Performance Analysis for LTE Wireless Communication

Transcription

1 IOP Conference Series: Materials Science and Engineering PAPER OPEN ACCESS Performance Analysis for LTE Wireless Communication To cite tis article: S Tolat and T C Tiong 2015 IOP Conf. Ser.: Mater. Sci. Eng Related content - Adaptive Modulation and Coding for LTE Wireless Communication S S Hadi and T C Tiong - TEMPERATURE REVERSAL IN A LTE ATMOSPHERE Stepen J. Hill - LTE turbo decoder design Yang Le, Ye Tiancun, Wu Bin et al. View te article online for updates and enancements. Tis content was downloaded from IP address on 24/02/2018 at 10:42

2 Performance Analysis for LTE Wireless Communication S Tolat and T C Tiong Department of Electrical and Computer Engineering, Faculty of Engineering and Science, Curtin University, Miri, Sarawak, Malaysia tiong.teck.cai@curtin.edu.my Abstract. Long Term Evolution (LTE) is te new upgrade pat for carrier wit bot GSM/UMTS networks and CDMA2000 networks. Te LTE is targeting to become te first global mobile pone standard regardless of te different LTE frequencies and bands use in oter countries barrier. Adaptive Modulation and Coding (AMC) is used to increase te network capacity or downlink data rates. Various modulation types are discussed suc as Quadrature Pase Sift Keying (QPSK) and Quadrature Amplitude Modulation (QAM). Spatial multiplexing tecniques for 4 4 MIMO antenna configurations are considered. Tis paper as outlined various estimation tecniques to increase te trougput of te LTE network by simulating te estimation tecniques wit various parameters in te LTE downlink mode 4 (spatial multiplexing). Tree tecniques i.e. cannel estimation tecnique, estimation of cannel models and MIMO receiver algoritm are simulated to provide te ideal LTE wireless communication system. 1. Introduction Te increase in broad band multi-media services including voice over internet protocol (VoIP), mobile TV, audio and video messaging and internet access demand a larger wireless cellular network wit better reliability. In addition, telecommunication operators are constrained wit limited frequency spectrums purcased from te government. Tus te need arises to searc for a feasible network wic can increase te trougput witout purcasing additional bandwidt and at providing good service quality. Te Long Term Evolution (LTE) standard, specified by te 3rd Generation Partnersip Project (3GPP) in Release 8, defines te next evolutionary step in 3G tecnology. LTE offers significant improvements over previous tecnologies suc as Universal Mobile Telecommunications System (UMTS) and Hig-Speed Packet Access (HSPA) by introducing a novel pysical layer and reforming te core network. Te main reasons for tese canges in te Radio Access Network (RAN) system design are te need to provide iger spectral efficiency, lower delay, and more multi-user flexibility tan te currently deployed networks [1]. In 2008 te International Telecommunication Union specified requirements for 4G standards and named it te International Mobile Telecommunication Advanced wic was expected to provide transmission rates of up to 1 Gbps. Tis paper analyzed LTE mode 4 (Closed-Loop Spatial Multiplexing) configurations of cannel estimation tecniques, cannel models and multiple-input multiple-output (MIMO) receiver algoritms gives a lower bit error rate (BER) as to compreend tese needs tat ave arisen for te telecommunication operators [2]-[4]. Content from tis work may be used under te terms of te Creative Commons Attribution 3.0 licence. Any furter distribution of tis work must maintain attribution to te autor(s) and te title of te work, journal citation and DOI. Publised under licence by Ltd 1

3 2. MIMO MIMO use multiple antennas for bot transmitting and receiving antennas. Te advantages of MIMO communication, wic exploits te pysical cannel between many transmit and receive antennas, are currently receiving significant attention [5]-[6]. Figure 1 sows a MIMO system consists of n receive antennas and m transmit antennas. Te connections from eac antenna are designated as H n n2 1 m 2m nm (1) Figure 1. MIMO system. Denote te receive vector y, transmit vector x and noise n. We ave y Hx n (2) 2.1. Cannel modes Flat fading. A transmitted signal undergoes flat fading, or sometimes called as narrowband cannels, wen te signal bandwidt (BW) is less tan te cannel coerence bandwidt. In flat fading, toug te spectrum of te transmitted signal is preserved, te received signal varies in amplitude and encounters deep fades of 20 to 30 db. Instantaneous amplitudes of tese received signals follow Rayleig distribution Frequency selective. For Frequency selective fading or wide band cannel, te signal bandwidt is greater tan te cannel coerence bandwidt. In frequency selective fading, time dispersion causes inters-symbol interference at te receiver making te multiple versions of te received signal attenuate and delay in time to a different degree LTE specific cannels. Tere are tree different multipat fading cannels for LTE specific cannels made by te Tird Generation Partnersip Project (3GPP). Namely, Extended Pedestrian A (EPA) 2

4 Extended Veicular A (EVA) Extended Typical Urban (ETA) 2.2. MIMO receivers Receiver use equalizer to produce te best estimate of te transmitted symbols. Tree algoritms are considered i.e. Zero Forcing (ZF) equalizer Minimum Mean Square Error (MMSE) equalizer Spere Decoder (SD) equalizer ZF equaliser. Te ZF receiver inverses te MIMO cannel matrix. Te major drawback of te ZF receiver is noise enancement. Altoug te ZF receiver eliminates te interference, performance becomes poor wen te cannel of te signal of interest is almost collinear to te interference subspace, or in oter words wen te cannel matrix is almost rank deficient MMSE equaliser. Te MMSE receiver minimises te average estimation error on te transmitted symbols. Te average is taken over te transmitted symbols and te noise : te mean square error (MSE) is E xˆ x, n x 2, were xˆ is estimated symbol and x is te transmitted symbol. Te ZF receiver also minimises te output MSE but under te constraint of complete inter symbol interference (ISI) elimination SD equaliser. SD receiver finds te transmitted signal vector wit minimum maximum likeliood (ML) metric, tat is, to find te ML solution vector. However it considers only a small set of vectors witin a given spere rater tan all possible transmitted signal vectors. SD adjusts te spere radius until tere exists a single vector (ML solution vector) witin a spere. It increases te radius wen tere exists no vector witin a spere, and decreases te radius wen tere exists multiple vectors witin te spere. 3. OFDM Ortogonal Frequency Division Multiplexing (OFDM) is a radical and revolutionary tecnology as it forms te basis for 4G tecnology. Te basic concept of OFDM is to send samples concurrently using multiple ortogonal sub-cannels rater tan sending a sample signal using te entire bandwidt. Te main advantage of OFDM is tat because of te use of many sub-cannels, only some of te subcannels are affected by interference or multi-pat effect [7]-[8] OFDM In OFDM a single cannel utilizes multiple overlapping sub-carriers as sown in Figure 2. However, because tese overlapping sub-carriers are ortogonal to eac oter, te sub-carriers will not interfere wit eac oter. OFDM system can maximize spectral efficiencies. Figure 2. OFDM cannel 3

5 3.2. Cyclic prefix To completely eliminate te Inter Symbol Interference (ISI), cyclic prefix (CP) is introduced into te OFDM. CP is a repetition of te last samples of te data portion tat is appended at te beginning of te data payload. As long as te CP duration is longer tan te cannel delay spread te ISI will be completely eliminated. 4. LTE pysical layer Designed to meet te present-day needs for ig-speed media and data transfer, te 3GPP Long Term Evolution, or more commonly known LTE, symbolizes a major advance in cellular tecnology. It also includes multimedia unicast and broadcast services in addition to ig-capacity voice support. Te LTE pysical layer (PHY) acts as a igly efficient means for transmission of bot data and control information between mobile user equipment (UE) and an enanced base station (enodeb). Te tecnologies employed by te LTE PHY include OFDM and MIMO Furtermore, it uses Ortogonal Frequency Division Multiple Access (OFDMA) on te downlink (DL) and Single Carrier Frequency Division Multiple Access (SC-FDMA) on te uplink (UL). Te function of OFDMA is to direct data to or from multiple users for a number of symbolic periods on a subcarrier-by-subcarrier basis Transmission Modes in LTE downlink Mode 1 (SIMO). Tere is a single transmit antenna and multiple receiver antennas. Tus Mode 1 of LTE as only one transmit antenna Mode 2 (transmit diversity). Te default MIMO mode, transmit diversity, sends te same data via various antennas. As Eac antenna stream is using different frequency resources and different coding it provides a stronger transmission and causes te SNR to improve. Transmit diversity is used as a fall-back option in LTE for some transmission modes. For example, wen spatial multiplexing (SM) cannot be applied, Control cannels, suc as pysical broadcast cannel (PBCH) and pysical downlink control cannel (PDCCH), is also transmitted using transmit diversity. A frequency-based version of te Alamouti codes space frequency block code (SFBC) is used for two antennas; wereas a combination of SFBC and frequency switced transmit diversity (FSTD) is used in te case of four antennas Mode 3 (open-loop spatial multiplexing). To attain iger date rates, tis mode supports spatial multiplexing of two to four layers tat are multiplexed to two to four antennas, respectively. Less user equipment (UE) feedback is required for it regarding te cannel situation as precoding matrix indicator is excluded. It is used wen cannel rapidly canges or wen tere is missing cannel information. An example of tis can be UEs moving wit ig velocity. Furtermore, tere is a specific delay in supplying te signal to every antenna known as te cyclic delay diversity ( CDD). Tis in turn artificially creates a frequency diversity Mode 4 (closed-loop spatial multiplexing). Tis mode supports spatial multiplexing wit up to four layers tat are multiplexed to up to four antennas respectively. Tis is in order to attain iger data rates. To allow cannel estimation at te receiver, cell-specific reference signals (RS) is transmitted by te base station and distributed over various resource elements (RE) and timeslots. A response regarding te cannel situation is ten sent by te UE. Tis includes information about wic precoding is ideal from te defined codebook. 4

6 5. Performance Analysis 5.1. Examining effects for cannel estimation tecniques Te cannel estimation tecniques are Ideal Cannel Estimation Cannel Estimation on interpolation Cannel estimation based on averaging over eac slot Cannel estimation based on averaging over eac subframe Figure 3. BER vs SNR for cannel estimation tecniques Table 1. Cannel estimation tecnique parameters. Parameter Value Transmission Mode 4 Number of transmit antenna 2 Number of receiver antenna 2 Cannel Bandwidt 10 MHz Modulation Sceme 16QAM Cannel Model frequency selective Figure 3 sows te ideal cannel estimation tecnique provides te best BER performance. As te variations witin slots and subframes are allowed by interpolation it as a iger performance in an MMSE error-minimization context. Toug te averaging metods don t improve te BER performance muc it provides a continuity needed for better perceptual results. Te best cannel estimation is Ideal cannel Estimation, followed by Estimation based on Interpolation and ten averaging over eac slot and finally averaging over eac sub frame. 5

7 5.2. Examining effects of cannel models Table 2. Effects of cannel model parameters. Parameter Value Transmission Mode 4 Number of transmit antenna 2 Number of receiver antenna 2 Cannel Bandwidt 10 MHz Modulation Sceme 16QAM Cannel estimation tecnique Ideal cannel estimation MIMO Receiver MMSE Figure 4. BER vs SNR for low correlated cannels Figure 4 sows tere is not muc difference in all te graps until around SNR 7dB. At SNR 12 db, As suc te best to worst type of cannel models can be listed as follows from te grap; EVA of a maximum Doppler sift of 5 Hz EVA of a maximum Doppler sift of 70 Hz EPA of a maximum Doppler sift of 5 Hz Flat Fading cannel of a Maximum Doppler sift of 0 Hz Frequency Selective Fading cannel of a maximum Doppler sift of 0 Hz As noise profile increase, te performance becomes consistently reduced. Higer mobility as indicated by iger Doppler sift also reduce performance. Above 15 db, Flat Fading cannel of a Maximum Doppler sift of 0 Hz outperform all oters. 6

8 5.3. Examining effects of MIMO receiver algoritms To examine te effects of MIMO receiver Algoritms, te MIMO receiver algoritms are compared i.e. ZF, MMSE and SD. Table 3. MIMO receiver algoritm parameters. Parameter Transmission Mode 4 Number of transmit antenna 2 Number of receiver antenna 2 Cannel Bandwidt Modulation Sceme Value 10 MHz 16QAM Cannel estimation tecnique Ideal cannel estimation As te ideal cannel estimation tecnique is te best estimation tecnique, Ideal cannel estimation tecnique is used as a simulation parameter for cannel estimation tecnique in tis simulation. Figure 5 sows tat SD as te best performance followed by MMSE and ZF. Figure 5. MIMO receiver algoritms comparison. Toug te SD outperforms MMSE and ZF, SD as iger computational complexity. Terefore Trade-off between performance and complexity decides te coice between SD and MMSE. 6. Conclusion LTE was introduced to meet te demands of te users to download and upload packet data faster tan its predecessor 3G. Tis paper as outlined various estimation tecniques to increase te trougput of te LTE network by simulating te estimation tecniques wit various parameters in te LTE downlink Mode 4 (closed-loop spatial multiplexing). 7

9 Tree tecniques are used in tis paper, cannel estimation tecnique, estimation of cannel models and MIMO receiver algoritm simulations. LTE pysical layer is simulated to select ideal combination. Te ideal combination is Ideal Cannel Estimation Tecnique, Low correlated EVA wit low Doppler sift and SD. References [1] Dalman E, Parkvall S, Skold J and Beming P Jul G Evolution: HSDPA and LTE for Mobile Broadband (Academic Press) [2] Foscini G J 1996 Layered space-time arcitecture for wire-less communication in a fading environment wen using multi-element antennas Bell Labs Tec. J. 1(2) pp [3] Telatar I E 1999 Capacity of multi-antenna gaussian cannels Eur. Trans. Telecommun.10 (6) pp [4] Bliss D W, Forsyte K W, Hero A O and Swindle-urst A L 29 Oct. 1 Nov MIMO Environmental Capacity Sensitivity Tirty-Fourt Asilomar Conf. on Signals, Systems & Computers vol. 1 pp [5] Bliss D W, Forsyte K W and Yegulalp A F 4 7 Nov MIMO communication capacity using infinite dimension ran - dom matrix eigenvalue distributions Tirty-Fift Asilomar Conf. on Signals, Systems & Computers vol 2 pp [6] Gos A and Ratasuk R 2011 Essentials of LTE and LTE-A Cambridge University Press [7] Sidrum B M, Mitun T P and Madukar M 2013 BER, SNR and PAPR analysis for multiple accesses in LTE Int. J. of Scientific & Engineering Researc vol 4 Issue 7 pp [8] Kebede G M and Paul O O 2010 Performance Evaluation of LTE Downlink wit MIMO Tecniques Blekinge Institute of Tecnology 8