PPLICTION OF OFDM TECHNIQUE TO UNDERWTER COUTIC DT TRNMIION 1 IWON KOCH K, HENRYK LOT Gdansk University of Tehnology 11/12, G. Narutowiza t., 80-233 Gda sk, Poland ia.kohanska@gmail.om; hlas@eti.pg.gda.pl Performanes of underwater aousti ommuniation (UC) digital systems are strongly related to speifi transmission properties of the underwater hannel. Depending on the harateristis of the hannel, an arhiteture and modulation tehniques are usually implemented that are known as reliable solutions for data transmission in diffiult radio hannels. The OFDM tehnique seems to be the most promising nowadays. The parameters of the physial layer and data link layer of IO/OI model are adapted to the needs of underwater ommuniations, so as to best protet the signal from distortions and interferenes typial for the underwater environment. This paper ompares the performane of UC and wireless teleommuniations systems working in hannels of vary different transmission parameters. Based on these parameters, some physial and link layer onfigurations of OFDM data transmission systems are proposed for implementation in a laboratory model and testing at a lake. INTRODUCTION Data transmission rates in underwater aousti ommuniation (UC) systems are muh lower than available in wireless ommuniations. The reason is not only that the arrier frequeny range of ultrasoni signals is situated muh lower than radio frequenies, resulting in proportionally lower effetive bandwidth - here some progress an be expeted from the tehnology of broadband transduers for aousti antennae; even more disadvantageous is the influene of ommuniation properties ultrasound meets in inland and sea waters. Due to the large variability of properties of underwater ommuniation hannels, mostly lassified as shallow underwater aousti (UW) hannels, there is a need for adaptive mathing of the UC systems ommuniation signaling to instantaneous, often strongly varying, hannel onditions. Without optimizing signal parameters in both physial and link 1 upported by a grant from the Ministry of iene and Higher Eduation, deision 4706/B/T02/2011/40. 91
HYDROCOUTIC Volume 14 layers of ommuniation protools, a ommuniation system would be fored to work with settings assuming the worst possible onditions; this would strongly limit its effetive bandwidth, range, speed and effiieny. daptation of modulation and oding tehniques, known from wireless teleommuniation systems, to the properties of the hannel should signifiantly improve the quality of ommuniation. 1. OFDM TECHNIQUE mong the designers of underwater ommuniation systems, onsiderable interest in the so-alled orthogonal frequeny division multiplexing (OFDM) tehnique is observed [1, 2]. OFDM is a tehnique of partiular interest to groups developing ontemporary wireless ommuniations standards. OFDM is a spetrally effiient and II-resistant modulation and oding tehnique, ommonly used in wide-band DL teleommuniation modems. Its flexibility allows to math the transmitted signal parameters to the urrent onditions in the ommuniation hannel. Mathing signal parameters to the hannel onditions requires knowledge of the harateristis of the hannel. Time-varying impulse response an be measured at the beginning of ommuniation, during the onnetion establishing proedure, known as a handshake, whih heks the link and allows the hannel to be tested in operation. The adaptation of OFDM tehnique to ultrasound underwater ommuniation requires assumption of a flexible hannel model, whih is a statistial model with assumption of wide sense stationarity of hannel harateristis (W) and unorrelated sattering (U) of multipath omponents arriving to the ommuniation system reeiver. tatistial parameters resulting from the instantaneous impulse response measurements and the adopted hannel model are essential for transmitted signal parameters alulation [3]. The hosen design methodology of OFDM modems transmission protool, inluding handshake, for underwater ommuniation is based on omparing the physial limitations of radio and underwater hannels. nalysis of the UC ommuniation onditions in terms of the radio OFDM systems hannels determines the physial layer parameters, suh as the OFDM symbol duration, length of yli prefix, and subarriers spaing. OFDM is a tehnique of partiular interest of ontemporary radio standards. The idea of orthogonal frequeny division multiplexing is a digital multi-arrier modulation method. It is based on splitting a high-speed broadband transmission into many slower narrow-band transmissions (Fig. 1). large number of losely-spaed orthogonal sub-arriers are used to arry data. OFDM modulator diagram is shown in Fig. 2. Fig. 1. OFDM signal spetrum [1] 92
Volume 14 HYDROCOUTIC The data is split into several parallel data streams, one for eah sub-arrier. Eah subarrier is modulated with a onventional modulation sheme (suh as QM, PK or DPK tehnique) at a low symbol rate, maintaining total data rates similar to onventional singlearrier modulation shemes in the same bandwidth. treams of omplex onstellation symbols are next proessed by the inverse Fourier transform, forming a sequene of omplex OFDM waveform samples. Than the sequene is onverted into an analog signal, whih modulates in quadrature the arrier of enter frequeny of OFDM signal spetrum. Bandwidth of OFDM modulated signal is the sum of adjaent subhannels, eah of bandwidth W. Time duration of a single OFDM symbol ist s 1/ W. Fig. 2. OFDM modulator sheme [4] Eah OFDM symbol is preeded by a yli prefix, i.e. a redundant repetition of the last segment of itself (Fig. 3). Cyli prefix of duration T G protets the OFDM signal against intersymbol interferenes. Fig. 3. OFDM symbol with yli prefix [1] Values of OFDM parameters are determined on the basis of the teleommuniations requirements of desired transmission bit rate and the harateristis of the ommuniation hannel. 2. COMPRION OF WIRELE ND UC OFDM YTEM Table 1 presents the omparison of data transmission harateristis of: (1) wireless loal area networks (WLN) [5], (2) the so-alled long term evolution (LTE) ellular networks [6 8], and (3) a hypothetial UC system designed on the basis of transmission properties of shallow underwater hannel measured in natural onditions of a river UW hannel as desribed in [9]. Whereas Doppler spread in radio hannels is in the order of 10-5 to 10-6 of the system bandwidth, in a shallow underwater aousti hannel it is in the order of 10-3 of the bandwidth. 93
HYDROCOUTIC Volume 14 Expeted data rates in UC system designed based on measurements in [9] does not exeed a few kilobits per seond. Higher data rates are reported in ases of underwater MIMO OFDM systems using multiple transmitters and reeivers [2, 10]. Table 1. Channel parameters of OFDM systems hannel onditions WLN LTE UC propagation veloity 8 3 10 m/s 8 3 10 m/s 1500 m/s arrier frequeny f 2.4 GHz 2.5 GHz 10 khz max movement veloity 2 m/s 36 m/s 2 m/s rms exess delay rms 100 s 1 ms 149 s oherene time t 26 ms 1.4 s 50 ms Doppler spread m 16 Hz ( 0.0001% B) 300 Hz ( 0.0015 % B) 3 Hz ( 0.15% B) oherene bandwidth 20 MHz 200 khz 2.2 khz f system bandwidth B 17 MHz 1.4, 3, 5,10,15, 20 MHz 2 khz Table 2. Performanes of OFDM systems system parameters WLN LTE UC OFDM symbol 4 s ( 400, t /6500) duration T rms 71.4 s( 72 rms, t /20) 12.5 m s ( 84 rms, t /4) OFDM subhannel 312.5 khz ( 20000 f m, 15 khz ( 50 f m, f /13) bandwidth f /64) 100kHz( 34f m, f /22) number subarriers 52 72, 180, 300, 600, 900, 1200 20 yli prefix T G 800 ns 4.7 s 2.5 ms modulation BPK, QPK, QPK, 16 QM, 64 - QM 16 QM, 64 - QM BPK, QPK, 16 - QM oding rate 1/ 2, 2/3, 3/4 1/3-1 1/ 2, 2/3, 3/4 data rate 6, 9,12,18, 24, Mbps 36, 48, 54 100 Mbps 0.4 3.6 kbps 3. PHYICL ND LINK LYER REQUIREMENT There is no typial underwater ommuniation hannel [9]. Eah environment has different harateristis that affet the behavior of a digital ommuniation system. However, there are lasses of underwater ommuniation hannels and it is possible to design the arhiteture of physial and link layers of IO/OI transmission system model, so as to operate in a satisfatory manner within the lass. In the physial layer, the hannel is understood through the theory of aousti wave propagation, its limitations and the measurement of impulse response. Knowledge of basi limitations of signaling in an underwater environment translates into inreasingly sophistiated digital ommuniation tehniques, tailored to the unique harateristis of the UC hannel. olutions adapted to the UC systems must have good adaptive and synhronization properties - so they have to meet the upper range of the radio ommuniation requirements. In the link layer of transmission, protool pakets of data are onstruted so as to be well suited to meeting the onstraints of half-duplex modems, limited bandwidth and variable 94
Volume 14 HYDROCOUTIC quality of servie. In a partiularly tough hannel, reliability is enhaned through the implementation of seletive automati repeat requests tehnique (seletive RQ). Based on the underwater ommuniation system requirements, measured hannel harateristis and W-U hannel model, OFDM parameters are determined as follows [3]: 1. Total bandwidth of OFDM signal should over the entire range of available frequenies for data transmission, determined by the bandwidth of transduer. 2. The orthogonality of arrier frequenies is provided by the IFFT algorithm applied in UC system transeiver. 3. The oherene bandwidth f sets an upper limit on a single OFDM subhannel bandwidth (the signaling rate W ). The bandwidth of single OFDM hannel should be muh larger than the Doppler spread f : m f W (1) f m 4. The rms exess delay rms (or maximum exess delay max ) sets an upper limit on the OFDM symbol duration. lso, the OFDM symbol should be muh longer than the hannel oherene time t : t (2) T 5. Extending the OFDM symbol by using the yli prefix, whih is a kind of redundany, is assoiated with loss of signal to noise ratio. It is desirable to redue to a minimum the throughput loss of NR; therefore a symbol period five to six times larger than the yli prefix period should be hosen, whih results in less than 1-dB loss in throughput [11]. 4. HNDKE PROTOCOL FOR UC OFDM YTEM Data transmission is followed by a handshake onsisting of two phases (Fig. 4). During the first phase the hannel impulse response is measured using the ML sequene. Good synhronization and orrelation properties of the reeived ML signal allow the system reipient to detet the beginning of signal transmission and measure the time-varying impulse response of the hannel. Measured impulse response is analyzed and the funtions and parameters of W-U hannel model are estimated. Next, estimated parameters are modulated by spread spetrum tehnique by onvolving data with ML sequene, and sent through the ontrol hannel to the sender of ommuniation system. During the handshake seond phase the sender alulates OFDM modulation parameters based on reeived information about W-U model estimates. Next, the testing signal is formed and sequene of OFDM symbols, followed by synhronization ML sequene, is sent to the reipient. Based on reeived testing signal the bit error rate for eah OFDM subhannel is alulated and the deision is made if BER for given subhannel is aeptable. Information about useful subhannels indies, modulated with spread spetrum tehnique, is sent bak to the sender. Data transmission is performed only at the subhannels indiated by the reipient as useful. Eah paket of OFDM symbols is followed by a synhronization ML sequene. few of the OFDM subhannels are used as pilot tones to ompensate for phase distortions, aused by the hannel influene and the differene between sampling frequenies of the sender and the reipient sound devies. In the ase of partiularly tough ommuniation onditions, it may be neessary to apply a RQ (seletive RQ) tehnique [12] illustrated in Fig. 5. When the transmission of OFDM 95 rms
HYDROCOUTIC Volume 14 data pakets proeeds, the BER is alulated for every paket and if the errors are unorretable by orretion oding, the OFDM paket should be transmitted again. n RQ paket is issued inluding indexes of pakets to be retransmitted. ender retransmits Q pakets speified by the RQ. If the number Q of inorret pakets is large, it will be neessary to return to the handshake phase and re-measure the transmission properties of the hannel. sender Impulse response measurement ML sequene reipient Impulse response measurement dapting OFDM modulation to hannel statistis Channel statistial parameters Calulating hannel statistial harateristss and parameters Testing OFDM subhannels OFDM testing signal Calulating BER in OFDM subhannels Forming data OFDM signal OFDM data transmission OFDM subhannels indexes OFDM data signal eleting OFDM subhannels for data transmission OFDM data reeption Fig. 4. Handshake of UC system laboratory model based on OFDM tehnique if Q > Q th W-U parameters ML sequene OFDM subarriers RQ OFDM testing signal N OFDM pakets Q OFDM pakets hannel impulse resposne measurement testing OFDM subhannels OFDM transmission OFDM retransmission D T H N D H K E T R N M I I O N Fig. 5. RQ mehanism for reliable data transmission 96
Volume 14 HYDROCOUTIC 5. LBORTORY MODEL OF OFDM UC YTEM The adaptive OFDM tehnique is implemented in a laboratory model of underwater ommuniation system. (Fig. 6). Both system sender and reipient onsist of portable omputer, aousti amplifier, eletroaousti transmitting transduer and hydroaousti reeiving transduer with preamplifier. amplifier Data transmission hannel preamplifier preamplifier Control hannel amplifier Fig. 6. Laboratory model of underwater OFDM ommuniation system Matlab omputing environment was hosen for software development of the data transmission system. Matlab environment meets the projet requirements. It performs matrix proessing and storing large amount of data, whih in the laboratory model orresponds to even a few minutes of sound signal sampled at 44.1 khz. The transmission is ontrolled with Data quisition Toolbox a set of development tools enabling Matlab support of the analog input and output of a omputer sound devie. Using the software interrupt system, offered by Data quisition Toolbox, real-time transmission in half-duplex mode is performed. 6. CONCLUION Depending on the lass of hannels, determined by the geometri situation, the use of different link-layer protools is required as well as orretly designed signal parameters at the physial layer data level. Radio systems offer more possibilities, while the UC ones offer less. When using the W-U model and designing appropriate protools, optimum transmission speeds an be ahieved in given loal onditions. Underwater ommuniations systems are needed that would adapt to the variability of the hannel. In the ase of the most favorable onditions of a short range ommuniation in deep water, the system would fully exploit its maximum predisposed data rate. The OFDM modulation and oding sheme turns out to be very effetive in exploiting the ommuniation potential of UC systems in natural transmission onditions of UW hannel. The most promising results have been reahed up till now when the ommuniation was additionally spatially split among several transmitters and reeivers, i.e. the so-alled multiple input multiple output (MIMO) approah was superimposed on the lassial onehannel OFDM tehnique. REFERENCE [1] Yong oo Cho, Jaekwon Kim, Won Young Yang, MIMO-OFDM Wireless Communiations with MTLB, Wiley 2010. 97
HYDROCOUTIC Volume 14 [2] tojanovi M., daptive Channel Estimation for Underwater ousti MIMO OFDM ystems, in Pro. IEEE DP/PE Workshop 2009, Maro Island, FL, January 2009. [3] Lasota H., Koha ska I., Transmission Parameters of Underwater Communiation Channels, Hydroaoustis, 2011, Vol. 14. [4] http://wapedia.mobi/pl/ofdm. [5] Prasad R., OFDM for Wireless Communiations ystems, rteh House, 2004, pp 24 25. [6] http://www.3gpp.org/lte. [7] http://www.radio-eletronis.om/info/ellularteleomms/lte-long-term-evolution/lte-ofdmofdma-sfdma.php. [8] esia., Toufik I., Baker M., The UMT Long Term Evolution: From Theory to Pratie, Wiley 2009, pp. 478 479. [9] Borowski B., Charaterization of a Very hallow Water ousti Communiation Channel, Proeedings of MT/IEEE OCEN 2009, Biloxi, Mississippi, Otober 2009. [10] Li B., Huang J., Zhou., Ball K., tojanovi M., Freitag L., Willett P., MIMO-OFDM for High Rate Underwater ousti Communiations, IEEE Journal of Oeani Engineering, Vol. 34, No 4, (2009), 634 645. [11] klar B., Digital Communiations: Fundamentals and ppliations (2nd Edition), Prentie-Hall, 2001, pp. 944 996. [12] Rie J., eaweb aousti ommuniation and navigation networks, Proeedings of the International Conferene Underwater ousti Measurements: Tehnologies & Results, Heraklion, Crete, Greee, 2005. 98