in appendix 1, the soft capacity is calculated as in table 2 for blocking probability of 1%.

Table 2 : Erlang calculation for soft capacity


The average minute of use for cell phone call is 475 minutes per month for a holding time of 3.25 minutes, meaning 146 calls per month. [16]

On average 90% of the traffic is during the working days (21 days) in a month. [17] This results to average of 6 calls /day/subscriber. Averagely, 2 of the calls are made during the busy hour. [18]

The traffic erlang for a single user = ………… (8)

With the figures specified above, equation (8) gives 0.1083 erlang of traffic

The user density in the city centre = 1000users/km2.

Traffic density = 1000 x 0.1083 = 108.33erlang/km2.

Area of a cell = Soft capacity / Traffic density

= 67.69erlang / 108.33erlang/km2 = 0.625km2

Using the Hexagonal shape model of cell dimension, the range of the cell = = 0.490km (well within the range for maximum path loss 1.225km).

Number of cells in the city centre = 100km2 / 0.490km2 = 204 cell sites

The user density in the suburban area = 100users/km2.

Traffic density = 100 x 0.1083 = 10.83erlang/km2.

Area of a cell = Soft capacity / Traffic density

= 67.69erlang / 10.83erlang/km2 = 6.25km2

The range of the cell = 1.55km (still within the range for maximum path loss 2.067km).

Number of cells in the suburban area = 400km2 / 6.25km2 = 64 cell sites

The user density in the sparsely populated area = 5users/km2.

Traffic density = 5 x 0.1083 = 0.5415erlang/km2.

Area of a cell = Soft capacity / Traffic density

= 67.69erlang / 0.5415erlang/km2 = 125.0km2

The range of the cell = 6.934km (less than link budget value of range for maximum path loss 7.163km).

Number of cells in the suburban area = 1000km2 / 125km2 = 8 cell sites

Total of 276 cell sites required to cover Voda in Orange Island.


Fig. 6 : Uplink network for data traffic (Mobile to Base station)

Figure 6 is the uplink network diagram of the design for a mobile communication system capable of providing 10kbps voice service and 150kbps non-realtime data service. Starting with a user who is uploading data at 150kbps rate, a 10ms frame containing 1500 bits undergoes cyclic coding with additional 12 CRC bits appended to it. The 1512 bits are segmented into 4 blocks of 378bits so as to match the latency of the interleaver (390bits) in the turbo encoder. Each block is channel coded using the rate 1/3 turbo encoder described in section 3.1. The output of the encoder is 4 blocks 1170 bits made up of the original 378 bits plus 12 tail bits multiplied by 3. The total output bit for radio frame equalization is 4680 bits. The bits are interleaved twice (depending on the transmission time interval) and mapped onto the physical layer. The bits are spread with channelization code separating the data and the control channels and the spreading factor is 8 for data. The output chip rate is 3.744Mcps and corresponds to the nominal bandwidth of about 4.6MHz using equation (1). The I and Q codes are multiplexed to avoid interference in the system during discontinuous transmission before data is spread with complex scrambling code. The scrambled signal then modulates the carrier for transmission. The downlink operation is about the same as the uplink with QPSK used in both the downlink and uplink.

The network diagram for the uplink voice traffic is shown in figure 7. The network is circuit switched and segmentation and equalization processes are not involved. Also skipped is the first interleaving to avoid delays that will degrade the quality of the voice service.

Fig. 7 : Uplink network for voice traffic (Mobile to Base station)


A 12 bit CRC code is used for error detection to reduce the proportion of undetected error in case of severe fading to 2-12 [14] considering the amount of data involved.

The turbo coding is one of the most efficient in error correction with BER lower than 10-5 and Eb/N0 of 0.7dB. [2] Figure 8 shows a block diagram of the turbo code. The parallel connected encoders are convolution encoders of K =7 resulting to total tail bit of 12 bits per code block. The coding is systematic and recursive while decoding is by continuous iteration. Table 3 shows the detailed output of the encoding scheme. The low latency of the interleaver 390 bits allows a maximum delay of 39ms for the 10kbps voice signal, making the encoder suitable for both data and voice.

Table 3 : Rate 1/3 Turbo code (Max. size of code Block 390 (i.e. 378 + 2 X 6 tail bits))




Data Data

Code Sequence

Fig. 8 : Rate 1/3 Turbo Encoder (Parallel Concatenated Convolutional encoder). [2]


QPSK with coherent detection is preferred as the modulation/demodulation technique for this design to maximize spectral efficiency while maintaining uniform spreading factor (96 & 8) and data rate (10kbps & 150kbps) in the uplink and downlink directions. The bit error rate performance of this technique is high because the signal does not suffer from the degradation experienced when differential detection is used. [18]


The link budget calculation for option 2 is located appendix 4 and 5 and coverage limitation due to maximum path loss is uplink bound.

At the allowed propagation losses are 140.5dB, from equations (2) - (4), the range of the cells corresponds to 1.07km, 2.36km and 9.31km for the city centre, the suburban area and the sparsely populated area respectively. These values give bounds in terms of path loss (coverage), but we need to check limitation due to capacity to determine the optimum size of the cells.

For voice only service with: W=3.744Mcps, R = 10kbps, = 7dB (5), β=50% (0.5) and =65% (0.65 i.e. 50% plus overhead during DTX ), using equation (5) the number of users per cell (single carrier one sector) in the uplink is :

NUL ≈77 user channels.

For data only service with: W=3.744Mcps, R = 150kbps, = 1dB (1.259), β=50% (0.5) and =100% (1.0),

NUL ≈14 user channels.

In the downlink, using equation (6), the capacity of a single carrier one sector cell for voice only service with: W=3.744Mcps, R = 10kbps, = 7dB (5), β=50% (0.5), =65% (0.65) & 𝜶 = 0.4,

NDL ≈ 105 user channels.

For data only service with: W=3.744Mcps, R = 150kbps, = 1dB (1.259), β=50% (0.5), =100% (1.0) & 𝜶 = 0.4,

NDL ≈ 19 user channels.

The calculations show that in both capacity and coverage, we are uplink limited. The design will focus on the uplink.

The pole capacity calculated above is not practical as it assumes that the mobile station has infinite transmission power and interference at node B receiver goes to infinity. [19] The 6dB noise rise corresponds to 75% load factor (ηL). [15]

Actual capacity in the uplink using equation (7) corresponds to approximately 58 voice only channels and 11 data channels.

The throughput in a cell is given by: = R x N (1- BLER) ........... (9)

where BLER = block error probability. For BLER of 1% and 10% for voice and data respectively, = 574.2kbps & 1485kbps in the uplink.

For a single carrier 3 sector cell, the number of channels scale to 174 voice channels and 33 data channels, while the throughput scale to 1.723Mbps and 4.455Mbps in the uplink.

3.5 MULTIPLE SERVICES (10kbps voice and 150kbps non-real time data)

So far in the design, we have been considering the network in terms of single services (voice or data). When mixed services are provided by a network, the WCDMA cell capacity cannot be referred to as a single number but a capacity region. [19] Figure 9 shows the capacity regions for pole capacity and 75% loading of the cell (matlab code in appendix 1).Every point in the bound represents a particular capacity mix for both services. The extreme cases of the bound at 75% loading is 174 voice user channels and no data user channel and 33 data user channels and no voice user channel.

To progress in the design, we need to define a traffic mix to enable capacity allocation and hence dimensioning of the cell. Priority is given to voice services which is more on demand. The point A in figure 8 defines a capacity resource schedule of 63% for voice service and 37% for data service which corresponds to approximately 109 voice user channels and 12 data user channels (throughput of 1,648,350bps out of the 4.455Mbps).


Fig. 9 : Uplink Capacity Region for Voice and Packet switched data Mixed Traffic


The soft capacity is shown below in table 3 using the procedure in appendix 1 for blocking probability of 1%. Note that there is no soft capacity for the packet switched data as shown in table 4 and figure 10.

Table 4 : Erlang calculation for soft capacity

Fig.10 : Increase in Erlang for voice services due to Interference sharing


The traffic intensity (A) as calculated for soft capacity is 96.353 Erlang.

The average number of call attempts per unit time (λ) is given by: where 1/µ is the average call length. Using the same holding time and busy hour call rate as in section 2.4, λ = 29.647 calls/minute.

The number of subscribers that can be accommodated in a cell becomes 29.647 x 30 = 889 users.

The medium rate of data usage is at 10Mbits per day. [12] Also the busy hour trend is the same for the same class of users (33% of usage in a day during busy hour as estimated for voice services). The average usage is about 3.3Mbits per hour and 916 bits per second.

With the throughput of 1,648,350bps, the number of users that can be accommodated in a cell is calculated as 1,648,350/916 = 1799 users. If we provide for burstiness factor [20] of 2, this will be 899 users

Proceeding with the hexagonal shape model. The area of a hexagon is given by 2.6R2. If 899 users may be accommodated in a cell, this means 346/R2 users per unit area.

For the densely populated area, 346/R2 = 1000 users/km2 giving R = 0.588km (well within the maximum path loss calculated for this area as 1.07km).

Area of cell = 0.899km2

Number of cell sites = 100km2/0.899km2 ≈ 111 cell sites

For the suburban area, 346/R2 = 100 users/km2 giving R = 1.86km (still within the maximum path loss for this area calculated as 2.36km).

Area of cell = 9.00km2

Number of cell sites = 400km2/9.00km2 ≈ 45 cell sites

For the sparsely populated area, 346/R2 = 5 users/km2 giving R = 8.319km (still within the maximum path loss for this area calculated for this area as 9.31km).

Area of cell = 180.00km2

Number of cell sites = 1000km2/180.00km2 ≈ 6 cell sites

Total of 121 cell sites required to cover Voda in Orange Island.


Expansion in the city centre will be covered through cell layering. Microcells and picocells will be installed where necessary to provide network for less mobile users along the streets and in hot spots. The suburban and the sparsely populated areas can accommodate more macrocells if population of users increase. However, a major growth in the number of users may require launching a second carrier.


The expectation is high for the future with recent technological advancements. A total packet switched network using internet protocols is to be deployed for voice, video and data networks maximizing the use of the available spectrum while availing users with very high data rates at low cost. With the target data rate already achieved in WiMax (worldwide interoperability for microwave access) and the progress made in WOBAN (Hybrid Wireless-Optical Broadband Access Network), it will be virtually possible to provide any data rate on demand for conceivable user services.


In this project, two network designs for a voice only service and voice and data services using code division multiplexing access was presented for a coverage area of 1500km2 with about 145,000 subscribers. Cost analysis was performed to determine the relative cost of bandwidth to base station site to enable good decision and trade-offs. Provisions were made for the effects of fading, busy hour traffic and bursty data to ensure that a robust network that will deliver quality service to the user at a low cost is provided. Link budget calculations were done to ensure adequate coverage of the cell edges and allowance made for expansion. The design was concluded with a brief discussion of expectation on the future of mobile communication.


The Economist, Dec 30th 2009 Mobile-phone culture: The Apparatgeist calls. Available at: Accessed : 3/2/2010

Burr, A. (2001) Turbo-codes: the ultimate error control codes? IEE Electronics & Communication Engineering Journal, 13, pp 155-165. Accessed : 14/04/2010 Accessed : 14/04/2010 Accessed : 14/04/2010 Accessed : 14/04/2010

David-Etim, R. (2007) Financing Telecoms: The Past, The Present and The future - A presentation at the GSM strategic conference. Available at :

Lee, D. (2003) Lessons from the Nigerian GSM auction. Science direct - Telecommunication Policy 27 (pp 407-416)

ITREALMS ONLINE (January 14, 2007) Mubadala gets unified license in Nigeria. Available at : Accessed : 22/4/2010

Daily summaries Spectrum auctions. Available at : static/archive/spectrumauctions/auction/auction_index.htm. Accessed : 22/4/2010

3G Spectrum. Available at : Accessed : 22/04/2010

Ofcom : Answers to stakeholder questions relating to cost modelling - Relating to the consultation on application of spectrum liberalisation and trading to the mobile sector(pp 3 & 10). Available at: liberalisation/app/supplement.pdf . Accessed : 22/4/2010

Johansson, K., Furuskar, A., Karlsson, P. & Zander, J. (2004) Relation between base station characteristics and cost structure in cellular systems. IEEE International Symposium, 4, pp 2627 - 2631

Schwartz, M. (2005) Mobile wireless communications (pp 174- 230). United States: Cambridge University Press.

Holma, H. & Toskala, Antti. (2004) WCDMA for UMTS, Radio Access for Third Generation Mobile Communications (pp 185-217). England: John Wiley & Sons, Ltd.

Matt Owings 2009, Biomechanics and Anatomical Kinesiology KINE3334-02 Course syllabus, The University of Texas at Tyler. Available at : Accessed : 3/2/2010

Smith, C. & Collins, D. (2002) 3G Wireless Networks (pp 393-476). United States: McGraw-Hill

Karim, M. R. & Sarraf, M. (2002) W-CDMA and cdma2000 for 3G Mobile Networks (pp 331-362). United States: McGraw-Hill

Qualcomm Engineering Services Group (2006) Air Interference Cell Capacity of WCDMA Systems 80-W0989-1 Rev A. PDF Available at : Accessed : 4/03/2010

EE428 Midterm-solution University of Bridgeport. Available at www.bridgeport. edu/sed/fcourses/ee428/Exams/midterm-solution.doc. Accessed : 7/04/2010


%Bandwidth relationship with number of users


N1= ((w/30832.21));

N1R = (N1*.75);

N2= ((w/283258.22));

N2R = (N2*.75);

figure (1)

plot (N1, w, 'kd-'); xlabel('Number of users'); ylabel ('Bandwidth');

title ('Bandwidth Pole Capacity'); grid on; zoom on;hold on;

plot (N2,w,'rs-');

legend('Number of 10kbps voice users','Number of 150kbps PS data users')


plot (N1R,w,'bs-'); xlabel('Number of users'); ylabel ('Bandwidth');

title ('Bandwidth 75% Loading Capacity'); grid on; zoom on;hold on;

plot (N2R,w,'rd-');

legend('Number of 10kbps voice users','Number of 150kbps PS data users')

%Capacity Region Plotting







figure (1)

plot (x1,y1,'-bd');xlabel('Number of 10kbps voice users');

ylabel('Number of 150kbps PS data users');

title ('Uplink Capacity Region');grid on; zoom on;hold on;

plot (x2,y2,'-rh'); legend('Pole Capacity','75 Percent Loading')

Procedure for estimating Soft capacity [15].

1.Calculate the number of channels per cell, N, in the equally loaded case, based on the

uplink load factor

2. Multiply that number of channels by 1 + to obtain the total channel pool in the soft

blocking case.

3. Calculate the maximum offered traffic from the Erlang B formula.

4. Divide the Erlang capacity by 1 + .

5. The trunking efficiency is defined as the hard blocked capacity divided by the number

of channels



Assumptions for mobile station

Voice Terminal

Max. Tx. Power: 23 dBm (200mW class 3 device)

Antenna gain: 0 dBi

Body loss : 3 dB

Assumptions for Base station

Amplifier Noise Figure : 4dB

Antenna gain : 18 dBi (3 sector base station)

Requirement : 7 dB

Cable loss : 2 dB

10kbps voice service (120km/h in-car user with soft handoff in the Capacity limited area (City centre))

Transmitter (Mobile station)

Max. Mobile Tx. Power [dBm] 23.0 a

Mobile Antenna gain [dBi] 0.0 b

Body loss [dB] 3.0 c

Equivalent Isotropic Radiated Power (EIRP) [dBm] 20.0 d = a +b - c

Receiver (Base station)

Thermal noise density [dBm/Hz] -174.0 e (KTBn where Bn =1Hz)

Base station Receiver Noise Figure [dB] 4.0 f

Receiver Noise density [dBm/Hz] -170.0 g = e + f

Receiver Noise power [dBm] -107.8 h = g +

Interference margin [dB] 6.0 i (75% loading maximum)

Total effective noise + Interference [dBm] -101.8 j = h +i

Processing gain [dB] 22.2 k =

Required [dB] 7.0 l (dependent on service)

Receiver sensitivity [dBm] -117.0 m = l - k + j

Base station antenna gain [dBi] 18.0 n

Cable loss in the base station [dB] 2.0 o

Fast fading margin [dB] 0.0 p

Maximum path loss [dB] 153.0 q = d - m + n - o - p

Log normal fading margin [dB] 7.5 r (93.4% coverage prob.)

Soft & Softer hand-off gain [dB] 3.0 s

In-car loss [dB] 8.0 t

Building Penetration loss [dB] 0.0 u

Allowed Propagation loss [dB] 140.5 v = q - r + s - t - u