Time-domain Complexity Analysis of Impulse Noise Sources for xDSL/PLC Systems

The main objective of this paper is to study the complexity of time domain patterns of impulse noises present in wireline (indoor) communication systems. We have explored usefulness of Second Order Difference Plots (SODP) images and an information theory based complexity measure, i.e., approximate entropy to quantify the time-domain complexity of impulse noises. The attributes derived from these measures are capable of distinguishing between deterministic systems, periodic/aperiodic systems, stochastic and mixed systems. We have demonstrated in this paper that these attributes are able to capture and quantify the variability of impulse noises generated by electronic appliances and thus may aid in identifying the primary disturber, i.e., the active impulse noise source.