Implementation and performance evaluation of scaled precision model for ofdm synchronization

Abstract

High spectral efficiency and its robustness to inter symbol interference behavior of the frequency selective wireless channels in any multi-users environment orthogonal frequency division multiple access (OFDM) has emerged as the most dominant wireless transmission technique for high speed wireless systems. The generic performance metrics of OFDM system include both the estimation techniques used and arithmetic model used for FFT computation. Accurate offset estimation is a major concern in OFDM system. Even though there are numerous methodologies available to estimate the offset values still OFDM synchronization face certain problems with erroneous fixed point model arithmetic. Due to word length constraints accurate offset estimation using fixed point FFT model is a challenging task since finite arithmetic model always cause computational error and hence significant performance loss over vehicular channels. Since OFDM is highly vulnerable to channel impairments several data aided and non-data-aided synchronization are investigated in many works to provide complementary solutions to synchronization related issues. In general non-data-aided synchronization makes use of the cyclic property of the guard interval which is less efficient and reduced the accuracy level with moderate channel length. On the other side data-aided synchronization accomplished this task with the aid of well known training symbols called pilot symbols which gives considerable quality improvement and the feasibility of estimation process in OFDM systems. In these work 802.11a standard preambles are used as training symbols which consists of ten identical short training symbols (t1, t2..., and t10) and two identical long training symbols. Frame detection is carried out with peak cross correlation function to estimate the offset values newline