Partial-Data Superimposed Training With Data Precoding for OFDM Systems

Abstract

Superimposed training (ST) is a recently addressed technique used for channel estimation where known training sequences are arithmetically added to data symbols, avoiding the use of dedicated pilot subcarriers, and thus, increasing the available bandwidth compared with traditional pilot symbol assisted modulation schemes. However, the system handles data interference over channel estimation as a result of the ST process; also, data detection is degraded by pilot interference. Recent ST methods have analyzed the data interference and presented schemes that deal with it. We propose a novel superimposed model over a precoded data scheme, named partial-data superimposed training (PDST), where an interference control factor assigns the adequate information level to be added to the training sequence in orthogonal frequency division multiplexing systems. Also, a data detection method is introduced to improve the symbol error rate performance. Moreover, a capacity analysis of the system has been derived. Finally, simulation results confirm that performance of PDST is superior to previous proposals.