International Journal of Computer Applications (0975 - 8887) National Conference on Advances in Communication and Computing (NCACC-2014) SVD -Low Complexity Channel Estimation in IEEE 802.16e-DL-PUSC System Mehul V. Desai P.G Student Electronics Department SVNIT Surat Gujarat-395007 Mrs.Shilpi Gupta Assistant Professor Electronics Department SVNIT Surat Gujarat-395007 Upena D. Dalal, Ph.D Associate Professor Electronics Department SVNIT Surat Gujarat-395007 ABSTRACT 802.16e system support high data rate transfer but at the same time user has high mobility and due to this effect,the channel response may vary rapidly.Also pilot signals are usually lim- ited in comb type arrangement to perform interpolation. So to perform channel estimation and interpolation with low complex- ity in high mobility scenario is the challenging problem. In this paper channel estimation of system is carried out by finding Channel Impulse Response (CIR) of pilot sub-carrier using LS, LMMSE and SVD algorithms and then finding Channel Fre- quency Response (CFR) at data sub-carrier is done by time and frequency Interpolation of Pilot CIR. This paper presents BER and MSE performance for 16QAM and 64QAM Coded OFDM System(C-OFDM) evaluated on standard channel ITU-A and ITU- B. Results show that by using low complexity SVD algorithm gives better performance than LS and LMMSE at lower SNR. General Terms: Orthogonal Frequency Division multiplexing(OFDM),Channel Estima- tion,Mobile WiMAX(802.16e) Keywords: Coded OFDM System (C-OFDM), Channel Estimation, LS (Least square Estimation), LMMSE (Linear minimum mean square error), SVD (Singular Value Decomposition), Channel Impulse Response (CIR), Channel Frequency Response (CFR), Down-link Partially Used Sub channelization (DL-PUSC) 1. INTRODUCTION Channel estimation is one of challenging problems in IEEE 802.16e Orthogonal Frequency Division Multiplexing Access (OFDMA) down-link system.[1] In order to facilitate the correct detection of transmitted symbols, the effects of the channel must be removed from the received signal. This is the task of equalizer block which perform normalization of the signal received by each sub-carrier with its channel transfer function.[2] Channel transfer estimates are generated within the channel estimation block. The approach for OFDMA mobile WiMAX system channel estimation is pilot assisted or training symbol based channel estimation which employ pilot symbols known both to the receiver and the transmit- ters [3] OFDM is key technology of standard (802.16e) Mobile WiMAX is based on orthogonality principle support to multicarrier trans- mission technique which is built by many orthogonal carriers that transmits simultaneously[1][3]. Idea behind OFDM is that to di- vide the transmitted bit stream in to many different sub-stream and send these over many sub channels, also the number of sub-stream is chosen to ensure that each sub channel has a bandwidth less than the coherence bandwidth of the channel, so that sub channels ex- perience relatively flat fading and due to this Inter Symbol Inter- ference (ISI) on each sub-channel is small. And the remaining ISI effect is eliminated by using cyclic prefix. [4] The estimated Channel Frequency Response (CFR) at pilot sub- carriers can be obtained by the Least Squares (LS) or the Min- imum Mean Square Error (MMSE) criteria [5][6]. However, the computational complexity of MMSE is too high and requires the knowledge of channel information and operating SNR, these make MMSE difficult in practical use. Linear Minimum Mean Square Er- ror (LMMSE) has a very low computational complexity compare to MMSE algorithm by taking some known values: one is β which depends upon different modulation; another being R HpHp , which is the auto correlation matrix of the channel. A low rank approxi- mation is applied to a LMMSE estimator which uses the frequency correlation of the channel. By using the Singular-Value Decom- position (SVD) an optimal low-rank estimator is derived, where performance is essentially preserved - even for low computational complexities[7][8]. In this paper CIR has been obtained by exploiting pilot by means channel estimation algorithms and then CFR can be obtained by exploiting pilot at symbol data with two dimension interpo- lation scheme called time and frequency interpolation in DL- PUSC.[9] Based on simulations over fast- and slow-fading mobile channels(ITU-R-A ,ITU-R-B), it has been observed that significant improvement in SNR with lowest computational complexity under SVD algorithm in terms of Bit Error Rate and Mean Square Er- ror performance for 16QAM and 64QAM systems which has been simulated on MATLAB. The paper is organized as follows. Section 2 explains the system description; section 3 discusses channel estimation method, with pilot based channel estimation method by LS, LMMSE, and SVD with Linear Interpolation. Results have been presented in section 4, which shows the effects of lower complexity and efficient algo- 1