Kanghang He's research while affiliated with University of Strathclyde and other places

Publications (12)

Article
Full-text available
With the growing application of undirected graphs for signal/image processing on graphs and distributed machine learning, we demonstrate that the shift-enabled condition is as necessary for undirected graphs as it is for directed graphs. It has recently been shown that, contrary to the widespread belief that a shift-enabled condition (necessary for...
Article
Full-text available
Building on recent unsupervised Non-intrusive load monitoring (NILM) algorithms that use graph Laplacian regularization (GLR) and achieve state-of-the-art performance, in this paper, we propose a novel unsupervised approach to design an underlying graph to model the correlation within time-series smart meter measurements. We propose a variable-leng...
Preprint
Full-text available
It has recently been shown that, contrary to the wide belief that a shift-enabled condition (necessary for any shift-invariant filter to be representable by a graph shift matrix) can be ignored because any non-shift-enabled matrix can be converted to a shift-enabled matrix, such a conversion in general may not hold for a directed graph with non-sym...
Preprint
Full-text available
It has recently been shown that, contrary to the wide belief that a shift-enabled condition (necessary for any shift-invariant filter to be representable by a graph shift matrix) can be ignored because any non-shift-enabled matrix can be converted to a shift-enabled matrix, such a conversion in general may not hold for a directed graph with non-sym...
Article
The large-scale deployment of smart metering worldwide has ignited renewed interest in electrical load disaggregation, or non-intrusive load monitoring (NILM). Most NILM algorithms disaggregate one appliance at a time, remove the estimated appliance contribution from the total load, and then move on to disaggregate the next appliance. On one hand,...
Preprint
Full-text available
In a 2013 paper by Sandryhaila and Moura, the authors introduced a condition (herein we will call it shift-enabled condition) that any shift invariant filter can be represented by the shift matrix if the condition is satisfied. In the same, the authors also argued that shift-enabled condition can be ignored as any non-shift-enabled matrix can be co...
Article
Full-text available
Large-scale smart energy metering deployment worldwide and integration of smart meters within the smart grid will enable two-way communication between the consumer and energy network, thus ensuring improved response to demand. Energy disaggregation or Non-intrusive load monitoring (NILM), namely disaggregation of the total metered electricity consu...
Article
We study detection of random signals corrupted by noise that over time switch their values (states) between a finite set of possible values, where the switchings occur at unknown points in time. We model such signals as hidden semi-Markov signals (HSMS), which generalize classical Markov chains by introducing explicit (possibly non-geometric) distr...
Conference Paper
Most current non-intrusive load monitoring (NILM) algorithms disaggregate one appliance at a time, remove the appliance contribution towards the total load, and then move on to the next appliance. On one hand, this is effective since it avoids multi-class classification, and analytical models for each appliance can be developed independently of oth...
Article
Full-text available
We study detection of random signals corrupted by noise that over time switch their values (states) from a finite set of possible values, where the switchings occur at unknown points in time. We model such signals by means of a random duration model that to each possible state assigns a probability mass function which controls the statistics of dur...
Article
With the large-scale roll-out of smart metering worldwide, there is a growing need to account for the individual contribution of appliances to the load demand. In this paper, we design a Graph signal processing (GSP)-based approach for non-intrusive appliance load monitoring (NILM), i.e., disaggregation of total energy consumption down to individua...

Citations

... Given a graph, necessary conditions for a graph filter to be representable as a polynomial of the graph shift matrix is discussed in [1] and [9], where the notion of shift-enabled graph is introduced as a graph where any shift-invariant filter H can be represented as a polynomial of the shift matrix. It is shown in [1] and [9] that the shift-enabled condition [1] is important for both directed and undirected graphs, and hence it needs to be taken into account. ...
... Meanwhile, various deep-learning-based methods have been proposed for the NILM approach. Classification methods based on graph signal processing (GSP) have been proposed and have shown good classification performance [22,23]. As an unsupervised approach, dynamic time warping (DTW) has been used to calculate similarities among signal patterns of appliances [22]. ...
... Load event (will be referred as "event" later) corresponds to the transition process of any two defined operating states of appliances, which is represented on the aggregated load power profile as a transient segment between two adjacent distinctly different steady-state segments, as shown schematically in Fig. 1. Different from the non-eventbased methods which require inference on all samples of the data, the event-based NILM methods take the load events as the objects of learning and inference, which has the significant advantages of low algorithm complexity and high operation efficiency [6]. Therefore, the event-based solutions are the mainstream choice for NILM technology implementations. ...
... Graph signal processing (GSP) [23] is a concept that effectively captures spatio-temporal correlation among data samples by embedding the structure of signals into a graph. Zhao et al. [27] proposed a low-resolution, event-based, unsupervised GSP approach. Recently, a modified cross-entropy method for event classification has been suggested [28], which is based on CO and formulates NILM as a cross-entropy problem. ...
... Hidden Markov models (HMMs) and various extensions of them are advocated in order to explore the possible combinations among the different appliances' state sequences [14,[29][30][31]. In this light, HMMs are state-based, so the studied appliances should have discrete states in their signatures [32]. ...
... More importantly, future iterations of this work should also consider performance metrics for energy estimation. While only a few authors have addressed the energy estimation step in event-based NILM (Giri & Bergés, 2015;He et al., 2018;Zhao et al., 2016), there are many eventless implementations (e.g., HMM and Deep ANNs) (Gomes & Pereira, 2020;Harell et al., 2019;Makonin, Popowich, et al., 2016;Murray et al., 2019) and public datasets (Pereira & Nunes, 2018) that can serve as a basis for this task. Furthermore, despite there is a category of metrics dedicated to energy estimation, it would be of crucial importance to also evaluate the potential of the metrics under the event detection category to assess the performance of energy estimation algorithms. ...
... where σ is the scaling factor, iteratively tuned as in [45] based on graph Fourier transform, and DTWdist(a, b) is the distance between sequences a and b. DTW is a commonly used metric for measuring similarity between two sequences of possibly different lengths. ...
... Support vector machines demonstrated good performance in [28][29][30][31]. K-nearest neighbors were explored in [32,33], decision tree in [34,35], k-means clustering in [36], and graph signal processing in [37,38]. More recently, deep learning approaches have prompted an upsurge in NILM research. ...