COMPRESSIVE SAMPLING FOR NON-INTRUSIVE APPLIANCE LOAD MONITORING (NALM) USING CURRENT WAVEFORMS

We propose a NALM technique by exploiting the compres-sive sampling and sparse reconstruction framework. We estimate the contribution of the individual appliances by measuring the current of the total load. We further assume to know the steady-state current waveform of each appli-ance. We exploit the sparsity of the current signal to com-press the measurement via random sampling, which lowers signicantly the processing complexity, the storage and the communication burden. Using the proposed sparse recon-struction approach, we can still identify the on/off status of each appliance from the compressed measurement as if the original non-compressed measurement is used