Schedule
| 0830 | Breakfast and registration - Robinson College Dining hall balcony |
| 0900 | Welcome |
| 0915 | Paul Harris, Phil Le Grice |
| 0945 | DAISY Sensing (Team Challenge 2024) |
| 1045 | Break |
| 1130 | Osarenkhoe Ogbeide |
| 1200 | Leon Barron |
| 1230 | Lunch |
| 1330 | Student talks (Wadood Tadbier, Anne-Pia Marty, Justas Brazauskas) |
| 1415 | Emily Lines |
| 1445 | Break |
| 1530 | Martin Peacock (Co-founder and CSO, Zimmer & Peacock) |
| 1545 | Lisa Hecker |
| 1615 | Shailaja Fennell |
| 1645 | Closing and Prizes (Sponsored by Marks & Clerk and Zimmer & Peacock) |
| 1700 - 1800 | Group photo, Drinks reception |
| 1900 | Dinner (by invitation, meet at the Dining Hall balcony at 1850) |
Invited speakers
Prof Paul Harris, Dr Phil Le Grice - 0915
Prof. Paul Harris is Principle Research Scientist at Rothamsted Research. He has several research themes: spatial/spatio-temporal statistics, geostatistics and geographically weighted (GW) models; development of R statistical packages for above themes (see GWmodel on CRAN); analyses of agricultural, marine, ecological, environmental and socio-economic data; analyses of remote sensing, crowd-sourced, land cover and land use data; hybridisation of agri-process-based models with statistical models; statistical and data mining methods for quality control of large, multivariate, spatio-temporal environmental data sets; visualisation; methodological development in robust, non-stationary and scale-dependent statistical models.
Dr Phil Le Grice is Scientific Specialist at Rothamsted Research. He is an Agriculturalist with a career that has spanned research in plant physiology, economic development project design and management, and senior academic leadership roles in the Agricultural and Land-Based sector. He has also held directorships and trusteeships for several agricultural organizations and charities. His current research interests include the emerging interfaces of Corporate Sustainability and Agri-Environmental Research and the role of data creation, curation, and harmonization in Agricultural and Environmental Research.
The North Wyke Farm Platform: An introduction and Future Initiatives
The North Wyke Farm Platform is a long-term experiment (established 2010) for the study of sustainable agricultural systems. The platform is divided into 15 distinct sub-catchments where three farming systems are managed under different farming regimes. An indoor farm creates a fourth system for comparison. Data is collected from these systems by a mixture of sensors and manual collection across multiple temporal and spatial scales. The full collection has just exceeded 100 million data points from over 400 deployed sensors. After 15 years of operation, we are able to reflect upon lessons learnt from the operation of the farm platform, its sensor deployment, its data management, its open data release and the role of long-term data infrastructure in the application of digital science to agri-environmental research.
DAISY Sensing (Team Challenge 2024) - 0945
Every Summer, our Sensor CDT MRes students complete a group project which aims to solve a real-world problem. This year, they worked in partnership with Rothamsted Research to design cattle-mounted pasture biodiversity monitoring sensors. You can read more about their project on their Team Challenge page.
Dr Osarenkhoe Ogbeide - 1130
Dr Osarenkhoe Ogbeide is a Junior Research Fellow at Churchill College, University of Cambridge and recipient of the William and Barbara Hawthorne Fellowship and Sydney Harvey Research Fellowship in Engineering. Ogbeide pursued his doctoral studies in Engineering at Churchill College, University of Cambridge with funding from the Engineering & Physical Sciences Research Council (EPSRC), supplemented by industrial support from Alphasense Ltd.
During his PhD journey, he delved into the realm of next-generation printed gas sensors using 2D materials, leading to the production of the most authoritative review in the field and the development of a predictive gas sensor that has the capacity to detect gases at untrained concentrations within mixed gas environments. His current research focus continues on the theme of gas sensing applications enabled through nanoengineering.
Inkjet Printed Graphene-Based Smart Gas Sensors
This presentation introduces innovative gas sensing technology using inkjet-printed, graphene-based materials. Tackling challenges in gas selectivity and mixed environments, we combine reduced graphene oxide (rGO) with binary metal oxides to develop highly sensitive, room-temperature sensors. By integrating machine learning, the platform can accurately detect specific gases and predict concentrations, even with interfering factors like humidity. Attendees will learn about the technology’s scalability and its broad potential applications, from environmental monitoring to industrial safety.
Dr Leon Barron - 1200
Dr. Leon Barron is a Reader in Analytical and Environmental Sciences at Imperial College London. He holds a BSc(hons) and PhD in analytical chemistry at Dublin City University, Ireland. He was Lecturer and then Senior lecturer in Forensic Science at King’s College London for 11 years until 2020 when he moved to Imperial to lead the Emerging Chemical Contaminants team within the Environmental Research Group. His research aims to further our understanding of the sources, risks and impacts existing, emerging and new chemical contaminants on environmental and public health. Recently, this work has been addressing the challenges with scaling up analytical approaches for chemicals in the environment.
Scaling up Chemical Pollution Monitoring in Water for ‘One Health’
One Health is an integrated and collaborative approach to balance and optimise the health of people, wildlife and the environment. The focus of this talk is on chemical pollution. The number of chemicals registered for manufacture and sale now exceeds 350,000. This poses a huge challenge to environmental scientists to not only detect and monitor them, but also their transformation products and their combined effects as complex mixtures on wildlife and humans. This talk will focus on recent advances to scale up analytical solutions for monitoring environmental contamination with large numbers of chemicals in the water cycle. Leon will focus on new low-cost tools for chemical sampling, the role of ‘citizen science’, rapid analysis and intelligent data analytics tools to enable environmental risk prioritisation of chemicals. In particular, the development and application of highly sensitive and rapid mass spectrometry-based analytical workflows for targeted analysis and suspect screening in surface, drinking and wastewater will be discussed, as well as its relevance in a One Health context. It is hoped that this technology can rapidly provide evidence to help understand what chemicals pose the largest risks, when and where they come from and to inform policies for chemical risk mitigation for a healthier world.
Dr Emily Lines - 1415
Dr. Emily Lines is an Associate Professor in Physical Geography at Cambridge University, UKRI Future Leaders' Fellow and Turing Fellow. She works at the intersection of forest ecology, remote sensing and data science. She studied Mathematics at Imperial College London, undertook a PhD in Plant Sciences at the University of Cambridge in collaboration with Microsoft Research, undertook a postdoctoral research associate at UCL working on the European Space Agency Data Assimilation project, and, prior to joining Cambridge, was a Senior Lecturer in Environmental Science at QMUL.
Ecological insights from high resolution remote sensing and AI in forests
Recent advances in remote sensing and data processing are revolutionising our ability to accurately measure tree and forest structure from leaves to landscapes. In this talk I will present our work using high resolution remote sensing to understand the structure, function and dynamics of forests, discuss technological choices and and demonstrate how AI can solve processing problems to enable us to scale the application of these technologies for use in practice.
Dr Lisa Hecker - 1545
Dr. Lisa Hecker joined the CDT in the 2016 cohort and completed her PhD in the Laser Analytics group of Prof. Clemens Kaminski. During a NERC policy internship with the Joint Nature Conservation Committee (JNCC) she discovered many similarities between Earth observation methods and the imaging techniques used in her PhD. She subsequently joined JNCC as an Earth observation specialist where she works with satellite imagery, aerial photography and LiDAR data to map habitats, assess their condition and detect changes. Using this information she provides scientific advice to the devolved governments in all four countries as well as to partners in the oversea territories and Latin America.
Sensing from Space: Using Earth Observation data to inform environmental policy
Prof. Shailaja Fennell - 1615
Prof. Shailaja Fennell is Professor of Economic Security and Resilience in the Department of Land Economy at the University of Cambridge. Her research interests include economic security and ecological change, community and national resilience, and institutional reform and regional transformation. She has a particular focus on local and sub-national decision making in rural and urban policy design, agricultural sustainability and food security; youth, migration and employment aspirations; provision of public goods in the spheres of education and health.
Student talks
(prizes sponsored by Marks & Clerk and Zimmer & Peacock)
Wadood Tadbier
Advancing Neurodegenerative Disease Diagnosis: Remote Gate Electrolyte-Gated Graphene Field Effect Transistors
Neurodegeneration signifies a persistent decline in neuronal function, characterised by irreversible damage, with Alzheimer’s Disease (AD) and Parkinson’s Disease (PD) being the most prevalent. The absence of a non-invasive, cost-effective, and portable diagnostic method impedes early detection efforts, delaying progress in treatment development. To address this challenge, we propose electrolyte-gated graphene field-effect transistors (EGFETs) as a promising solution, leveraging their exceptional surface area, high transconductance, sensitivity, and biocompatibility. However, the reproducibility of graphene and its stability in bio-interfacing environments pose significant obstacles to real-world applications. In this study, we introduce a remote gate EGFET design that isolates the sensing electrolyte from the operational electrolyte. By optimising the fabrication process, including graphene transfer, substrate selection, contact resistance minimisation, and bottom gate dielectric enhancement, we successfully reduce hysteresis and enhance stability. Our developed remote gate EGFET enables the utilisation of a single EGFET with multiple remote gates, eliminating the need for multiple fabrications and enhancing system reproducibility. Our results demonstrate that remote gate EGFETs offer stable, reproducible and portable solutions which can be used for neurodegenerative disease diagnosis, promising significant advancements in this field.
Anne-Pia Marty
Microscopes but Cooler: High-Resolution Microscopy for 0°C Imaging to Shed a New Light on Cold-Adapted Systems.
The Southern Ocean is one of the only places on earth where biological processes routinely happen below 0°C.[1] At these temperatures, our models for life fall apart. Diffusion, protein folding, energy and metabolism are all processes that are poorly understood in the cold, where dynamics are slow and energy is scarce.[2] This gap in knowledge is partly due to the lack of suitable equipment to study cold-adapted systems in physiological conditions. The low working distances required by high- and super-resolution microscopy meet issues of condensation and very fast heat transfer occurring between the microscope and the sample. My PhD aims to address this instrumentation problem by building tools for microscopy to perform high resolution imaging of live samples (primary cell lines of Antarctic fish) at 0°C. The development of this new method offers perspectives for studying the rate of molecular, cellular and sub-cellular processes. Observing cells from Antarctic animals in vivo reveals some surprising dynamics, in particular when comparing the size and shapes of cold adapted organelles, and the speed of their movement in the cold. This opens avenues to investigate how the energy budget of cold-adapted cells is spent, and what survival strategies have evolved from such a constrained environment. The applications of this microscopy methods span the fields of evolutionary biology, fundamental biophysics and biotechnology.
[1] J. Schram, K. Schoenrock et al., Ocean warming and acidification alter Antarctic macroalgal biochemical composition but not amphipod grazer feeding preferences. Mar. Ecol. Prog. Ser., vol. 581, p. 45-56, oct. 2017, doi: 10.3354/meps12308.
[2] Peck, Lloyd S. A cold limit to adaptation in the sea. Trends in Ecology & Evolution 31.1 (2016): 13-26.
Justas Brazauskas
Human-centric Design of Real-time Digital Twins
This research examines digital twins (DTs) in operational environments, focusing on intelligent edge AI sensors, their applications, and user group needs. Successful DT deployment depends on adapting to the end-user environment. This doctoral research project has three key contributions:
Digital Twin Construction: A dense sensor network with over 40 environmental sensors is deployed in a lecture theatre alongside the Cerberus privacy-preserving camera system. This network feeds into the Adaptive City Platform, providing real-time data with high spatial resolution. The digital twin includes 17 unique visualisations and a proprietary classification framework.
Visualisation Taxonomy: A taxonomy categorises DT visualisations by fidelity, timeliness, and data aggregation, helping designers leverage dense sensor networks across various applications. The taxonomy describes how sensor data management practices should respond to possible challenges at the application stage.
User Evaluation Case Study: Interviews with facility managers and building occupants assess user perceptions and co-design tools that democratise in-building data, aligning DT systems with user needs to improve efficiency and decision-making while maintaining focus on privacy.
Posters
(prizes sponsored by Marks & Clerk and Zimmer & Peacock)
Eleni Papafilippou
Characterising the rupture, fatigue and recovery of intercellular junctions using a stochastic bond model
Soft tissues' mechanical properties are increasingly recognized as a regulator of tissue homeostasis [1]. When tissues become injured or damaged, a cascade of cellular, molecular and mechanical responses is triggered, thereby extending the material's lifespan and maintaining its structural integrity even under continuous stress [2-3]. While it is understood that soft tissues are resilient to rupture, the mechanisms of load distribution, remodelling and healing remain unclear. This has led to increasing interest in characterising rupture and repair in soft tissues. A well accepted theoretical framework for studying intercellular junction dynamics was proposed by Bell in 1978, which assumes that cell adhesion is mediated by reversible bonds between surface molecules [4]. Here we use a stochastic model inspired by Bell’s work, to explore the force-dissociation kinetics of junction proteins. Our hypothesis is that cyclic loading can probe the reversible kinetics of individual bonds and predict the rupture and recovery dynamics of intercellular junctions. Our key findings indicate that (i) intercellular junctions have characteristic rupture and recovery timescales, (ii) cyclic loading perturbs these characteristic timescales and (iii) maps with universal features can be used to predict intercellular junction behaviours. Overall, this theoretical framework may be used to overcome the challenges of experimentally characterising the remodelling and recovery of active biomaterials.
[1] DuFort, C.C., Paszek, M.J. and Weaver, V.M., 2011. Balancing forces: architectural control of mechanotransduction. Nature reviews Molecular cell biology, 12(5), pp.308-319.
[2] Barcellos-Hoff, M.H., 1998. How do tissues respond to damage at the cellular level? The role of cytokines in irradiated tissues. Radiation research, 150(5s), pp.S109-S120.
[3] Bonfanti, A., Duque, J., Kabla, A. and Charras, G., 2022. Fracture in living tissues. Trends in Cell Biology, 32(6), pp.537-551.
[4] Bell, G.I., 1978. Models for the specific adhesion of cells to cells: a theoretical framework for adhesion mediated by reversible bonds between cell surface molecules. Science, 200(4342), pp.618-627
Hayley Gilbert
Using Lasers and X-rays to Study Defects in Next-Generation Solar Cells Materials
Mixed composition lead halide perovskites have emerged as a successful material in the development of solar cell devices. The tunability of the band gap (Eg) of these materials through compositional engineering is highly desirable for tandem devices, having enabled these devices to exceed power conversion efficiencies of 30% [1]. Increasing the Eg (by increasing the bromide content) is currently viable up to a threshold of ~1.75eV [2] due to halide segregation occurring upon light irradiation [3,4], which reduces the PCEs of devices [2,5]. Currently, the mechanism behind halide segregation is not fully understood [6].
To elucidate the mechanism of halide segregation in alloyed perovskites thin films, we used correlative hyperspectral photoluminescence (PL) microscopy and synchrotron X-ray mapping techniques. Techniques including X-ray fluorescence (XRF), X-ray diffraction (XRD), and X-ray absorption near edge spectroscopy (XANES) are used as a chemical probe into localised regions across the film and provide insight into chemical structure, environment, and compositional distribution. Through correlation with PL microscopy, which provides spatially resolved information on the optoelectronic performance, ion migration, and charge carrier trapping, a deeper understanding into key properties such as performance and longevity can be gained.
[1] J. Tong, Q. Jiang, F. Zhang, S.B. Kang, D.H. Kim, K. Zhu,Wide-Bandgap Metal Halide Perovskites for Tandem Solar Celss, ACS Energy Lett. 6 (2021) 232-248
[2] W. Yang, H. Long, X. Sha, J. Sun, Y. Zhao, C. Guo, X. Peng, C. Shou, X. Yang, J. Sheng, Z. Yang, B. Yan, J. Ye, Unlocking Voltage Potentials of Mixed-Halide Perovskite Solar Cells via Phase Segregation Suppression, Ad. Funct. Mater. 32 (2022) 2110698
[3] K. Frohna, M. Anaya, S. Macpherson, J. Sung, T.A.S. Doherty, Y. Chiang, A.J. Winchester, K.W.P. Orr, J.E. Parker, P.D. Quinn, K.M. Dani, A. Rao, S.D. Stranks, Nat. Nanotechnol. 17 (2022) 190-196
[4] E.T. Hoke, D.J. Slotcavage, E.R. Dohner, A.R. Bowring, H.I. Karunadasa, M.D. McGehee, Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics, Chem. Sci. 6 (2015) 613-617
[5] T. Leijtens, K.A. Bush, R. Prasanna, M.D. McGehee, Opportunities and challenges for tandem solar cells using metal halide perovskite semiconductors, Nat. Energy, 3 (2018) 828-838
[6] A. Kerner, Z. Xu, B.W. Larson, B.P. Rand, The role of halide oxidation in perovskite halide separation, Joule. 5 (2021) 2273-229
Matthew Ellis
Highly Tunable Ag Seed Mediated Synthesis and Characterisation of Ag-Au Alloy Nanoparticles Through Controlled Au Growth
Alloy nanoparticles containing a homogenous mixture of Ag and Au are increasingly sought after due their enhanced plasmonic and catalytic properties. Synthesis of these alloys is made challenging due to the difference in reduction potential between Ag and Au, which results in the galvanic replacement of Ag with Au. Co-reduction of Ag and Au in varying compositions is one of the more popular means of synthesising alloys but often this results in a lack in control of nanoparticle size and morphology. Seed-based approaches have the opportunity to overcome these limitations but also come with their own challenges and require robust experimental methods in order to create
nanoparticles of desired morphology. In our work, we explore the use of the surfactant CTAC and HEPES buffer to create two separate novel approaches for the directed growth and subsequent homogenisation of Au onto AgNP seeds of around 10 nm. The Au growth proceeds at room temperature, allowing for in-situ spectroscopic monitorisation and mechanistic insights
to be elucidated. Alloy homogenisation can be achieved using relatively low temperatures (70°C) within a few hours and allows for tunability of the LSPR across different ratios of Ag-Au. Our work aims to develop a greater understanding of the factors involved in controlling Au growth to make the synthesis of Ag-Au nanoparticles of defined size and composition more accessible to be used for a wide range of applications.
Melissa Watt
Quantitative spectroscopy in deep tissue for next generation wearables
From diagnostics to surgery and therapeutics, optical technologies have become ubiquitous in healthcare. Wearable technologies, or wearables, now allow continuous monitoring of vital signs and biomarkers using optical sensors, promising to facilitate proactive health monitoring, disease prevention, and management. Relying on mass-market light sources and sensors that probe visible wavelengths, current wearables are limited in their capabilities by probing only surface vasculature and are susceptible to measurement bias due to melanin absorption in the skin. Due to lower scattering, near and short-wave infrared (NIR and SWIR) wavelengths have been proposed to penetrate deeper into biological tissue, and could avoid confounding skin tone bias, but have yet to be used in commercial wearables.
Pigmented skin layer tissue phantoms were designed using a co-polymer-in-oil material formulation with tunable optical and acoustic properties, suitable for biophotonic and hybrid photoacoustic imaging applications. The absorption and scattering properties of the phantom materials were characterised at visible and NIR wavelengths with a double-integrating sphere system, providing the desired characteristics. Further work will explore the potential of these phantoms to optimise target wavelengths for wearable sensors and will explore strategies to equitably deliver the next generation of wearables.
[1] Hacker 2021 IEEE Trans. Med. Imaging 40(12) 3593–3603.
[2] Else 2023 J. Biomed. Opt. 29 S11506.
Sofia Kapsiani
Deep learning for enhancing fluorescence lifetime imaging microscopy in photon-starved conditions
Fluorescence lifetime imaging microscopy (FLIM) is a powerful optical technique that provides valuable insights into a sample’s molecular environment by measuring the time a fluorophore remains in its excited state before emitting a photon. FLIM reveals subtle changes in a fluorophore's microenvironment, including variations in temperature, ion concentrations, protein conformations, and molecular interactions, which are often undetected by intensity-based microscopy alone. However, the application of time-domain FLIM in high-throughput settings, such as drug screening, is hindered by the long acquisition times required to capture a single image. This limitation arises because traditional analysis methods, like multi-exponential fitting, rely on fluorescence lifetime decay curves with high photon counts for accurate lifetime calculations, which require longer data acquisition times to capture. Low photon count data often result in noisier signals, reducing lifetime estimation accuracy. In this study, we address this limitation by employing deep learning to predict pixel-wise fluorescence lifetime values from raw time-correlated single photon counting (TCSPC) FLIM data. Our results demonstrate that the deep learning model can accurately predict fluorescence lifetimes under photon-starved conditions while preserving both spatial and temporal resolution.
Sotirios Vavaroutas
Uncertainty Estimation for Sequence-to-Sequence Regression on Sparse Time Series
Machine Learning models typically assume that time series are regularly spaced, however this is often unrealistic in data from real-world sensors, where missing recordings are common. In this context, uncertainty estimates play a pivotal role, as they can enable confident and non-confident predictions to be distinguished. This work explores a novel uncertainty-aware sequence-to-sequence prediction method for sparse time series data. Specifically, it enhances the state-of-the-art evidential regression framework, widely used for uncertainty estimation, to handle missing sensors data. Following data imputation with an Akima spline-based method, the loss function of evidential regression is modified by assigning different weights to imputed and observed data points, to offer more reliable uncertainty estimates. Additionally, the proposal examines a variety of metrics for assessing the success of uncertainty estimations on sequence-to-sequence predictions, providing a reliable way to evaluate the models in a medical setting. The proposal is demonstrated in two real-world health applications, achieving improvements up to 30% in the accuracy of uncertainty-aware time series predictions.
