LAUSR

One of the challenges faced by indoor power line communication systems is the frequency-selective channels that are time varying and dependent on a large number of variabilities. A probable solution is to extract as much determinism as possible through prediction and statistical analysis of the frequency-selective notches. However, such deterministic tools are available only for simple open- and short-circuit branches and not for complex loads and topologies. This paper proposes an alternate method to predict and analyze notches using a minimum of four parameters without evaluating the transfer function. Termed as the load frequency mapping, the method is applicable for any frequency-dependent time-invariant loads and topologies and also capable of performing statistical analysis of random channels. A high decrease in prediction error (99.46–93.63%) is found for all the channels analyzed. Statistical analysis of 26 power line cables using random loads shows that some cables and loads offer more variations in frequency selectivity than others. For capacitive loads, the variation is more for the low-frequency notches and for those modeled as parallel resonant circuits at the frequencies near the resonance. Maximum variation is found for cables with high characteristic impedance with loads having high resonant frequencies and low quality factor and least for inductive loads. The power line therefore has considerable amount of determinism, and this can be incorporated to complement for fading channels or analysis of variability optimized for dual purpose of power delivery and data transfer.