SYNCHRONIZATION AND CHANNEL ESTIMATION TECHNIQUE ON MIMO-GFDM SYSTEM

• Main Author: Vincent
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/70727
• Institution: Institut Teknologi Bandung

The discrete Fourier transform (DFT)-based channel estimation has been shown to improve least square (LS) channel estimation in generalized frequency division multiplexing (GFDM) system. However, the evaluation assumed perfect channel impulse response (CIR) length knowledge on the receiver. Consequently, the improvement offered by DFT-based channel estimation cannot be extracted properly. Generally, a receiver does not have any knowledge on CIR length, thus an estimation must be done. A threshold can be constructed using several samples from LS channel estimation to set channel and undesired components apart. The designed adaptive scale could improve the CIR length estimation by adjusting the threshold based on the signal-to-interference-ratio (SINR) of the signal. To accommodate its implementation, a sparse-preamble is designed to allow the receiver to estimate the SINR. The exponentially-modeled adaptive scale with convergence rate controller of 0.05 is found to fit best the urban and rural environment, and the estimated CIR length could realize DFT-based channel estimation. The result is identical for urban environment, and it requires 0.5 dB more power to achieve same symbol error rate (SER) level with ideal condition at 10-4 . In addition, the designed sparse-preamble requires 5.5 dB less power than conventional preamble to achive MSE of STO at 10-1 at the cost of saturation of carrier frequency offset (CFO) estimation at -50 dB for signal-to-noise ratio (SNR) above 20 dB. Furthermore, the proposed methods can be easily implemented in multiple-input multiple-output (MIMO)-GFDM system, and it provides improvement for the reliability of the system for both environments.