Demonstration of a Novel SLM-Based PAPR Reduction Method for GFDM Using USRP Testbed

The remarkable attributes like elevated transmission rates, low latency, minimal out band emission and enhanced spectral efficiency of the generalized frequency division multiplexing (GFDM) system have lured the significant attention from research community. However, the inherent superposition of multiple symbols within the time domain results in a notable amount of peak-to-average power ratio (PAPR) that affects the performance due to non-linear front end circuitry. The selected mapping (SLM) technique resolves this problem by employing phase rotation vectors to generate alternative signal representations. Conventionally, phase sequence vectors possess uniform magnitudes, thereby constraining the achievable PAPR reduction. To address this limitation, we propose the use of normalized rows from the Riemann matrix as phase sequence vectors, given their inclusion of components with substantial magnitudes. This approach mitigates excessive peaks and consequently yields superior PAPR reduction gains. To optimize real-time system performance, side information (SI) is also integrated into the transmitted signals alongside the normalized Riemann matrix sequences. Indeed, for validation purpose, an experimental setup utilizing the universal software radio peripheral (USRP) 2953R in conjunction with LabVIEW software is employed for real-time transmission. The observed spectral response from the USRP closely aligned with simulation results, affirming the efficacy of the proposed method.