Subsea power cables are critical assets for electrical transmission and distribution networks, and highly relevant to regional, national, and international energy security and decarbonization given the growth in offshore renewable energy generation. Existing condition monitoring techniques are restricted to highly constrained online monitoring systems that only prioritize internal failure modes, representing only 30% of cable failure mechanisms, and has limited capacity to provide precursor indicators of such failures or damages. To overcome these limitations, we propose an innovative fusion prognostics approach that can provide the in situ integrity analysis of the subsea cable. In this paper, we developed low-frequency wide-band sonar (LFWBS) technology to collect acoustic response data from different subsea power cable sample types, with different inner structure configurations, and collate signatures from induced physical failure modes as to obtain integrity data at various cable degradation levels. We demonstrate how a machine learning approach, e.g., SVM, KNN, BP, and CNN algorithms, can be used for integrity analysis under a hybrid, holistic condition monitoring framework. The results of data analysis demonstrate the ability to distinguish subsea cables by differences of 5 mm in diameter and cable types, as well as achieving an overall 95%+ accuracy rate to detect different cable degradation stages. We also present a tailored, hybrid prognostic and health management solution for subsea cables, for cable remaining useful life (RUL) prediction. Our findings addresses a clear capability and knowledge gap in evaluating and forecasting subsea cable RUL. Thus, supporting a more advanced asset management and planning capability for critical subsea power cables.

Agroforestry can be defined as an agroecosystem whereby soil is used holistically and synergistically by various stakeholders including farmers, livestock, and plants. As such, agroforestry offers numerous benefits that include conservation of biodiversity, regulation of pests and diseases, increased quality of soil, air and water, efficient cycling of nutrients, and resilience to climate change. Review of published studies in agroforestry shows however that research in this area could benefit from increased real-time, spatial and temporal measurements. This situation is to be contrasted with that of precision agriculture in monocultures and precision livestock farming where progress made in sensor systems has attracted considerable research interest. It is advocated in this review article that wireless sensor networks could also significantly impact agroforestry through the monitoring of the local real-time interactions that occur between the various components constituting agroforestry systems. This review article proposes therefore the new field of data-driven agroforestry which lies at the intersection of precision agriculture, precision livestock farming, permaculture, and agroforestry. Data-driven agroforestry has the potential to not only help farmers harness the interactions between the different components of an agroforestry system to their advantage but also shine light on fundamental interactions between soil, plants, trees, and livestock while offering a sustainable agricultural method beneficial to all agroforestry stakeholders.

Convolutional Neural Networks (CNNs) are a popular choice for estimating Direction of Arrival (DoA) without explicitly estimating delays between multiple microphones. The CNN method first optimises unknown filter weights (of a CNN) by using observations and ground-truth directional information. This trained CNN is then used to predict incident directions given test observations. Most existing methods train using spectrally-flat random signals and test using speech. In this paper, which focuses on single source DoA estimation, we find that training with speech or music signals produces a relative improvement in DoA accuracy for a variety of audio classes across 16 acoustic conditions and 9 DoAs, amounting to an average improvement of around 17% and 19% respectively when compared to training with spectrally flat random signals. This improvement is also observed in scenarios in which the speech and music signals are synthesised using, for example, a Generative Adversarial Network (GAN). When the acoustic environments during test and training are similar and reverberant, training a CNN with speech outperforms Generalized Cross Correlation (GCC) methods by about 125%. When the test conditions are different, a CNN performs comparably. This paper takes a step towards answering open questions in the literature regarding the nature of the signals used during training, as well as the amount of data required for estimating DoA using CNNs.

Inspection of masonry buildings, typically railway bridges, for crack detection is currently performed by humans under tedious and sometimes dangerous working conditions. Over the past years, computer vision based techniques have been developed to automate structure visual inspections. These techniques could be integrated with (semi) autonomous drone surveillance to collect images of assets for full automation of simultaneous inspection and crack detection in railway bridges. In this study we have adopted the architecture of Faster R-CNN object detectors to provide crack detection in images. In this architecture, we have tested three networks (Mobilenetv2, Resnet50 and ZF512) to be utilised as feature extractors in a limited resource system for crack detection. We propose a new way of performing detection that we call Progressive Detection to increase the robustness of detection, considering otherwise only partially detected cracks. Since one of the main goals of visual inspection is checking the health of every single defect, we have revisited binary classification of images with and without cracks from a detection point of view, with the objective of minimising crack missing rates. Results show that Mobilenetv2 performs both successfully and fast enough to be applied in a drone application as a feature extractor network, achieving a close level of performance to the more sophisticated network Resnet50 with half its inference time. Regarding classification, Mobilenetv2 achieves its best performance in the early stages of its training process, showcasing 93 % accuracy and a crack miss rate ranging from 1 % to 15%. These results are comparable to Resnet's and better than ZF512's.

Subsea power cables are critical infrastructure for the continuity of energy supply and are a key enabler to the global growth in offshore renewable energy generation. Capital projects for long range, greater than 60km distances, for transmission networks can cost in excess of £1billion. In this paper, we have extensively reviewed the data within academia and industry with respect to the practices and challenges of subsea power cable management. With a detailed focus on 15 years of historical cable failure data from the UKs largest owner of subsea power cables, we identified that existing commercial monitoring systems do not monitor about 70% of subsea power cable failure modes. To overcome the challenges this represents to delivering cost effective and timely intervention to subsea power cables, we present a fusion prognostic model to enable predictive forecast on cable failure modes, include location and rates of degradation. In our model, we incorporate physical models to simulate the process where common cable failure modes lead to cable damage, such as abrasion and corrosion. In addition, we implemented multi-physics modelling techniques to model cable displacement and scouring, taking into consideration different environmental condition profiles. We also demonstrate how new sensing technologies can be integrated into this sensor agnostic model in order to enhance lifetime prediction accuracy. An operational decision support system is implemented within this work to integrate these different physical failure models, using a fusion model approach which integrates in-situ inspection data from sonar, autonomous underwater vehicle (AUV) inspection mission planning, and data analysis results into a holistic subsea cable remaining useful life prediction capability.

Subsea power cables are essential assets for the electrical transmission and distribution networks. They are crucial in ensuring the security of electricity supply and supporting the global expansion in offshore renewable energy generation. After reviewing historical data on subsea cable failure modes, we established that existing monitoring systems do not account for over 70% of subsea cable failure modes. The current technologies focus on electrical failure modes and subsea cable asset management strategies are typically reactive or time based, with inspection limited to diver and/or ROV supported video footage which has several limitations, such as requiring good visibility, access to the cable, challenges in locating the cable and inability to identify failure modes at the interface of the seabed. To overcome these limitations, we propose an innovative sensor technology that can provide the in-situ integrity analysis of the subsea cable. In this paper, we applied low frequency sonar technology to undertake detailed and in-situ assessment of subsea cable integrity. Specifically, in our work, a wideband low frequency (LF) sonar scanning system is manufactured to collect acoustic response from different subsea power cable samples with different inner structure and external failure modes. In addition, accelerated life cycle testing was conducted by manually introduce controlled stages of corrosion and abrasion to the cables to obtain integrity data at various cable degradation levels. Seminal results provide a detailed library of LF sonar responses to cable type and failure mode variations. The results of preliminary data analysis demonstrate the ability to distinguish subsea cables by differences in diameter and cable types and achieve an overall 95%+ accuracy rate to detect different cable degradation stages.

The study presents an underwater acoustic signal processing approach for the identification of the material of an object using wideband chirp pulses. The echo from the target is formed by a number of processes that occur during reflection of the pulse from an object. The timing of these reflected components and geometry of their wave paths determine speed of sound in the wave propagating material. The calculated speed of sound is then used to identify the material. The novel method enables automated material identification without any training data. The object of the material identification is a two-layer metal spherical shell filled with liquid and placed in a fresh water tank. The presented approach identifies the shell and filler materials of the sphere. In this work, the sphere thickness is limited by 2% in relation to the radius. Results are evaluated for the sphere's radius in a range from 0.05 to 0.15 m. The approach yields 83.5% success rate for the shell material identification and 79.9% success rate for the filler material on the synthetic data. For the experimental data, 67.3 and 80% success rates were obtained for shell and filler material identification, respectively.

One way to recognise an object is to study how the echo has been shaped during the interaction with the target. Wideband sonar allows the study of the energy distribution for a large range of frequencies. The frequency distribution contains information about an object, including its inner structure. This information is a key for automatic recognition. The scattering by a target can be quantitatively described by its Form Function. The Form Function can be calculated based on the data of the initial pulse, reflected pulse and parameters of a medium where the pulse is propagating. In this work spherical objects are classified based on their filler material - water or air. We limit the study to spherical 2 layered targets immersed in water. The Form Function is used as a descriptor and fed into a Neural Network classifier, Multilayer Perceptron (MLP). The performance of the classifier is compared with Support Vector Machine (SVM) and the Form Function descriptor is examined in contrast to the Time and Frequency Representation of the echo.

Subsea power cables are critical assets within the distribution and transmission infrastructure of electrical networks. Based on review of historical failure data, we have discovered that 70% of their failure modes are not monitored by current commercial monitoring systems, which predominately focus on the internal failure modes associated with partial discharge or overheating using online monitoring methods or embedded fibre-optics. In this paper, we proposed a fusion prognostic model for subsea cable lifetime prediction, employing an analytical model which is supported by accelerated aging data. In addition, we presented the preliminary results of our low frequency (LF) sonar analysis of subsea cables with a view to insitu integrity monitoring of cables. Seminal analysis on the echo data allowed us to distinguish cable samples with different types and diameter differentials of 2mm. Future work will focus on the integration and validation of this sonar data into the predictions of a cable's remaining useful life

As part of the UK’s programme for a sustainable, low-carbon future, increasing use will be made of wind energy. Presently the UK has 5 GW of wind energy, enough to power about 3 million homes1. This could rise to 40 GW in the coming decades, most of which would be in the form of offshorewind farms. The capital investment required for this UK offshore infrastructure could be as much as£ 120 billion. Operation and Maintenance (O&M) in the UK could be an industry worth as much as£ 2 billion per year by 2025. With the UK being a world-leader in offshore engineering, this presents huge potential for UK industry to establish itself, become leading exporter of services in this fieldand support 60,000 direct and indirect jobs in the UK by 20322. Present industry practice in the sector is good, though there is substantial opportunity to adopt best practice from other sectors, particularly the offshore oil and gas industry …

Accurate estimation of Time-Difference of Arrivals (TDOAs) is necessary to perform accurate sound source localization. The problem has traditionally been solved by using methods such as Generalized Cross-Correlation, which uses the entire signal to accurately estimate TDOAs. However, this could pose a problem in distributed sensor networks in which the amount of data that can be transmitted from each sensor to a fusion center is limited, such as in underwater scenarios or other challenging environments. Inspired by approaches from computer vision, in this paper we identify Scale-Invariant Feature Transform (SIFT) keypoints in the signal spectrogram. We perform cross-correlation on the signal using only the information available at those extracted keypoints. We test our algorithm in scenarios featuring different noise and reverberation conditions, and using different speech signals and source locations. We show that our algorithm can estimate Time-Difference of Arrivals (TDOAs) and the source location within an acceptable error range at a compression ratio of 40 : 1.

In-situ and above water radiometers are a critical for validating Ocean Color Satellite measurements, used to monitor in-water constituents of the global ocean. The calibration process, instrument response characterization, and environmental measurement all contribute to the overall uncertainty budget of the radiometric measurement. An integral part of this uncertainty traceability chain is accurate laboratory calibration of radiometric sensors. Over its lifetime, the Sea-Bird Scientific Halifax site (formerly Satlantic, LP) participated in inter-laboratory comparisons to ensure the quality of its calibrations. These include: NASA’s Seventh SeaWIFS Intercalibration Round-Robin Experiment (SIRREX-7, Hooker et al. 2002), conducted in 1999, compared Halifax to the Center for Hydro-Optics and Remote Sensing (CHORS, San Diego State University, California, USA) and the Joint Research Centre (JRC, Ispra, Italy). More recently, Sea-Bird Scientific participated in the European Space Agency (ESA) sponsored Fiducial Reference Measurements for Satellite Ocean Colour (FRM4SOC) program. In 2017, Sea-Bird Scientific transitioned the manufacturing and calibration of radiometric products from the facility located in Halifax (HAL), Nova Scotia CA to the facility located in Philomath (PHI), Oregon USA (formerly WET Labs, Inc.). As part of this transition, the radiometer calibration facility was reproduced at the Philomath site and Sea-Bird Scientific conducted an extensive cross facility set of experiments to: 1. Quantify relative calibration uncertainties within and between Halifax and Philomath laboratories; 2. Quantify differences in repeatability relative to Halifax (established standard); 3. Compare relative laboratory calibration uncertainties to budget of estimated uncertainty sources; 4. Verify successful transfer of build and calibration processes at Philomath site.

This paper is focused on improving the self-awareness of autonomous underwater vehicles (AUVs) Operating in unknown environments. A runtime estimation framework is introduced to derive energy usage and navigation performance metrics in the presence of external disturbances, such as slowly varying sea currents. These are calculated by a state-of-the-art nonlinear regression algorithm (LWPR) using measurements commonly available on-board modern AUVs without relying on external sensors or a priori knowledge about the environment. The proposed framework is validated on two vehicles, an IVER3 AUV and a Nessie VII AUV, in the context of real sea trials with no modification required for the AUVs or their missions. Derived metrics are used to estimate the feasibility of underwater missions employing the concept of probability of mission completion (PoMC). If environmental effects modify the vehicle's effectiveness, a mission plan update is performed. This is based on an energy-aware route optimization algorithm that is also introduced in the paper. This algorithm, known as energy-aware orienteering problem (EA-OP), shows a practical usage for the runtime metrics. It allows an AUV to optimize its navigation and to maximize its mission's outcome according to measured performances. Simulation results are also presented for inspection scenarios. These show average improvements of 5%-20% for the mission's outcome when using the proposed strategy in the presence of environmental disturbances.

Acoustic source localization (ASL) is an important problem. Despite much attention over the past few decades, rapid and robust ASL still remains elusive. A popular approach is to use a circular array of microphones to record the acoustic signal followed by some form of optimization to deduce the most likely location of the source. In this paper, we study the impact of the configuration of microphones on the accuracy of localization. We perform experiments using simulation as well as real measurements using a 72-microphone acoustic camera which confirm that circular configurations lead to higher localization error than spiral and wheel configurations when considering large regions of space. Moreover, the configuration of choice is intricately tied to the optimization scheme. We show that direct optimization of well known formulations for ASL yield errors similar to the state of the art (steered response power) with 6 × less computation.

This paper presents classification of spherical objects with different physical properties. The classification is based on the energy distribution in wideband pulses that have been scattered from objects. The echo is represented in Time-Frequency Domain (TFD), using Short Time Fourier Transform (STFT) with different window lengths, and is fed into a Convolution Neural Network (CNN) for classification. The results for different window lengths are analysed to study the influence of time and frequency resolution in classification. The CNN performs the best results with accuracy of (98.44 ± 0.8)% over 5 object classes trained on grayscale TFD images with 0.1 ms window length of STFT. The CNN is compared with a Multilayer Perceptron classifier, Support Vector Machine, and Gradient Boosting.

This paper presents a study of different feature extraction methods for sea floor recognition. This analysis is based on the frequency components of the return signals from a bio-inspired wideband sonar. Part of the sea floor of Tallinn Bay was surveyed using a wideband sonar with linear chirp pulses. The area contains citadel constructions and sediment regions. The recognition task is to distinguish between these two types of the sea floor. It was observed that the frequency spectrum of the signals returned from these two surfaces has different distributions. The work presents three frequency based approaches for the feature extraction process: Time-Frequency Moment Singular Value Decomposition (TFM-SVD), Energy Vector and Complete Frequency Spectrum. Comparison of the techniques is made in terms of the classification results.

This paper explores the possibility of using the relationship between the frequency of the transmit waveform and the location of grating lobes to achieve beam steering. The performance of the proposed beam steering technique is compared with conventional phase steering in terms of hardware requirements, beamwidths and angular coverage.

Multiple-input-multiple-output (MIMO) sonar systems offer new perspectives for target detection and area surveillance. This paper introduces a unified formulation for sonar MIMO systems and focuses on the target detection and recognition capability of these systems. The multiplication of the number of transmitters and receivers not only provides a greater variety in terms of target view angles but also provides meaningful statistics on the target itself. Assuming that views are independent and the MIMO system is large enough, we demonstrate that target recognition is possible with only one MIMO snapshot. By studying the detection performance of MIMO sonars we also demonstrate that such systems solve the speckle noise and decorrelate individual scatterers inside one cell resolution, leading to super-resolution imaging. We also show that, if carefully designed, MIMO systems can surpass the resolution of a synthetic aperture sonar (SAS) system using the same bandwidth. All the discussed properties are derived from the independent view assumption. Fulfilling this assumption drives the design and efficiency of such systems.

Wideband bio-inspired sonar allows to study echo structure over a broad range of frequencies. The emitted wideband pulses have a linear chirp design. The structure of reflected pulses from the object provides information about a target’s content. Dips and peaks in the echo structure provide information about the target characteristic. The echo is highly dependent on the emitted pulse. Therefore classification result can be improved by an adaptive approach where the emitted pulse is being refined based on echo processing. This research presents a novel approach of wideband pulse adaptation for better object characterisation. The paper describes an adaptation of a pulse’s slope values. Image processing techniques are applied to localize peaks and dips in the Time-Frequency Representation of the echo signal.