Sky DSP logo  
Index    Chapter 1    Chapter 2    Chapter 3    Conclusion    Appendix    Matlab Code    Audio Demo   


The suitability of OFDM as a modulation technique for wireless telecommunications, with a CDMA comparison.

Revised: 16/10/2001 (Fixed up several mistakes in the simulations and fixed up spelling and grammar).

Print Friendly Version in PDF (519 kB)
This is a PDF of the complete thesis, with no Matlab Code
Original version of the thesis (966 kB)
Thesis submitted by Eric Lawrey in October 1997 in partial fulfilment of the requirements for the Degree of Bachelor of Engineering with Honours in Computer Systems Engineering at James Cook University.

This thesis investigates the effectiveness of Orthogonal Frequency Division Multiplexing (OFDM) as a modulation technique for wireless radio applications. The main aim was to assess the suitability of OFDM as a modulation technique for a fixed wireless phone system for rural areas of Australia. However, its suitability for more general wireless applications is also assessed.

Several of the main factors affecting the performance of a OFDM system were measured, including multipath delay spread, channel noise, distortion (clipping), and timing requirements. The performance of OFDM was assessed by using computer simulations performed using Matlab, and practical measurements. These measurements were performed by recording a low bandwidth (audio) OFDM signal, generated using Matlab, on to a tape player. This recorded signal was then played back and recorded using the sound card of a PC. This was then decoded using a Matlab script.

Most third generation mobile phone systems are proposing to use Code Division Multiple Access (CDMA) as their modulation technique. For this reason, CDMA was also investigated so that the performance of CDMA could be compared with OFDM. 

It was found that OFDM performs extremely well compared with CDMA, providing a very high tolerance to multipath delay spread, peak power clipping, and channel noise. In addition to this it provides a high spectral efficiency.

OFDM was found to have total immunity to multipath delay spread provided the reflection time is less than the guard period used in the OFDM signal. In fact, multipath signals lead to a strengthening of the received signal, improving the performance. In a typical system a delay spread of up to 100 msec could be tolerated, corresponding to multipath reflections of 30 km. The only problem caused by multipath is frequency selective fading, which can result in carriers being heavily attenuated due to destructive interference at the receiver. This can result in the carriers being lost in the noise. 

For the modulation schemes investigated (BPSK, QPSK and 16 QAM), clipping of the OFDM signal was found to have little effect on the performance of the system, allowing the peak power of the signal to be clipped by up to 6 - 9dB before the symbol error rate became significant. This tolerance to clipping reduces the dynamic range overhead required in output stages of OFDM transmitters.

The noise performance of OFDM was found to depend solely on the modulation technique used for modulating each carrier of the signal. The performance of the OFDM signal was found to be the same as for a single carrier system, using the same modulation technique. The minimum signal to noise ratio (SNR) required for BPSK was ~7 dB, where as it was ~12 dB for QPSK and ~25 dB for 16PSK.

CDMA was found to perform poorly in a single cellular system, with each cell only allowing 7-16 simultaneous users in a cell, compared with 128 for OFDM. This was for a 1.25 MHz bandwidth and 19.5 kbps user data rate. This low cell capacity of CDMA was attributed to the use of non-orthogonal codes used in the reverse transmission link, leading to a high level of inter-user interference.

The only main weak point that was found with using OFDM, was that it is very sensitive to frequency, and phase errors between the transmitter and receiver. The main sources of these errors are frequency stability problems; phase noise of the transmitter; and any frequency offset errors between the transmitter and receiver. This problem can be mostly overcome by synchronizing the clocks between the transmitter and receiver, by designing the system appropriately, or by reducing the number of carriers used.
Table of Contents

Download OFDM Matlab Code
This is the code that was used in the thesis.

Audio Data Demonstration
Shows the effect of noise and signal clipping on audio data sent using OFDM

1. Introduction

1.1 Third Generation Wireless Networks 1.1.1 Evolution of Telecommunication Systems.
1.1.2 Overall Aims of Universal Mobile Telecommunications System
1.1.3 Teleservices
1.1.4 UMTS Environments
1.1.5 Cell types
1.1.6 Radio Interface
1.1.7 Satellite Networking
1.1.8 Timetable for System Implementation
1.1.9 Conclusion
1.2 Propagation Characteristics of mobile radio channels 1.2.1 Attenuation
1.2.2 Multipath Effects
1.2.3 Doppler Shift
1.3 Multiple Access Techniques 1.3.1 Frequency Division Multiple Access
1.3.2 Time Division Multiple Access
1.3.3 Code Division Multiple Access
1.3.4 CDMA Process Gain
1.3.5 CDMA Generation
1.3.6 CDMA Forward Link Encoding
1.3.7 CDMA Reverse Link Encoding
1.3.8 Orthogonal Frequency Division Multiplexing
1.3.9 OFDM generation
1.3.10 Adding a Guard Period to OFDM
2. OFDM Results 2.1 OFDM Model Used 2.1.1 Serial to Parallel Conversion
2.1.2 Modulation of Data
2.1.3 Inverse Fourier Transform
2.1.4 Guard Period
2.1.5 Channel
2.1.6 Receiver
2.1.7 OFDM Simulation Parameters
2.2 OFDM Simulated Results 2.2.1 Multipath Delay Spread Immunity
2.2.2 Peak Power Clipping
2.2.3 Gaussian Noise Tolerance of OFDM
2.2.4 Timing Requirements
2.3 Practical Measurements 2.3.1 Extended Model
2.3.2 Transmission Protocol
2.3.3 Video Recorder
2.3.4 Peak OFDM Performance for the VCR link
2.3.5 Audio Tape Player
2.4 Picture quality verse signal to noise ratio 2.4.1 Experimental comparison between QPSK and power averaged 256PSK
2.4.2 Results
2.5 Mathematical Model for OFDM performance 2.5.1 RMS Demodulated Phase Error
2.5.2 BER verses Channel Noise
2.6 OFDM system implementation 2.6.1 Using general purpose DSPs
2.6.2 Future DSP Processing Power
2.6.3 Hardware FFT Implementation
3. CDMA Results 3.1 Simulated Model 3.1.1 Forward Link
3.1.2 Reverse Path
3.2 Simulation Results 3.2.1 BER verses the number of users in a cell 3.3 Mathematical Model for Reverse Link 3.3.1 Cell Capacity for a CDMA system
3.3.2 Capacity of a single CDMA cell
3.3.3 Capacity of CDMA and OFDM with Multiple Cells
4. Conclusion


Appendix I. Acronyms

Appendix II. OFDM Guassian Noise Performance Prediction

Appendix III. BER verses Eb/No for a CDMA system


I would like to thank my supervisor Prof. Keith Kikkert for the endless hours of help, suggestions, ideas and advise during the development of this thesis. I would also like to thank Prof. Greg Allen for his advise, focus, and pep talks to keep me on track.

Finally, I would like to thank my computer for only crashing seriously once during the writing of this thesis.

Copyright 1997-2001 Eric Lawrey,   Last Updated: 2nd December 2001
Copyright © 2001 Sky DSP