PERFORMANCE EVALUATION OF IS-95A CDMA TRAFFIC CHANNELS USING SIMULINK

Muhammad Umair Siddiqui, Qazi Safiullah Shahrukh, Muhammad Usman Supervised by Associate Professor Dr. Zainab Zaidi Department of Electronic Engineering, NED University of Engineering & Technology, Karachi umairsiddiqui@engineer.com, le.reaver.de.soul@gmail.com

ABSTRACT The paper before you is an evaluation of the IS-95A CDMA forward link and reverse link performance. We have devised some numerical results based on the real time simulation of the forward link and the reverse link under constraints imposed by the regular AWGN and REYLEIGH faded channels along with the mobile users imposed limitation. This simulation can also be used to deduce results for practical CDMA system designing. We have come up with graphical relationship between the BER (Bit Error Rate) and SNR (signal to noise ratio). Key words: CDMA, Rate Set, Convolutional Encoder, Orthogonal Spreading Codes, Pseudonoise (PN), CRC (Cyclic Redundancy Check), Vocoder, Chip. 1. INTRODUCTION Code Division Multiple Access (CDMA), as defined in Interim Standard 95 (IS-95), describes a digital air interface standard for mobile equipment that enhanced the capacity of older analog methods with greatly improved transmission quality. cdmaOne is the brand name1 for the complete wireless telephone system that incorporates the IS-95 interface. CDMA systems were serving over 65 million subscribers worldwide by June, 2000. CDMA has proven itself as a successful wireless access technology in 2nd generation networks. Furthermore, the evolving third generation systems will rely on CDMA techniques for radio access. However, the structure of the physical layer of a cdmaOne network is significantly different than its GSM or IS136 counterparts. Figure 1 gives you a brief idea.

Figure 1 Basic overview of CDMA 2. FORWARD AND REVERSE TRAFFIC CHANNELS 2.1 Forward Traffic Channel For IS-95A CDMA The forward channel is the link between the base station and the mobile user. Data is sent on RATE SET 1 (9600 bps). The data or voice generated by the vocoder is passed in to the CRC generator which attaches its generator bits along with the data or voice streams. The baseband data plus the generator bits from CRC is convolutionally encoded for further strengthening error protection. Since rate set 1 is used therefore rate of encoder is . After convolutional encoding the symbol are repeated by the symbol repetition block only when data rate is below FULL RATE only to reduce interference power. After symbol repetition the data is interleaved to combat fading. Then the interleaved data is scrambled by a decimated long PN sequence which is generated by a ling code generator at 1.288 Mcps. After this the PCBs (power control bits) are multiplexed with the scrambled scheme at 19.2

Ksps and can be punctured into any one of the first 16 bits position of a PCG (which contains 24 bits). After this the data is orthogonally spread by the assigned Walsh function which is used here for channelization. The data stream is further spread by the assigned short PN sequence of transmitting sector. The output of the logical channel thus formed is then fed into a QPSK modulator. The gain of each channel is dynamically adjusted and indicates the amount of power transmitted for that channel. After this on the receiver end the signal after demodulation is passed into the rake receiver which due to diversity advantage turns multipath signals into one strong signal. After this in the descrambling section, the short PN sequence is descrambled first, then the Walsh code and in the end the long PN sequence. Then follows the de-inter-leaver and the derepeater which de-repeats only when the data rate is below FULL rate. For decoding the VITERBI algorithm is used and then the frame quality is checked for evaluating power handling instructions.[1,2,3]

Here channelization is done by long PN sequence instead of Walsh codes. Here Walsh codes are used for orthogonal modulation which is sending a Walsh code instead of a group of six symbols or bits which makes detection at the receiving end easier. As the reverse link uses OQPSK modulation the data is further scrambled into I and Q paths by short PN sequences which is running at 1.2288 Mcps. This is done for increasing mobiles battery life. The receiving end operates in the same manner as the forward link.[1,2,3]

Figure 2.2 Reverse Link 3. SIMULATION TOOLS Simulation is performed with the help of MATLAB and SIMULINK. CDMA Reference Blockset, Communication Blockset and Signal Processing Blockset are used to build the Forward and Reverse Traffic Channel models. 3.1 What Is SIMULINK? Simulink is a software package for modeling, simulating, and analyzing dynamic systems. It supports linear and nonlinear systems, modeled in continuous time, sampled time, or a hybrid of the two. Systems can also be multirate, i.e., have different parts that are sampled or updated at different rates. For modeling, Simulink provides a graphical user interface (GUI) for building models as block diagrams, using click-and-drag mouse operations. Simulink includes a comprehensive block library of sinks, sources, linear and nonlinear components, and connectors. You can also customize and create your own blocks. Models are hierarchical, so you can build models using both top-down and bottom-up approaches. After you define a model, you can simulate

Figure 2.1 Forward Link 2.2 Reverse Traffic Channel For IS-95A CDMA The reverse link indicates the link between the mobile user and the base station usually this is the limiting link. The data or voice from the mobile user has to pass through the same stages in the reverse link as in the forward link. The difference lies in the modulation scheme and the spreading. Here the rate of convolutional encoder is lowered from to 1/3 which makes the error protection more robust on the reverse link

it, using a choice of integration methods, either from the Simulink menus or by entering commands in the MATLAB Command Window. 3.2 What Is the CDMA Reference Blockset? The CDMA Reference Blockset is a collection of Simulink blocks designed to help you develop and simulate CDMA wireless communication systems, based on the current North American IS-95A CDMA (code division multiple access) standard. With the CDMA Reference Blockset, you can construct block diagram models of wireless systems quickly and easily using click-and-drag mouse operations. You can then run simulations on those models and change parameters as needed. The blocks in the CDMA Reference Blockset encompass the complete functionality required by the IS-95A standard. 3.3 What Is the Communications Blockset? The Communications Blockset extends Simulink with a comprehensive library of blocks to design and simulate the physical layer of communication systems and components. The blockset helps you design communications systems and their semiconductor components, such as commercial or defense wireless and wireline systems. The key features of the blockset are: [4] Blocks for designing and simulating the physical layer of communications systems, including modulation, source and channel encoding, channels, and equalization. Graphical user interface for dynamically tuning models and visualizing the results. Hierarchical, block-based models for visually conveying complex designs. 3.4 What Is the Signal Processing Blockset? The Signal Processing Blockset is a tool for digital signal processing algorithm simulation and code generation. It is made up of blocks contained within block libraries. You can interconnect these blocks to create sophisticated models capable of operations such as speech and audio processing, wireless digital communications, radar/sonar, and medical electronics. [4] 4. SIMULATION MODEL 4.1 Forward Traffic Channel Model The Simulink model of IS-95A Forward Traffic Channel is given in figure 4.1 (see last page).

The IS-95A Forward Traffic Channel Model shows the channel coding, modulation and spreading of the data symbols at the base station and the corresponding decoding as well as despreading and noncoherent demodulation at the mobile station receiver. The transmitter encodes the data, and then symbols are modulated by a Walsh modulator and spread by a PN sequence. The channel model adds noise to simulate errors in transmission. The receiver side retrieves the information bits by performing the decoding and demodulation. The bit error rate for the data is displayed in the simulation. The IS-95A Forward Traffic Channel Model uses these library blocks from the CDMA Reference Blockset IS-95A CRC Generator IS-95A Frame Quality Detector IS-95A Fwd Ch Convolutional Encoder IS-95A Fwd Ch Interleaver/Deinterleaver IS-95A Fwd Ch Repeater/Derepeater IS-95A Fwd Ch Viterbi Decoder IS-95A Fwd Ch Base Station Transmitter Interface IS-95A Fwd Ch Detector IS-95A Long Code Generator IS-95A Short Code Generator IS-95A Walsh Code Generator IS-95A Fwd Ch Scrambler Two library blocks from the Communication Blockset AWGN Channel Rayleigh Multipath Fading Channel and two from the Signal Processing Blockset Transmit Filter Receive Filter The base station transmitter section performs the CRC (cyclic redundancy check) generation, convolutional encoding, symbol repetition, and interleaving. The Random Binary Frame Generator masked subsystem generates random data that act as information bits. The Base Station Transmitter Data Rate masked subsystem provides the selection of the data rate. The IS-95A CRC Generator library block appends the CRC bits to the information bits. These CRC bits are used to detect errors in the data frame at the receiver. The IS-95A Fwd Ch Convolutional Encoder library block convolutionally encodes the data using a 1/2-rate encoder for protection against channel errors. Because IS-95A supports variable data rate operation, the data frame at this stage can have a number of different sizes. Depending on the data rate, the IS-95A Fwd Ch Repeater library block may repeat the bits it receives to create a data frame of 384 symbols. Then the IS-95A

Fwd Ch Interleaver library block interleaves the data frame for protection against the localized error bursts that can occur in fading channel conditions. The IS-95A Long Code Generator library block generates the long code used to scramble the data. The IS-95A Fwd Ch Scrambler library block uses the decimated long code input to scramble the input traffic frame and insert power bits. The scrambled traffic channel data and the data symbols from the various other types of channel sources are input to the Spreading and Modulation subsystem. These inputs are orthogonally encoded by their respective Walsh codes, added, and spread with the in-phase and quadrature components of the short PN sequence. Part of this is accomplished by the IS-95A Fwd Ch Base Station Transmitter Interface library block, which is inside the Spreading and Modulation subsystem. The signal generated is processed by the pulse shaping Transmit Filter block, which generates the modulated I and Q waveforms. The Rayleigh Multipath and AWGN Channel subsystem simulates the propagation through multiple paths of a Rayleigh fading channel. Complex white Gaussian noise is added to the channel output. This noise represents the interference generated by other base stations that are using the same frequency band. The interfering signal subsystem simulates interference generated by other users in the same base station. The receiver section of the system is responsible for the recovery of the data symbols transmitted on the traffic channel. The operations performed in this section include the receive filtering, the rake correlator, the rake demodulator, and descrambling. The Receive Filter block performs FIR [5] filtering on the I and Q sample streams with a filter that is matched to the transmit filter to maximize the in-band signal-to-noise ratio. The IS-95A Fwd Ch Detector library block is a masked subsystem with several components inside. The rake receiver computes symbol duration correlations for the Traffic data and Pilot symbols. These correlation values are used by the IS-95A Fwd Ch Rake Demodulator to recover the Traffic channel symbols. The Traffic symbols are further processed by the IS-95A Fwd Ch Descrambler to obtain the decision values for the original transmitted data symbols. The IS-95A Fwd Ch Deinterleaver library block deinterleaves the input data to restore the original symbol ordering. The IS-95A Fwd Ch Derepeater library block derepeats the symbols depending on the symbol rate, which involves averaging the symbols that

were repeated. The resulting frame is then provided as input to the IS-95A Fwd Ch Viterbi Decoder library block, which retrieves the information that was previously encoded. The decoded information bits and the CRC bits are provided to the IS-95A Frame Quality Detector library block. The final metrics from the IS95A Fwd Ch Viterbi Decoder block are also input to the IS-95A Frame Quality Detector block, which decides whether the frame was correctly received. The IS-95A Frame Quality Detector block outputs the Quality Indicator signal, as well as the information bits without the CRC bits. One Error Rate Calculation block compares the information bits to the bits generated at the source. Finally, the resultant bit error rate is displayed. This simulation uses the raw BER as the measure of the performance under the channel and noise conditions selected. [4] 4.2 Reverse Traffic Channel Model The Simulink model of IS-95A Reverse Traffic Channel is given in figure 4.2 (see last page). The IS-95A Reverse Traffic Channel Model shows the channel coding, modulation and spreading of the data symbols at the mobile station and the corresponding decoding as well as despreading and noncoherent demodulation at the base station receiver. The transmitter encodes the data, then symbols are modulated by a Walsh modulator and spread, and a randomized gating is applied to the transmit bursts. The channel model adds noise to simulate errors in transmission. The receiver side retrieves the information bits by performing the decoding and demodulation. The bit error rate for the data is displayed in the simulation. The IS-95A Reverse Traffic Channel Model uses these library blocks from the CDMA Reference Blockset: IS-95A CRC Generator IS-95A Frame Quality Detector IS-95A Rev Ch Convolutional Encoder IS-95A Rev Ch Interleaver/Deinterleaver IS-95A Rev Ch Repeater/Derepeater IS-95A Rev Ch Viterbi Decoder IS-95A Rev Ch Detector IS-95A Rev Ch Walsh Modulation and Spreading IS-95A Short Code Generator

Two library blocks from the Communication Blockset AWGN Channel

Rayleigh Multipath Fading Channel

and two from the Signal Processing Blockset Transmit Filter Receive Filter The mobile station transmitter section performs the Cyclic Redundancy Check (CRC) generation, convolutional encoding, symbol repetition, and interleaving. The Random Binary Frame Generator masked subsystem generates random data that act as information bits. The Mobile Station Transmitter Data Rate masked subsystem provides the selection of the data rate. The IS-95A CRC Generator library block appends the CRC bits to the information bits. These CRC bits are used to detect errors in the data frame at the receiver. The IS-95A Rev Ch Convolutional Encoder library block convolutionally encodes the data using a 1/2-rate encoder for protection against channel errors. Because IS-95A supports variable data rate operation, the data frame at this stage can have a number of different sizes. Depending on the data rate, the IS-95A Rev Ch Repeater library block may repeat the bits it receives to create a data frame of 384 symbols. Then the IS-95A Rev Ch Interleaver library block interleaves the data frame for protection against the localized error bursts that can occur in fading channel conditions. The short pseudonoise (PN) code is used for the inphase and quadrature spreading of the signal. The IS95A Short Code Generator library block generates the short PN code. The Spreading and Modulation subsystem contains several blocks that are responsible for the Walsh modulation, and the spreading with the long and short PN codes. The IS-95A Rev Ch Walsh Modulation and Spreading library block contains the IS-95A Rev Ch Burst Randomizer library block, which processes the long code and generates a gating signal based on the long code and the data rate of the input frame. The IS95A Rev Ch Walsh Modulation and Spreading library block also groups the input data in 6-bit groups, maps each group of 6 bits to a 64-symbol Walsh code, upsamples these 64 symbols by a factor of 4 to bring the result to chip rate, and then spreads the upsampled symbols with the gated long code. This gating ensures that the transmission is only performed for a fraction of the frame duration (half the time for half rate, and so on). Finally, other portions of the Spreading and Modulation subsystem spreads the data in quadrature by the PN code. The Transmit Filter block generates the I and Q waveforms. The Q waveform is delayed by a 1/2-chip duration relative to the I waveform.

The channel used is the same as that in the forward link which is implicit as the user and the base station are in the same environment. In the receiver section, the incoming signal is first filtered by the Receive Filter block, which implements a filter matched to the transmit filter. The filters in this model are designed to maximize the signal power within the desired frequency band. Then the filtered signal is sent to the IS-95A Rev Ch Detector library block, which contains the reverse channel rake receiver. The rake receiver consists of three rake fingers that are set to different delays to handle up to three multipaths. Each active rake finger performs the despreading of the input data with the short PN sequence, followed by despreading with the long code. This is followed by the correlation with the entire set of 64 Walsh codes. The energies in the I and Q components are added, and the results from the fingers are added together. This is processed by the Walsh demodulator, which generates decisions in groups of 6 bits, the size used for modulation. The Walsh demodulator outputs both soft decisions and bipolar-valued hard decisions. Both are gated by the data burst randomizer signal. The IS-95A Rev Ch Deinterleaver library block deinterleaves the input data to restore the original symbol ordering. The IS-95A Rev Ch Derepeater library block derepeats the data depending on the symbol rate, which involves averaging the symbols that were repeated. The resulting frame is then provided as input to the IS-95A Rev Ch Viterbi Decoder library block, which retrieves the information that was previously encoded. The decoded information bits and the CRC bits are provided to the IS-95A Frame Quality Detector library block. The final metrics from the IS95A Rev Ch Viterbi Decoder block are also input to the IS-95A Frame Quality Detector block, which decides whether the frame was correctly received. The IS-95A Frame Quality Detector block outputs the Quality Indicator signal, as well as the information bits without the CRC bits. Error Rate Calculation block compares the information bits to the bits generated at the source. Finally, the resultant bit error rate is displayed. 5. OBSERVATIONS AND RESULTS The basic properties that are studied are: Noise Effect in the channel Multipath Fading Effect in the channel Effect of Multiuser interference All these properties are studied by measuring BER under different conditions. The simulation is done for 50 frames of data.

5.1 Noise Analysis with AWGN Channel BER [6] is measured under different SNR [6] environments. The following graph is plotted with the results obtained.
BER VS SNR WITH AWGN 50 45 40 35 30 BER in % 25 20 15 10 5 0 -20 -19 -18 -17 -16 -15 -14 SNR in dB -13 -12 -11 -10

As shown in the graph, the BER performance of the system increases dramatically with the increase in number of fingers involved in the Rake receiver. The improve in BER performance according to increase in the number of Rake fingers is due to the improved combined received signal quality after maximal combining as the number of fingers in the rake receiver increases. Also maximal combing improves the strength of the received signal by combining all the correlators output according to their signal strength, and also, the received signal will contain much less error than many of the multipath signals received. In common IS-95 CDMA forward link, 3-finger Rake receiver is widely adopted and is still capable in delivering high received signal quality in extremely poor SNR environments like SNR = -10 dB, as shown by the graph.
BER VS. SNR WITH RAYLEIGH FADING CHANNEL 50 45 40 35 BER in % 30 25 20 15 10 5 0 -20 -18 -16 -14 SNR in dB -12 -10 -8

Figure 5.1 BER Vs SNR (AWGN) The results obtained have been accurate and effective in demonstrating the effect of noise in affecting the BER performance of the communication system. As shown by the results, the simulated system is able to achieve 0 BER at SNR>=-13.5 dB and starts to degrade gradually from SNR=-15 dB. 5.2 Multipath Faidng Effect with Rayleigh Fading Channel Two tests are performed in this simulation 1. BER versus Number of Rake Receiver Fingers 2. BER versus SNR The following graphs are plotted from the results obtained:
BER VS. NO. OF RAKE FINGERS 30

Figure 5.3 BER Vs SNR (Reyleigh Fading) BER starts to degrade with increasing SNR and becomes zero at SNR>=-9.5 dB. Comparing with the result of noise analysis with AWGN, we see that in the presence of multipath signals, the BER is higher at the same value of SNR. With AWGN BER=0 at SNR=-13 dB but with Rayleigh Multipath BER=0 at SNR=-9.5 dB.

25 SNR = -10 dB 20 BER in %

15

10

2 No. of Rake Receiver Fingers

Figure 5.2 BER Vs No. of Rake Fingers

COMPARISION OF AWGN AND RAYLEIGH FADING CHANNEL 50 45 40 35 BER in % 30 25 20 15 10 5 0 -20 -18 -16 -14 SNR in dB -12 -10 -8 AWGN Channel Rayleigh Fading Channel

the communication at least above -5 dB, as proven by the result. 6. CONCLUSION It has been deduced that by keeping proper SNR we are able to reduce AWGN to a tolerable extent. Similarly, by choosing three rake fingers and keeping a proper value of SNR the Reyleigh fading effect has been minimized. It has been observed that out of optimal 42 users in a sector, 20 users can be supported with minimum error rate and even after reaching the optimum count the BER is within tolerable limits. 7. REFERENCES [1]. Vijay K. Garg, IS-95 CDMA and cdma 2000; Pearson Education, Inc, India, First Edition 2000. [2]. William Stallings, Wireless Communication And Networks, Pearson Education, Inc, India, First Reprint, 2002. [3]. Samuel C. Yang, CDMA RF System Engineering; Artech House, Inc, Boston, London, 1998. [4]. CDMA Reference Block Set Users Guide, The MATHWORKS, Inc, and algorex, Inc, 2000-2001. [5]. John G Proakis, Digital Signal Processing, Pearson Education, Inc, 3rd Edition, 1996. [6]. A.Bruce Carlson, Communication Systems, McGraw Hill, 4th Edition, 2002. 8. ABOUT THE AUTHORS

Figure 5.4 Comparision Of AWGN And Reyleigh Faded Channel 5.3 Multiuser Interference Analysis
BER VS> NO. OF USERS 22 20 18 16 14 BER in % 12 10 8 6 4 2 0 0 5 10 15 20 25 No. of users 30 35 40 45

Figure 5.5 BER Vs No. Of Users It is shown from the graph that the simulated IS-95 CDMA communication system is immune to multiuser interference in SNR environments with SNR>=-5 dB. It can support 25 users with BER=0 and approximately 40 users BER<=20%. However, when the number of interfering signals in a channel increases, the noise level in the system also increases and hence the SNR in the channel. As, a result the BER performance of the communication system starts to degrade. The systems immunity to multiuser interference is due to the fact that interfering signals of other users uses (1) different long PN sequences, (2) different Walsh code and (3) different short PN sequences. Although these interfering signals have contribution in raising the noise background, however they are unable to affect the BER performance of the system, provided that the transmitter transmits at an adequate power level to keep the SNR in Muhammad Umair Siddiqui is currently a student of final year Electronic Engg. at NED University of Engg. & Tech. Karachi. His areas of interest are 4G Communication Systems and Wireless Comminucations. Qazi Safiullah Shahrukh is currently a student of final year Electronic Engg. at NED University of Engg. & Tech. Karachi. His areas of interest are Wireless Comminucations and UMTS. Muhammad Usman is currently a student of final year Electronic Engg. at NED University of Engg. & Tech. Karachi. His areas of interest are emerging Communication technologies and Mobile Communication.

IS-95A Forward Traffic Channel (From BS to MS) Model

[Rate]

Rate

Rate

Rate

Rate

Rate IS-95A Fwd Ch Interleaver/ Deinterleaver Buffer IS-95A Fwd Ch Interleaver/Deinterleaver IS-95A Long Code Generator IS-95A Long Code Generator BER Random Power Bits To Frame To Frame Data In Long Code Power Bits IS-95A Fwd Ch Scrambler

Raw Data

Frame In IS-95A CRC Generator

Frame In IS-95A Fwd Ch Convolutional Encoder

Frame In IS-95A Fwd Ch Repeater/Derepeater

Combine with Sync & Paging --SS Modulate Spreading and Modulation Transmit Filter

Data Source

Multipath Rayleigh Fading Channel

Multipath Rayleigh Fading

Tx Error Rate Calculation Sel Rx [Rate]

Rate Idx

Bit Error Rate Number of Errors Number of Bits

AWGN Channel AWGN
Users

Determine Data Index

Error Rate Calculation

[Rate]
Rate Walsh Seq

Interfering Signal from other users IS-95A Walsh Code Generator [ 96 1 116 1 239 1 0 0] Initial Phases and Finger Enables (0 0) Short PN Mask Receive Filter

Quality Indicator

Frame Out

Rate

Rate

Frame In Frame Out Metric

Metric

Frame In

Frame In IS-95A Fwd Ch Derepeater

IS-95A Fwd Ch Interleaver/ Deinterleaver IS-95A Fwd Ch Deinterleaver

Rx Signal Path Delay /Enable Short PN Mask

IS-95A Frame Quality Detector

IS-95A Fwd Ch Viterbi Decoder

IS-95A Fwd Ch Detector

Figure 4.1 IS-95A Forward Traffic Channel Model in Simulink

IS-95A Reverse Traffic Channel (From MS to BS) Model

[Rate]

Mobile Station Transmitter Data Rate

Rate

Rate

Rate

Rate

Rate Frame In Out

Frame In Random Binary Frame Generator IS-95A CRC Generator

Frame In IS-95A Rev Ch Convolutional Encoder

Frame In IS-95A Rev Ch Repeater

Frame In IS-95A Rev Ch Interleaver PN I

PN I PN Q

Spreading and Modulation

Pulse Shaping Transmit Filter

[0 0] PN Mask
PN Q IS-95A Short Code Generator BER Tx Error Rate Calculation Sel Rx [Rate]
Rate Idx

Multipath Rayleigh Fading

Multipath Rayleigh Fading

Bit Error Rate Number of Errors Number of Bits

AWGN Channel AWGN

Determine Data Index

Error Rate Calculation

[Rate]

Users

Quality Indicator

Rate

Frame Out

Rate

Rate

Rate

Interfering Signal from other users

Frame In Frame Out Metric

Rate

Hard Dec Rx Signal

Metric

Frame In

Frame In IS-95A Rev Ch Derepeater

Frame In IS-95A Rev Ch Deinterleaver

Sof t Dec

Path Delay /Enable Short PN Mask

IS-95A Frame Quality Detector

IS-95A Fwd Ch Viterbi Decoder

Pulse Shaping Receive Filter1

IS-95A Rev Ch Detector [ 96 1 116 1 0 0] (0 0) Short PN Mask Initial Phases and Finger Enables

Figue 4.2 IS-95A Reverse Traffic Channel Model in Simulink