Comparison of Orthogonal Frequency-Division Multiplexing and Pulse-Amplitude Modulation in Indoor Optical Wireless Links Daniel J

Comparison of Orthogonal Frequency-Division Multiplexing and Pulse-Amplitude Modulation in Indoor Optical Wireless Links Daniel J

IEEE TRANSACTIONS ON COMMUNICATIONS, VOL. 60, NO. 1, JANUARY 2012 153 Comparison of Orthogonal Frequency-Division Multiplexing and Pulse-Amplitude Modulation in Indoor Optical Wireless Links Daniel J. F. Barros, Sarah K. Wilson, Senior Member, IEEE, and Joseph M. Kahn, Fellow, IEEE Abstract—We evaluate the performance of three direct- that is unregulated worldwide. Light in the infrared or visible detection orthogonal frequency-division multiplexing (OFDM) range penetrates through glass, but not through walls or other schemes in combating multipath distortion in indoor opti- opaque barriers, so that optical wireless transmissions are cal wireless links, comparing them to unipolar M-ary pulse- amplitude modulation (M-PAM) with minimum mean-square confined to the room in which they originate. Furthermore, error decision-feedback equalization (MMSE-DFE). The three in a visible or infrared wireless link employing intensity OFDM techniques are DC-clipped OFDM and asymmetrically modulation with direct detection (IM/DD), the short carrier clipped optical OFDM (ACO-OFDM) and PAM-modulated dis- wavelength and large-area photodetector lead to efficient spa- crete multitone (PAM-DMT). We describe an iterative procedure tial diversity that prevents fading. Nevertheless, the existence to achieve optimal power allocation for DC-OFDM. For each modulation method, we quantify the received electrical SNR of multiple paths between the transmitter and receiver causes required at a given bit rate on a given channel, considering an multipath distortion, particularly in links using non-directional ensemble of 170 indoor wireless channels. When using the same transmitters and receivers, or in links relying upon non-line- symbol rate for all modulation methods, M-PAM with MMSE- of-sight propagation [1],[2]. This multipath distortion can lead DFE has better performance than any OFDM format over a to intersymbol interference (ISI) at high bit rates. range of spectral efficiencies, with the advantage of (M-PAM) increasing at high spectral efficiency. ACO-OFDM and PAM- Multicarrier modulation has been proposed to combat ISI DMT have practically identical performance at any spectral in optical wireless links, since the symbol period of each efficiency. They are the best OFDM formats at low spectral subcarrier can be made long compared to the delay spread efficiency, whereas DC-OFDM is best at high spectral efficiency. When ACO-OFDM or PAM-DMT are allowed to use twice caused by multipath distortion [5]. Multicarrier modulation is the symbol rate of M-PAM, these OFDM formats have better usually implemented by orthogonal frequency-division multi- performance than M-PAM. When channel state information is plexing (OFDM) [3],[6],[7]. The main drawback of multicar- unavailable at the transmitter, however, M-PAM significantly rier modulation in systems using intensity modulation (IM) is outperforms all OFDM formats. When using the same symbol the high DC bias required to make the multicarrier waveform rate for all modulation methods, M-PAM requires approximately three times more computational complexity per processor than nonnegative. There have been several approaches for reducing all OFDM formats and 63% faster analog-to-digital converters, the DC bias in IM OFDM systems. The first technique uses assuming oversampling ratios of 1.23 and 2 for ACO-OFDM hard-clipping on the negative signal peaks in order to reduce and M-PAM, respectively. When OFDM uses twice the symbol the DC bias required. This method is called DC-clipped rate of M-PAM, OFDM requires 23% faster analog-to-digital OFDM (DC-OFDM) [3],[6]. Another technique clips the converters than M-PAM but OFDM requires approximately 40% less computational complexity than M-PAM per processor. entire negative excursion of the waveform. Impairment from clipping noise is avoided by appropriate choice of the data- Index Terms—Optical wireless, infrared wireless, visible light bearing subcarrier frequencies [7]. This technique is called communications, communications system performance, multi- carrier optical systems, orthogonal frequency-division multiplex- asymmetrically clipped optical OFDM (ACO-OFDM) [7]. A ing, power allocation, intensity modulation with direct detection, third technique also clips the entire negative excursion, but PAM-DMT, ACO-OFDM, OOK, OFDM, MMSE equalizers. modulates only the imaginary parts of the subcarriers such that the clipping noise becomes orthogonal to the desired signal. I. INTRODUCTION This technique is called PAM-modulated discrete multitone (PAM-DMT) [8]. NDOOR optical wireless transmission has been studied I extensively in recent decades [1]–[4]. The visible and There have been several studies comparing the performance infrared spectral regions offer virtually unlimited bandwidth of different OFDM techniques (e.g., [7]) but these compar- isons have been made for ideal additive white Gaussian noise Paper approved by A. Bononi, the Editor for Optical Transmission and (AWGN) channels. To our knowledge, previous studies have Networks of the IEEE Communications Society. Manuscript received October 21, 2010; revised July 9, 2011. not compared the OFDM methods to conventional baseband This work was supported by the Portuguese Foundation for Science and methods, such as on-off keying (OOK) or unipolar pulse- Technology scholarship SFRH/BD/22547/2005, and by a Stanford Graduate amplitude modulation (PAM), nor have they considered the Fellowship. D. J. F. Barros and J. M. Kahn are with the E. L. Ginzton Laboratory, dispersive nature of optical wireless channels. Furthermore, Department of Electrical Engineering, Stanford University, Stanford, CA in previous work, the powers of the subcarriers and the DC 94305-9515 USA (e-mail: [email protected]; [email protected]). bias for DC-OFDM were not jointly optimized according to S. K. Wilson is with the Department of Electrical Engineering, Santa Clara University, Santa Clara, CA, USA (e-mail: [email protected]). the channel frequency response in order to obtain the lowest Digital Object Identifier 10.1109/TCOMM.2011.112311.100538 required optical power. We present an iterative procedure for 0090-6778/12$31.00 ⃝c 2012 IEEE 154 IEEE TRANSACTIONS ON COMMUNICATIONS, VOL. 60, NO. 1, JANUARY 2012 electrical OFDM signal is demodulated and equalized with a single-tap equalizer on each subcarrier to compensate for channel distortion [9]. B. Optical Wireless Channel Multipath propagation in an indoor optical wireless channel [10],[11] can be described by an impulse response ℎ() or ∫by the corresponding baseband frequency response ()= ∞ ℎ () −2 −∞ . Including noise, the baseband channel model is [2]: Fig. 1. System model for indoor optical wireless with direct-detection OFDM. ()= ⋅ () ∗ ℎ ()+ () , (1) where () is the detected photocurrent, () is the transmitted DC-OFDM based on known bit-loading algorithms with a new intensity waveform, is the photodetector responsivity, and modification, the bias ratio, in order to obtain the optimum () represents ambient light shot noise and thermal noise. power allocation. Optical wireless channels differ from electrical or radio fre- The optimum detection technique for unipolar PAM in the quency channels because the channel input () represents presence of ISI is maximum-likelihood sequence detection instantaneous optical power. Hence, the channel input is (MLSD), but its computational complexity increases exponen- nonnegative ( () ≥ 0) and the average transmitted optical tially with the channel memory. ISI in optical wireless links power is given by is well-approximated as linear in the instantaneous power [2], ∫ and for typical wireless links, PAM with minimum mean- 1 = lim () , (2) square error decision-feedback equalization (MMSE-DFE) →∞ 2 − achieves nearly the same performance as MLSD and requires 2 rather than the usual time-average of ∣()∣ , which is appro- far less computational complexity. Hence, we compare the priate when () represents amplitude. The average received performance of the three aforementioned OFDM techniques optical power can be written as using optimized power allocations to the performance of PAM with MMSE-DFE at different spectral efficiencies. = (0) , (3) This paper is organized as follows. We present our system where (0) is the DC gain of the channel, i.e., (0) = and indoor optical wireless models in Section II. In Section ∫ ∞ ℎ () . III, we review the different OFDM formats. In Section IV, −∞ We use the methodology developed by Barry et al [10] we compare the receiver electrical SNR required to transmit to simulate the impulse responses of indoor optical wireless at several bit rates for the different OFDM formats and for channels, taking account of multiple bounces. A similar model unipolar M-PAM with MMSE-DFE equalization at different can be found in [12]. The algorithm in [10] partitions a room spectral efficiencies. Furthermore, we compare the receiver into many elementary reflectors and sums up the impulse electrical SNR required for the different modulation formats response contributions from ℎ-order bounces, ℎ() () ,= when there is no channel state information (CSI) available at 0, 1, 2,.... More recently, Carruthers developed an iterative the transmitter. We also compare the computational complex- version of the multi-bounce impulse response algorithm which ity required for OFDM and M-PAM at different bit rates. We greatly reduces the computational time required to accurately present conclusions in Section V. model optical wireless channels [13]. In this study, we use a

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    11 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us