Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol.5, No.6, 2014 46 Two Channel Estimation Methods for MIMO-OFDM System D.B. Bhoyar Yeshwantrao Chavan College of Engineering,Nagpur,India E-mail: dinesh_bhoyar@rediffmail.com C.G. Dethe 2 UGC Academic Staff College RTM Nagpur University,Nagpur,India M.M. Mushrif Yeshwantrao Chavan College of Engineering,Nagpur,India Abstract Adaptive Filter is a part of the modern communication system. The applications of the adaptive filters are channel equalization, noise cancellation, system identification and adaptive beam forming. So the proper implementation of adaptive filter is a great deal. The intersymbol interference (ISI) caused by the multipath in band limited frequency selective time dispersion channel distort the transmitted signal. In this paper, we have concentrated on modifying the algorithm for the adaptive filter. The proposed VSS-LLMS and Modified Variable Step Size Leaky LMS (MVSS-LLMS) which improves the channel estimation in the noisy environment. Also we compared the results of our proposed algorithms with the LMS, RLS and VLLMS and observed that it improves in computational complexity and Bit Error Rate (BER) performance. Keywords: Adaptive Channel Estimation, Adaptive filter, LMS, RLS, VLLMS and MVSS-LLMS 1. Introduction Modern wireless communication systems require higher data rate technologies to meet the demand for high data rate services such as multimedia, VOIP etc. In order to meet this enormous demand of higher data rate and better coverage of wireless network the channel bandwidth is increased. But this approach is not practical as the frequency spectrum is very expensive, even though by increasing the more complex modulation scheme for improving the throughput it is not used due to increased complexity of radio systems and therefore its cost. For transmitting large amount of data various modulation schemes are employed. Recent approach is to employ Orthogonal Frequency Division Multiplexing (OFDM) [2-9,11]. OFDM modulation turns the frequency- selective channel into a set of parallel flat fading channels to overcome the ISI. Least mean square (LMS), Kalman filter, Normalized Least mean square (NLMS)[4] and variable step size LMS (VSSLMS) [1] are some of the numerical computational techniques that have been applied for channel estimation in wireless communication. Along with it, fuzzy logic, neural networks, simulated annealing are also some of the new techniques used for frequency estimation. The LMS algorithm is mostly used conventional algorithm for the designing of adaptive filters due to its computational simplicity, ease of implementation and unbiased convergence [10]. In this paper, we have given more emphasis on the LMS algorithm. We have modified the LMS algorithm and proposed new VSS-LLMS and Modified Variable Step Size Leaky LMS (MVSS-LLMS) algorithms for channel estimation in noisy environment. The remaining portion of this paper is organized as follows; M.M. Mushrif 3 Section II describes the generalized adaptive channel estimation method. Section III describes proposed VSS-LLMS and MVSS-LLMS algorithms. The performance analysis is represented in section IV and in section V gives the conclusion that we have derived from the results. 2. Adaptive Channel Estimation Method