IEEE COMMUNICATION LETTERS, VOL. 4, NO. 9, SEPTEMBER 2000 277 Comparison of Convolutional and Turbo Codes for OFDM with Antenna Diversity in High-Bit-Rate Wireless Applications Lang Lin, Student Member, IEEE, Leonard J. Cimini, Jr., Fellow, IEEE, and Justin C.-I. Chuang, Fellow, IEEE Abstract—This letter presents simulation results for the applica- tion of Turbo coding to an OFDM system with diversity. First, re- sults are presented for convolutional and Reed–Solomon codes. It is shown that even with short constraint lengths, convolutional codes have the potential to outperform Reed–Solomon codes, provided that sufficient precision is used in the soft decoder. We then eval- uate the performance of Turbo codes under slow fading conditions and study the effects of varying codeword size. Increasing code- word size theoretically provides better interleaving between the two component codes. However, this advantage is less clear when the fading rate is significantly lower than the symbol rate, which is typical of the high-data-rate systems considered here. Under such conditions, the advantage of using two component convolutional codes in Turbo codes is limited. A single convolutional code with a long constraint length may be a better choice. I. INTRODUCTION T HE PUBLIC’S desire for mobile communications and computing, as evidenced by the popularity of cellular phones, pagers, and laptop computers, combined with the rapid growth in demand for Internet access, suggest a very promising future for wireless data services. The key to realizing this potential is the development and deployment of high-per- formance radio systems, supporting high data rates with wide-area coverage. One of the more promising candidates for achieving high-data-rate transmission in a mobile environment is orthogonal frequency division multiplexing (OFDM) [1], [2]. In OFDM, the signal bandwidth is divided into many narrow subchannels which are transmitted in parallel. Each subchannel is typically chosen narrow enough to eliminate the effects of delay spread. In [3], a technique for achieving a reliable, high-speed (1–2 Mb/s) wireless data service for mobile and portable cellular systems was proposed. By combining OFDM with multiple transmit antennas, receive antenna diversity, and Reed–Solomon coding, the link-budget and dispersive-fading limitations of the cellular mobile radio environment can be overcome and the effects of co-channel interference can be reduced. A significant advantage to combining the use of multiple transmission antennas with OFDM is that interleaving Manuscript received November 2, 1999. The associate editor coordinating the review of this letter and approving it for publication was Prof. M. Fossorier. L. Lin is with the Wireless Information Network Laboratory, Rutgers The State University of New Jersey, NJ 08854-8060 USA (e-mail: llin@winlab.rut- gers.edu). L. J. Cimini, Jr. and J. C.-I. Chuang are with the Wireless Systems Research Department, AT&T Laboratories - Research, Red Bank, NJ 07701 USA. Publisher Item Identifier S 1089-7798(00)07668-7. Fig. 1. OFDM system with antenna diversity. can be performed in time, frequency, and space. An important design consideration, therefore, is how to assign coded bits to transmission antennas and subchannels to best randomize the fading experienced by different symbols in a codeword [4]. In [5], a powerful coding technique, Turbo coding, has been shown to perform near the Shannon capacity limit in an addi- tive white Gaussian noise (AWGN) channel. Motivated by the performance advantages of OFDM and the potential of Turbo coding, in this letter, we consider the application of Turbo codes to the OFDM system described in [3]. II. OFDM SYSTEM ARCHITECTURE In order to provide a fair comparison, we evaluate the per- formance of both convolutional codes and Turbo codes using the system parameters in [3]. The architecture of the OFDM system is shown in Fig. 1. At the transmitter, the encoded data stream is sent to four OFDM transmission branches. The OFDM signal bandwidth, centered at 2 GHz, is divided into 120 6.25-kHz wide subchannels with QPSK modulation on each subchannel. Both differential and coherent demodulaton are considered. At the receiver, the demodulated signals from two receiving branches are combined using maximal ratio combining and then decoded. With a symbol period of 200 s (including a 40- s guard interval) and 1/2-rate coding, a maximum information rate of 600 kbps can be achieved in an 800-kHz bandwidth. In the simulations, we model the wireless channel as a Rayleigh-fading channel with a two-ray multipath delay profile. A 40- s impulse separation and a 200-Hz maximum Doppler frequency are assumed. Since the fading rate is signif- icantly lower than the 5-kHz symbol rate, the channel varies 1089–7798/00$10.00 © 2000 IEEE