Visual place recognition (VPR) is the process of recognising a previously visited place using visual information, often under varying appearance conditions and viewpoint changes and with computational constraints. VPR is related to the concepts of localisation, loop closure, image retrieval and is a critical component of many autonomous navigation systems ranging from autonomous vehicles to drones and computer vision systems. While the concept of place recognition has been around for many years, VPR research has grown rapidly as a field over the past decade due to improving camera hardware and its potential for deep learning-based techniques, and has become a widely studied topic in both the computer vision and robotics communities. This growth however has led to fragmentation and a lack of standardisation in the field, especially concerning performance evaluation. Moreover, the notion of viewpoint and illumination invariance of VPR techniques has largely been assessed qualitatively and hence ambiguously in the past. In this paper, we address these gaps through a new comprehensive open-source framework for assessing the performance of VPR techniques, dubbed “VPR-Bench”. VPR-Bench (Open-sourced at: https://github.com/MubarizZaffar/VPR-Bench) introduces two much-needed capabilities for VPR researchers: firstly, it contains a benchmark of 12 fully-integrated datasets and 10 VPR techniques, and secondly, it integrates a comprehensive variation-quantified dataset for quantifying viewpoint and illumination invariance. We apply and analyse popular evaluation metrics for VPR from both the computer vision and robotics communities, and discuss how these different metrics complement and/or replace each other, depending upon the underlying applications and system requirements. Our analysis reveals that no universal SOTA VPR technique exists, since: (a) state-of-the-art (SOTA) performance is achieved by 8 out of the 10 techniques on at least one dataset, (b) SOTA technique in one community does not necessarily yield SOTA performance in the other given the differences in datasets and metrics. Furthermore, we identify key open challenges since: (c) all 10 techniques suffer greatly in perceptually-aliased and less-structured environments, (d) all techniques suffer from viewpoint variance where lateral change has less effect than 3D change, and (e) directional illumination change has more adverse effects on matching confidence than uniform illumination change. We also present detailed meta-analyses regarding the roles of varying ground-truths, platforms, application requirements and technique parameters. Finally, VPR-Bench provides a unified implementation to deploy these VPR techniques, metrics and datasets, and is extensible through templates.@en

The transition towards a more decarbonised, resilient and distributed energy system requires local initiatives, such as Smart Local Energy Systems (SLES), which lead communities to gain self-sufficiency and become electricity islands. Although many SLES projects have been recently deployed, only a few of them have managed to be successful, mostly due to an initial knowledge gap in the SLES planning and deployment phases. This paper leverages the knowledge from the UK's largest SLES demonstrator in the Orkney Islands, named the Responsive FLEXibility (ReFLEX) project, to propose a framework that will help communities to successfully implement a SLES. First, this paper describes how the multi-services electrical SLES implemented in Orkney reduces the impact of the energy transition on the electrical infrastructure. We identify and discuss the main enablers and barriers to a successful SLES, based on a review of SLES projects in the UK. Second, to help future communities to implement SLES, we extend the Smart Grid Architecture Model (SGAM) into a comprehensive multi-vector Smart Local Energy Architecture Model (SLEAM) that includes all main energy services, namely power, heat and transport. This extended architecture model describes the main components and interaction layers that need to be addressed in a comprehensive SLES. Next, to inform successful deployment of SLES, an extensive list of key performance indicators for SLES is proposed and implemented for the ReFLEX project. Finally, we discuss lessons learnt from the ReFLEX project and we list required future technologies that enable communities, energy policy makers and regulatory bodies to best prepare for the energy transition.@en

The UK has set plans to increase offshore wind capacity from 22GW to 154GW by 2030. With such tremendous growth, the sector is now looking to Robotics and Artificial Intelligence (RAI) in order to tackle lifecycle service barriers as to support sustainable and profitable offshore wind energy production. Today, RAI applications are predominately being used to support short term objectives in operation and maintenance. However, moving forward, RAI has the potential to play a critical role throughout the full lifecycle of offshore wind infrastructure, from surveying, planning, design, logistics, operational support, training and decommissioning. This paper presents one of the first systematic reviews of RAI for the offshore renewable energy sector. The state-of-the-art in RAI is analyzed with respect to offshore energy requirements, from both industry and academia, in terms of current and future requirements. Our review also includes a detailed evaluation of investment, regulation and skills development required to support the adoption of RAI. The key trends identified through a detailed analysis of patent and academic publication databases provide insights to barriers such as certification of autonomous platforms for safety compliance and reliability, the need for digital architectures for scalability in autonomous fleets, adaptive mission planning for resilient resident operations and optimization of human machine interaction for trusted partnerships between people and autonomous assistants. Our study concludes with identification of technological priorities and outlines their integration into a new ‘symbiotic digital architecture’ to deliver the future of offshore wind farm lifecycle management.@en

Given the ongoing transition towards a more decentralised and adaptive energy system, the potential of blockchain-enabled smart contracts for the energy sector is being increasingly recognised. Due to their self-executing, customisable and tamper-proof nature, they are seen as a key technology for enabling the transition to a more efficient, transparent and transactive energy market. The applications of smart contracts include coordination of smart electric vehicle charging, automated demand-side response, peer-to-peer energy trading and allocation of the control duties amongst the network operators. Nevertheless, their use in the energy sector is still in its early stages as there are many open challenges related to security, privacy, scalability and billing. In this paper, we systematically review 178 peer-reviewed publications and 13 innovation projects, providing a thorough analysis of the strengths and weaknesses of smart contracts used in the energy sector. This work offers a broad perspective on the opportunities and challenges that stakeholders using this technology face, in both current and emergent markets, such as peer-to-peer energy trading platforms. To provide a roadmap for researchers and practitioners interested in the technology, we propose a systematic model of the smart contracting process, by developing a novel 6-layer architecture, as well as presenting a sample energy contract in pseudocode form and as open-source code. Our analysis focuses on the two mainstream application areas we identify for smart contract use in this area: energy and flexibility trading, and distributed control. The paper concludes with a comprehensive, critical discussion of the advantages and challenges that must be addressed in the area of smart contracts and blockchains in energy, and a set of recommendations that researchers and developers should consider when applying smart contracts to energy system settings.@en

This paper reviews the recent literature on machine learning (ML) models that have been used for condition monitoring in wind turbines (e.g. blade fault detection or generator temperature monitoring). We classify these models by typical ML steps, including data sources, feature selection and extraction, model selection (classification, regression), validation and decision-making. Our findings show that most models use SCADA or simulated data, with almost two-thirds of methods using classification and the rest relying on regression. Neural networks, support vector machines and decision trees are most commonly used. We conclude with a discussion of the main areas for future work in this domain.@en

This paper reviews the recent literature on machine learning (ML) models that have been used for condition monitoring in wind turbines (e.g. blade fault detection or generator temperature monitoring). We classify these models by typical ML steps, including data sources, feature selection and extraction, model selection (classification, regression), validation and decision-making. Our findings show that most models use SCADA or simulated data, with almost two-thirds of methods using classification and the rest relying on regression. Neural networks, support vector machines and decision trees are most commonly used. We conclude with a discussion of the main areas for future work in this domain.@en

Blockchains or distributed ledgers are an emerging technology that has drawn considerable interest from energy supply firms, startups, technology developers, financial institutions, national governments and the academic community. Numerous sources coming from these backgrounds identify blockchains as having the potential to bring significant benefits and innovation. Blockchains promise transparent, tamper-proof and secure systems that can enable novel business solutions, especially when combined with smart contracts. This work provides a comprehensive overview of fundamental principles that underpin blockchain technologies, such as system architectures and distributed consensus algorithms. Next, we focus on blockchain solutions for the energy industry and inform the state-of-the-art by thoroughly reviewing the literature and current business cases. To our knowledge, this is one of the first academic, peer-reviewed works to provide a systematic review of blockchain activities and initiatives in the energy sector. Our study reviews 140 blockchain research projects and startups from which we construct a map of the potential and relevance of blockchains for energy applications. These initiatives were systematically classified into different groups according to the field of activity, implementation platform and consensus strategy used.1 Opportunities, potential challenges and limitations for a number of use cases are discussed, ranging from emerging peer-to-peer (P2P) energy trading and Internet of Things (IoT) applications, to decentralised marketplaces, electric vehicle charging and e-mobility. For each of these use cases, our contribution is twofold: first, in identifying the technical challenges that blockchain technology can solve for that application as well as its potential drawbacks, and second in briefly presenting the research and industrial projects and startups that are currently applying blockchain technology to that area. The paper ends with a discussion of challenges and market barriers the technology needs to overcome to get past the hype phase, prove its commercial viability and finally be adopted in the mainstream.@en

Recent years have seen an increasing interest in Demand Response (DR) as a means to provide flexibility, and hence improve the reliability of energy systems in a cost-effective way. Yet, the high complexity of the tasks associated with DR, combined with their use of large-scale data and the frequent need for near real-time decisions, means that Artificial Intelligence (AI) and Machine Learning (ML) — a branch of AI — have recently emerged as key technologies for enabling demand-side response. AI methods can be used to tackle various challenges, ranging from selecting the optimal set of consumers to respond, learning their attributes and preferences, dynamic pricing, scheduling and control of devices, learning how to incentivise participants in the DR schemes and how to reward them in a fair and economically efficient way. This work provides an overview of AI methods utilised for DR applications, based on a systematic review of over 160 papers, 40 companies and commercial initiatives, and 21 large-scale projects. The papers are classified with regards to both the AI/ML algorithm(s) used and the application area in energy DR. Next, commercial initiatives are presented (including both start-ups and established companies) and large-scale innovation projects, where AI methods have been used for energy DR. The paper concludes with a discussion of advantages and potential limitations of reviewed AI techniques for different DR tasks, and outlines directions for future research in this fast-growing area.@en

Recent years have seen an increasing interest in Demand Response (DR) as a means to provide flexibility, and hence improve the reliability of energy systems in a cost-effective way. Yet, the high complexity of the tasks associated with DR, combined with their use of large-scale data and the frequent need for near real-time decisions, means that Artificial Intelligence (AI) and Machine Learning (ML) — a branch of AI — have recently emerged as key technologies for enabling demand-side response. AI methods can be used to tackle various challenges, ranging from selecting the optimal set of consumers to respond, learning their attributes and preferences, dynamic pricing, scheduling and control of devices, learning how to incentivise participants in the DR schemes and how to reward them in a fair and economically efficient way. This work provides an overview of AI methods utilised for DR applications, based on a systematic review of over 160 papers, 40 companies and commercial initiatives, and 21 large-scale projects. The papers are classified with regards to both the AI/ML algorithm(s) used and the application area in energy DR. Next, commercial initiatives are presented (including both start-ups and established companies) and large-scale innovation projects, where AI methods have been used for energy DR. The paper concludes with a discussion of advantages and potential limitations of reviewed AI techniques for different DR tasks, and outlines directions for future research in this fast-growing area.@en

Context: Demonstrating high reliability and safety for safety-critical systems (SCSs) remains a hard problem. Diverse evidence needs to be combined in a rigorous way: in particular, results of operational testing with other evidence from design and verification. Growing use of machine learning in SCSs, by precluding most established methods for gaining assurance, makes evidence from operational testing even more important for supporting safety and reliability claims. Objective: We revisit the problem of using operational testing to demonstrate high reliability. We use Autonomous Vehicles (AVs) as a current example. AVs are making their debut on public roads: methods for assessing whether an AV is safe enough are urgently needed. We demonstrate how to answer 5 questions that would arise in assessing an AV type, starting with those proposed by a highly-cited study. Method: We apply new theorems extending our Conservative Bayesian Inference (CBI) approach, which exploit the rigour of Bayesian methods while reducing the risk of involuntary misuse associated (we argue) with now-common applications of Bayesian inference; we define additional conditions needed for applying these methods to AVs. Results: Prior knowledge can bring substantial advantages if the AV design allows strong expectations of safety before road testing. We also show how naive attempts at conservative assessment may lead to over-optimism instead; why extrapolating the trend of disengagements (take-overs by human drivers) is not suitable for safety claims; use of knowledge that an AV has moved to a “less stressful” environment. Conclusion: While some reliability targets will remain too high to be practically verifiable, our CBI approach removes a major source of doubt: it allows use of prior knowledge without inducing dangerously optimistic biases. For certain ranges of required reliability and prior beliefs, CBI thus supports feasible, sound arguments. Useful conservative claims can be derived from limited prior knowledge.@en

This paper details the results of a survey on safety and reliability of commercial Unmanned Aerial Systems (UAS), issued to 1500 companies in the UK, with a 10% response rate. The results of this study concur with previous studies that commercial UAS have a failure rate of 1/103 flight hours. Furthermore, we found a reported 49% of companies believe a critical failure will occur within every 500 hours of flight. We investigate the responses concerning UAS subsystems; the experience of onboard intelligent systems, current practices around the frequency of maintenance and servicing of subsystems, and subsystems identified in order of the highest likelihood of failure. Informed by the results of this survey and a state of the art literature review we identify emerging technologies and methods as candidate solutions to the respondents reported challenges, such as; Integrated Vehicle Health Management (IVHM), Formal Methods, Simulations and Fault Tolerant Control (FTC). The integration of these techniques, in a systematic framework with a supporting automatic tool-chain, is identified as a candidate solution to reliability and certification challenges in UAS.@en

This paper details the results of a survey on safety and reliability of commercial Unmanned Aerial Systems (UAS), issued to 1500 companies in the UK, with a 10% response rate. The results of this study concur with previous studies that commercial UAS have a failure rate of 1/103 flight hours. Furthermore, we found a reported 49% of companies believe a critical failure will occur within every 500 hours of flight. We investigate the responses concerning UAS subsystems; the experience of onboard intelligent systems, current practices around the frequency of maintenance and servicing of subsystems, and subsystems identified in order of the highest likelihood of failure. Informed by the results of this survey and a state of the art literature review we identify emerging technologies and methods as candidate solutions to the respondents reported challenges, such as; Integrated Vehicle Health Management (IVHM), Formal Methods, Simulations and Fault Tolerant Control (FTC). The integration of these techniques, in a systematic framework with a supporting automatic tool-chain, is identified as a candidate solution to reliability and certification challenges in UAS.@en

Recent years have seen an increasing interest in Demand Response (DR), as a means to satisfy the growing flexibility needs of modern power grids. This increased flexibility is required due to the growing proportion of intermittent renewable energy generation into the energy mix, and increasing complexity in demand profiles from the electrification of transport networks. Currently, less than 2% of the global potential for demand-side flexibility is currently utilised, but a more widespread adoption of residential consumers as flexibility resources can lead to substantially higher utilisation of the demand-side flexibility potential. In order to achieve this target, acquiring a better understanding of how residential DR participants respond in DR events is essential – and recent advances in novel machine learning and statistical AI provide promising tools to address this challenge. This study provides an in-depth analysis of how residential customers have responded in incentive-based DR, utilising household-related data from a large-scale, real-world trial: the Smart Grid, Smart City (SGSC) project. Using a number of different machine learning approaches, we model the relationship between a household’s response and household-related features. Moreover, we examine the potential effects of households’ features on the residential response behaviour, and highlight a number of key insights which raise questions about the reported level of consumers’ engagement in DR schemes, and the motivation for different customers’ response level. Finally, we explore the temporal structure of the response – and although we found no supporting evidence of DR responders learning over time for the available data from this trial, the proposed methodologies could be used for longer-term longitudinal DR studies. Our study concludes with a broader discussion of our findings and potential paths for future research in this emerging area.@en

Electromagnetic Relays (Electromagnetic Relay (EMR)s) are omnipresent in electrical systems, ranging from mass-produced consumer products to highly specialised, safety-critical industrial systems. Our detailed literature review focused on EMR reliability highlighting the methods used to estimate the State of Health or the Remaining Useful Life emphasises the limited analysis and understanding of expressive EMR degradation indicators, as well as accessibility and use of EMR life cycle data sets. Prioritising these open challenges, a deep learning pipeline is presented in a prognostic context termed Electromagnetic Relay Useful Actuation Pipeline (EMRUA). Leveraging the attributes of causal convolution, a Temporal Convolutional Network (TCN) based architecture integrates an arbitrary long sequence of multiple features to produce a remaining useful switching actuations forecast. These features are extracted from raw, high volume life cycle data sets, namely EMR switching data (Contact-Voltage, Contact-Current). Monte-Carlo Dropout is utilised to estimate uncertainty during inference. The TCN hyperparameter space, as well as various methods to select and analyse long sequences of multivariate time series data are investigated. Subsequently, our results demonstrate improvements using the developed statistical feature-set over traditional, time-based features, commonly found in literature. EMRUA achieves an average forecasting mean absolute percentage error of ±12 % over the course of the entire EMR life.@en

With increasing decarbonisation and accessibility to our energy systems and markets, there is a need to understand and optimise the value proposition for different stakeholders. Game-theoretic models represent a promising approach to study strategic interactions between self-interested private energy system investors. In this work, we design and evaluate a game-theoretic framework to study strategic interactions between profit-maximising players that invest in network, renewable generation and storage capacity. Specifically, we study the case where grid capacity is developed by a private renewable investor, but line access is shared with competing renewable and storage investors, thus enabling them to export energy and access electricity demand. We model the problem of deducing how much capacity each player should build as a non-cooperative Stackelberg-Cournot game between a dominant player (leader) who builds the power line and renewable generation capacity, and local renewable and storage investors (multiple followers), who react to the installation of the line by increasing their own capacity. Using data-driven analysis and simulations, we developed an empirical search method for estimating the game equilibrium, where the payoffs capture the realistic operation and control of the energy system under study. A practical demonstration of the underlying methodology is shown for a real-world grid reinforcement project in the UK. The methodology provides a realistic mechanism to analyse investor decision-making and investigate feasible tariffs that encourage distributed renewable investment, with sharing of grid access.@en

The energy landscape for the Low-Voltage (LV) networks is undergoing rapid changes. These changes are driven by the increased penetration of distributed Low Carbon Technologies, both on the generation side (i.e. adoption of micro-renewables) and demand side (i.e. electric vehicle charging). The previously passive ‘fit-and-forget’ approach to LV network management is becoming increasing inefficient to ensure its effective operation. A more agile approach to operation and planning is needed, that includes pro-active prediction and mitigation of risks to local sub-networks (such as risk of voltage deviations out of legal limits). The mass rollout of smart meters (SMs) and advances in metering infrastructure holds the promise for smarter network management. However, many of the proposed methods require full observability, yet the expectation of being able to collect complete, error free data from every smart meter is unrealistic in operational reality. Furthermore, the smart meter (SM) roll-out has encountered significant issues, with the current voluntary nature of installation in the UK and in many other countries resulting in low-likelihood of full SM coverage for all LV networks. Even with a comprehensive SM roll-out privacy restrictions, constrain data availability from meters. To address these issues, this paper proposes the use of a Deep Learning Neural Network architecture to predict the voltage distribution with partial SM coverage on actual network operator LV circuits. The results show that SM measurements from key locations are sufficient for effective prediction of the voltage distribution, even without the use of the high granularity personal power demand data from individual customers.@en

Recent years have seen a surge of interest in distributed residential batteries for households with renewable generation. Yet, assuring battery assets are profitable for their owners requires a complex optimisation of the battery asset and additional revenue sources, such as novel ways to access wholesale energy markets. In this paper, we propose a framework in which wholesale market bids are placed on forward energy markets by an aggregator of distributed residential batteries that are controlled in real time by a novel Home Energy Management System (HEMS) control algorithm to meet the market commitments, while maximising local self-consumption. The proposed framework consists of three stages. In the first stage, an optimal day-ahead or intra-day scheduling of the aggregated storage assets is computed centrally. For the second stage, a bidding strategy is developed for wholesale energy markets. Finally, in the third stage, a novel HEMS real-time control algorithm based on a smart contract allows coordination of residential batteries to meet the market commitments and maximise self-consumption of local production. Using a case study provided by a large U.K.-based energy demonstrator, we apply the framework to an aggregator with 70 residential batteries. Experimental analysis is done using real per minute data for demand and production. Results indicate that the proposed approach increases the aggregator's revenues by 35% compared to a case without residential flexibility, and increases the self-consumption rate of the households by a factor of two. The robustness of the results to uncertainty, forecast errors and to communication latency is also demonstrated.@en

The trend of decentralization of energy services has given rise to community energy systems. These energy communities aim to maximize the self-consumption of local renewable energy generated and stored in assets that are typically connected to low-voltage (LV) distribution networks. Energy community schemes often involve jointly owned assets such as community-owned solar photo-voltaic panels (PVs), wind turbines and/or shared battery storage. This raises the question of how these assets should be controlled in real-time, and how the energy outputs from these jointly owned assets should be shared fairly among heterogeneous community members. Crucially, such real-time control and fair sharing of energy must also consider the technical constraints of the community, such as the local LV network characteristics, voltage limits and power ratings of electric cables and transformers. In this paper, we design and analyze a heuristic-based battery control algorithm that considers the influence of battery life degradation, and the resultant increase in local renewable energy consumption within local operating constraints of the LV network. We provide a model that first studies the techno-economic benefits of community-owned versus individually-owned energy assets considering the network/grid constraints. Then, using the methodology and principles from cooperative game theory, we propose a redistribution model for benefits in a community based on the marginal contribution of each household. The results from our study demonstrate that the redistribution mechanism is fairer and computationally tractable compared to the existing state-of-the-art methods. Thus, our methodology is more scalable with respect to modeling the economic sharing of joint assets in community energy systems. @en