Non-intrusive appliance load monitoring is the process of disaggregating a household's total electricity consumption into its contributing appliances. In this paper we propose an unsupervised training method for non-intrusive monitoring which, unlike existing supervised approaches, does not require training data to be collected by sub-metering individual appliances, nor does it require appliances to be manually labelled for the households in which disaggregation is performed. Instead, we propose an approach which combines a one-off supervised learning process over existing labelled appliance data sets, with an unsupervised learning method over unlabelled household aggregate data. First, we propose an approach which uses the Tracebase data set to build probabilistic appliance models which generalise to previously unseen households, which we empirically evaluate through cross validation. Second, we use the Reference Energy Disaggregation Data set to evaluate the accuracy with which these general models can be tuned to the appliances within a specific household using only aggregate data. Our empirical evaluation demonstrates that general appliance models can be constructed using data from only a small number of appliances (typically 3-6 appliances), and furthermore that 28-99% of the remaining behaviour which is specific to a single household can be learned using only aggregate data from existing smart meters.$ A preliminary version of this work appeared in [31]

No abstract

Monitoring the health of the elderly living independently in their own homes is a key issue in building sustainable healthcare models which support a country's ageing population. Existing approaches have typically proposed remotely monitoring the behaviour of a household's occupants through the use of additional sensors. However the costs and privacy concerns of such sensors have significantly limited their potential for widespread adoption. In contrast, in this paper we propose an approach which detects Activities of Daily Living, which we use as a proxy for the health of the household residents. Our approach detects appliance usage from existing smart meter data, from which the unique daily routines of the household occupants are learned automatically via a log Gaussian Cox process. We evaluate our approach using two real-world data sets, and show it is able to detect over 80% of kettle uses while generating less than 10% false positives. Furthermore, our approach allows earlier interventions in households with a consistent routine and fewer false alarms in the remaining households, relative to a fixed-time intervention benchmark.

Non-intrusive load monitoring, or energy disaggregation, aims to separate household energy consumption data collected from a single point of measurement into appliancelevel consumption data. In recent years, the field has rapidly expanded due to increased interest as national deployments of smart meters have begun in many countries. However, empirically comparing disaggregation algorithms is currently virtually impossible. This is due to the different data sets used, the lack of reference implementations of these algorithms and the variety of accuracy metrics employed. To address this challenge, we present the Non-intrusive Load Monitoring Toolkit (NILMTK); an open source toolkit designed specifically to enable the comparison of energy disaggregation algorithms in a reproducible manner. This work is the first research to compare multiple disaggregation approaches across multiple publicly available data sets. Our toolkit includes parsers for a range of existing data sets, a collection of preprocessing algorithms, a set of statistics for describing data sets, two reference benchmark disaggregation algorithms and a suite of accuracy metrics. We demonstrate the range of reproducible analyses which are made possible by our toolkit, including the analysis of six publicly available data sets and the evaluation of both benchmark disaggregation algorithms across such data sets.

Abstract-Non-intrusive load monitoring (NILM), or energy disaggregation, is the process of using signal processing and machine learning to separate the energy consumption of a building into individual appliances. In recent years, a number of data sets have been released in order to evaluate such approaches, which contain both building-level and appliance-level energy data. However, these data sets typically cover less than 10 households due to the financial cost of such deployments, and are not released in a format which allows the data sets to be easily used by energy disaggregation researchers. To this end, the Dataport database was created by Pecan Street Inc, which contains 1 minute circuit-level and building-level electricity data from 722 households. Furthermore, the non-intrusive load monitoring toolkit (NILMTK) was released in 2014, which provides software infrastructure to support energy disaggregation research, such as data set parsers, benchmark disaggregation algorithms and accuracy metrics. This paper describes the release of a subset of the Dataport database in NILMTK format, containing one month of electricity data from 669 households. Through the release of this Dataport data in NILMTK format, we pose a challenge to the signal processing community to produce energy disaggregation algorithms which are both accurate and scalable.

Non-intrusive appliance load monitoring is the process of disaggregating a household's total electricity consumption into its contributing appliances. In this paper we propose an approach by which individual appliances can be iteratively separated from an aggregate load. Unlike existing approaches, our approach does not require training data to be collected by sub-metering individual appliances, nor does it assume complete knowledge of the appliances present in the household. Instead, we propose an approach in which prior models of general appliance types are tuned to specific appliance instances using only signatures extracted from the aggregate load. The tuned appliance models are then used to estimate each appliance's load, which is subsequently subtracted from the aggregate load. This process is applied iteratively until all appliances for which prior behaviour models are known have been disaggregated. We evaluate the accuracy of our approach using the REDD data set, and show the disaggregation performance when using our training approach is comparable to when sub-metered training data is used. We also present a deployment of our system as a live application and demonstrate the potential for personalised energy saving feedback.

In recent years, the field of computational sustainability has striven to apply artificial intelligence techniques to solve ecological and environmental problems. In ecology, a key issue for the safeguarding of our planet is the monitoring of biodiversity. Automated acoustic recognition of species aims to provide a cost-effective method for biodiversity monitoring. This is particularly appealing for detecting endangered animals with a distinctive call, such as the New Forest cicada. To this end, we pursue a crowdsourcing approach, whereby the millions of visitors to the New Forest, where this insect was historically found, will help to monitor its presence by means of a smartphone app that can detect its mating call. Existing research in the field of acoustic insect detection has typically focused upon the classification of recordings collected from fixed field microphones. Such approaches segment a lengthy audio recording into individual segments of insect activity, which are independently classified using cepstral coefficients extracted from the recording as features. This paper reports on a contrasting approach, whereby we use crowdsourcing to collect recordings via a smartphone app, and present an immediate feedback to the users as to whether an insect has been found. Our classification approach does not remove silent parts of the recording via segmentation, but instead uses the temporal patterns throughout each recording to classify the insects present. We show that our approach can successfully discriminate between the call of the New Forest cicada and similar insects found in the New Forest, and is robust to common types of environment noise. A large scale trial deployment of our smartphone app collected over 6000 reports of insect activity from over 1000 users. Despite the cicada not having been rediscovered in the New Forest, the effectiveness of this approach was confirmed for both the detection algorithm, which successfully identified the same cicada through the app in countries where the same species is still present, and of the crowdsourcing methodology, which collected a vast number of recordings and involved thousands of contributors.

We present a system and study of personalized energyrelated recommendation. AgentSwitch utilizes electricity usage data collected from users' households over a period of time to realize a range of smart energy-related recommendations on energy tariffs, load detection and usage shifting. The web service is driven by a third party real-time energy tariff API (uSwitch), an energy data store, a set of algorithms for usage prediction, and appliance-level load disaggregation. We present the system design and user evaluation consisting of interviews and interface walkthroughs. We recruited participants from a previous study during which three months of their household's energy use was recorded to evaluate personalized recommendations in AgentSwitch. Our contributions are a) a systems architecture for personalized energy services; and b) findings from the evaluation that reveal challenges in designing energy-related recommender systems. In response to the challenges we formulate design recommendations to mitigate barriers to switching tariffs, to incentivize load shifting, and to automate energy management.