Compressive sensing based power signal compression in advanced metering infrastructure

In this paper, we consider a compressive sensing (CS) based compression method for aggregated power signals of smart meters in advanced metering infrastructure (AMI). If detailed power signals with high sampling rates are available for energy management systems, advanced services such as demand response and power disaggregation can be provided. However, typical smart meters have limited bandwidth resources and may require aggressive data compression for supporting the intensively sampled data. In order to alleviate the overhead for the detailed power signal, a CS based compression scheme is proposed in this study to take advantage of the sparsity of power signals based on the appliance load superposition modeling. Through simulations, the CS based compression schemes demonstrate improved energy efficiency while maintaining good reconstruction performance, compared to the wavelet based compression.