Abstract
This dissertation presents the design and experimental evaluation of a software-defined radio
(SDR) based M-ary Phase Shift Keying (M-PSK) transceiver for indoor narrowband power line
communication (NB-PLC) systems. Impulsive noise is a common phenomenon in residential
and industrial power line networks that affects the physical layer reliability, and this study will
examine its impact. The focus is on two key performance measures: bit error rate (BER) and
synchronization accuracy.
A reconfigurable transceiver was developed using MATLAB and Simulink together with Universal
Software Radio Peripheral (USRP) hardware. This setup enabled realistic hardware-inthe-
loop testing under controlled laboratory conditions. The system supports 4-PSK, 8-PSK,
and 16-PSK modulation schemes, allowing a direct comparison of spectral efficiency and noise
robustness at different signal-to-noise ratios (SNRs). Synchronization was achieved using the
Gardner non-data-aided timing recovery algorithm and a 13-bit Barker sequence for frame
detection, ensuring consistency with standard NB-PLC frame structures.
Experimental and simulation results show that impulsive noise causes both symbol errors and
frame misalignments. Synchronization failures alone accounted for up to 80 percent of total
bit errors under severe noise conditions. Theoretical and measured BER results under additive
white Gaussian noise (AWGN) agreed closely, while real impulsive environments produced
additional deviations that matched non-Gaussian noise characteristics. Higher-order modulations,
such as 16-PSK, were found to be more sensitive to timing and phase disturbances,
confirming the trade-off between spectral efficiency and system robustness.
The research contributes a validated SDR-based testbed for NB-PLC analysis, a dual-metric
framework linking BER and synchronization reliability, and practical insights for improving
synchronization under impulsive noise. The findings support the use of adaptive thresholding,
longer preambles, and lightweight error-control coding to improve the resilience of real-world
NB-PLC systems.