Creating and sharing knowledge for telecommunications

Adaptive Equalization Techniques with Imperfect Channel Estimates for Future 5G Systems

Rui Pedro Nóbrega Chanoca
Supervisors: Adão Paulo Soares da Silva, Pedro Miguel Ferreira de Oliveira Pedrosa
University: Universidade de Aveiro
Keywords: OFDM, OFDMA, SC-FDMA, MIMO, -DFE Abstract Wireless communication networks have been continuously experiencing an exponential growth since their inception. The overwhelming demand for high data rates, support of a large number of users while mitigating disruptive interference are the constant research focus and it has led to the creation of new technologies and efficient techniques. Orthogonal frequency division multiplexing (OFDM) is the most common example of a technology that has come to the fore in this past decade as it provided a simple and generally ideal platform for wireless data transmission. It’s drawback of a rather high peak-to-average power ratio (PAPR) and sensitivity to phase noise, which in turn led to the adoption of alternative techniques, such as the single carrier systems with frequency domain equalization (SC-FDE) or the multi carrier systems with code division multiple access (-CDMA), but the nonlinear Frequency Domain Equalizers (FDE) have been of special note due to their improved performance. From these, the Iterative Block Decision Feedback Equalizer (-DFE) has proven itself especially promising due to its compatibility with space diversity, MIMO systems and CDMA schemes. However, the -DFE requires the system to have constant knowledge of the communication channel properties, that is, to have constantly perfect Channel State Information (CSI), which is both unrealistic and impractical to implement. In this dissertation we shall design an altered -DFE receiver that is able to properly detect signals from SC-FDMA based transmitters, even with constantly erroneous channel states. The results shall demonstrate that the proposed equalization scheme is robust to imperfect CSI (-CSI) situations, since its performance is constantly close to the perfect CSI case, within just a few iterations.
Abstract: PDF download