Abstract-Energy disaggregation is the task of estimating power consumption of each individual appliance from the whole-house electric signals. In this paper, we study this task based on deep learning methods which have achieved a lot of success in various domains recently. We introduce the feature extraction method that uses multiple parallel convolutional layers of varying filter sizes and present an LSTM (Long Short Term Memory) based recurrent network model as well as an auto-encoder model for energy disaggregation. Then we evaluate the proposed methods using the largest dataset available. And experimental results show the superiority of our feature extraction method and the LSTM based model.

No abstract

No abstract

The application of Internet of Things in smart buildings is becoming more and more extensive. The energy supply problem of end nodes has always been a concern of the field of research and industry areas, and an adaptive radio frequency energy conversion system is proposed for the low power node of the Internet of Things. By using the unipolar transistor control, the threshold voltage of the rectifier changes dynamically, through RF matching, and double voltage rectification measures the supply for the low power nodes. The proposed circuit is aimed at reducing the threshold voltage of the transistor forward bias to increase power and output voltage of the harvester and increasing the reverse bias of the threshold voltage for reducing the leakage current. It prevents the loss of stored energy. This paper presents a 12 stages adaptive thresholdcompensated rectifier circuit has reached the maximum power conversion efficiency 35.3% at −14 dBm. Meanwhile, when the input power is −12dBm, the output voltage is 3.2 V at the output load 1 M. The low power characteristics of the nodes at the end of the Internet of Things make it possible to convert the RF energy into electric energy. INDEX TERMS Low power node, smart building, radio frequency energy harvester, energy conversion efficiency.