Experimental Validation of Wavelet-Based Smart Metering Data Compression over SDR Links

This study investigates wavelet-based compression of smart-metering data transmitted through a software-defined radio chain implemented in LabVIEW with QPSK modulation and USRP platforms. The objective is to reduce the transmitted payload while preserving the fidelity of the reconstructed electrical load profile. The work combines a mathematical formulation of the DWT-based compression and reconstruction process, a controlled scenario evaluation, and an experimental validation on an SDR testbed. The scenario analysis shows that the compression–reconstruction trade-off is best achieved in an intermediate operating region, where excessive coefficient removal increases reconstruction error despite higher nominal reduction. In the laboratory SDR campaign, Haar wavelet order 1 at the LabVIEW coefficient-retention setting 59 was selected as the most balanced representative configuration, achieving a 60.2% unit-based compression ratio, 10.61% relative error, (Formula presented.) and (Formula presented.). This selection refers to the physical SDR implementation and should not be confused with the public-dataset validation, where bior4.4 level 8 with 40% retained coefficients provided the best offline compression–reconstruction trade-off. Under the tested USRP/LabVIEW configuration, the 5 GHz setup showed shorter channel occupation time than the 915 MHz setup, with lower measured coverage in the same laboratory campaign. The additional validation using the public UCI Individual Household Electric Power Consumption dataset confirmed that DWT compression can preserve load-profile structure under substantial coefficient reduction. Overall, the results indicate that wavelet compression is technically feasible for smart-metering transmission over SDR links when the wavelet family, order, coefficient-retention setting, and radio-link operating conditions are jointly considered.