PAPR Reduction in Filtered OFDM using Companding Techniques

Abstract

Filtered Orthogonal Frequency Division Multiplexing is a strong contender waveform for fifth generation communications. In F-OFDM, total allocated bandwidth is riven into numerous sub-bands and diverse kinds of services are engaged in these sub-bands by using appropriate waveform and numerology, resulting in better spectrum utilization. F-OFDM avoids the restrictions of OFDM and also achieves the novel challenges encountered in 5G waveform. Just like OFDM, F-OFDM suffer from a high PAPR, which requires the usage of HPA with extended dynamic range at the transmitter, which increases both cost and power. There are many reduction techniques that lessen PAPR, but at the cost of degradation in BER. Partial Transmission Sequence (PTS) and Selective Matching (SLM) methods are widely used to decrease PAPR, but these two algorithms are complex and inefficient, because secondary information is required to be transmitted to receive the information at the receiver. Clipping and filtering are simplest of all reduction methods, but they involve signal distortion that causes data loss. Tone injection and tone reservation techniques are less complex but due to some data loss, the degradation of BER is even greater. The precoding method is easy to implement, but there is a certain data rate loss and also computationally complex. In this thesis significant companding techniques have been considered for PAPR reduction in FOFDM. Both A-law and µ-law companding techniques have proven decent reduction in PAPR however their BER degradation is high. Exponential companding has the maximum BER degradation among the companding techniques mentioned in this thesis however has appropriate PAPR reduction.