A novel blind STO estimation for UW-GFDM systems

• Main Authors: valluri, Siva Prasad, Mani, V. V.
• Format: Conference Proceeding
• Language: English
• Subjects: Correlation; Encoding; Estimation; Frequency-domain analysis; Indexes; Internet of Things
• ISSN: 2162-1039
• DOI: 10.1109/ATC.2018.8587431

5G applications like internet of things (IoT) has attracted attention of all researchers with spectacular characteristics like massive connectivity and lightly loaded packets. Synchronization is a perturbing barrier for these systems, because their small packet sizes cannot accommodate room required for training symbols. Recently the usage of deterministic sequence, named as unique word (UW), instead of well known cyclic prefix (CP) is introduced for one of the promising contenders of 5G physical layer named as generalized frequency division multiplexing (GFDM) system. The primary motive behind introduction of UW is to exploit the correlations in frequency domain which results in coding gain. However, perfect symbol time offset (STO) compensation is a prerequisite and is the first task to be done. This article presents a blind metric for the estimation of STO using the structure of UW-GFDM system without requirement of pilot symbols present in UW positions. The developed algorithm can blindly estimate the block duration, symbol duration, number of redundant subcarriers and number of subcarriers. The obtained simulation results verify the feasibility of the proposed algorithm in various channel conditions.