20505 research outputs found
Sort by
Swarm drones - efficient machine learning and informatics
No full textIn 2020, worldwide consumer drone unit shipment was 5 million, which is expected to be
9.6 million in 2030. This generates a global drone market with 26.3 billion USD in revenue. The popularity of drones in civilian and professional environments has changed the
way humans live and work. However, they have also brought new challenges and threats
to the environment and society (e.g., property and personal damage caused by inappropriate drone operations or hostile drones) which requires more robust regulatory mechanisms and advanced technologies of drones. Supervision of tiny shape, high-mobility
drones by humans is inefficient and inaccurate, whereas Artificial Intelligence (AI) methods, especially deep learning (DL) show potential in drone detection, classification and
tracking. However, as data-driven models, the performance of DL models is decided by
the quality of data and model structure. At the same time, the structural complexity of
black-box DL models affects their explainability and energy efficiency. These factors affect the willingness of people to trust DL models. Therefore, this thesis aims to analyze
the impact of drone informatics on DL behaviour, and achieve more efficient training of
high-trustworthiness DL models by efficient drone informatics. This will require research
on DL explainability, trustworthiness and efficiency. The aforementioned researches are
all interrelated and highly relevant to the data.
In this thesis, firstly, explainable AI (XAI) and DL trust factors are reviewed. A
theoretical DL trustworthiness metric Quality of Trust (QoT) and a lifelong AI trust-
worthiness supervision protocol are proposed. Secondly, a novel partially explainable
Gaussian-process-based neural network structure is proposed. Compared with conventional machine learning methods, it is more transparent and without any sacrifice in accuracy. Thirdly, a GAN-TDA method is proposed to analyze the learning efficiency of
convolutional layers on drone images and guide the collection of new data. Collecting
new data with direction could boost the DL model performance more efficiently in time
and cost. Fourthly, a transistor operations (TOs) model is proposed to analyze the DL
energy consumption scaling law to different model architectures and settings. Finally,
a physical visual neural stealth drone canopy is designed with the hard-to-learn design
features analyzed by GAN-TDA and painted with adversarial evasion features to escape
DL drone detection and classification. The canopy design method is further extended to
swarm drone scenarios.
This thesis shows: 1) both model explainability and performance are related to DL
trustworthiness, and need a trade-off according to the QoT of different tasks; 2) combining
human-understandable efficient drone informatics and the understanding of DL energy
scaling laws can find high-efficiency datasets and network structures, resulting in efficient
DL models with high trustworthiness; 3) The above knowledge can be used to formulate
attacks on drone-related DL models to reduce their trustworthiness.PhD in Aerospac
Brief Communication: measurement of boundary layer turbulent transition using an acoustic microphone unit
No full textIn this article the transition of a laminar boundary layer (BL) over a flat plate is characterized using an acoustic technique with a pitot probe linked to a microphone unit. The probe was traversed along a BL plate at a fixed wind tunnel flow velocity of 5.5 m/s. A spectral analysis of the acoustic fluctuations showed that this setup can estimate the streamwise location and length of the BL transition region, as well as the BL thickness, by using the intermittency similitude approach. Further work is required to quantify the uncertainty caused by signal attenuation within the data acquisition system.SAE International Journal of Aerospac
Novel combustion flame luminosity methodology to track gun propellant combustion progress in closed vessel and semi-closed vessel
Get PDFAkhavan, Jacqueline - Associate supervisorThis research project investigated a novel technique aiming to evaluate the performance of gun propellant combustion. The voids between propellant particles cause some difficulties in accurately measuring the performance of powder propellants. Closed vessels and strand burners are methods capable of measuring propellant burn rate, however, cannot consistently control the non-linear progression of gun propellants due to the voids. Combustion models have been employed to solve this problem; however, several models still produce unrealistic predictions, hence the necessity for an experimental technique.
The combustion parameters commonly associated with propellant combustion were recorded and their correlation with the luminosity was investigated. A novel technique was designed intended to contribute to the knowledge obtained during
certain combustion processes. The luminosity generated by the combustion flame is associated with the chemistry of the propellants and can be measured by cameras from safe distances. A camera was utilised to record the luminosity generated by the combustion and the data was compared to the pressure and combustion time. Several modes of operation were utilised in the form of closed vessel, vented vessel, and open vessel to study the difference between these modes on the parameter’s correlation.
It was observed that the luminosity data was closely correlated to the pressure, especially the peak luminosity to the peak pressure, obtained when the burn rate was at its highest conditions. Despite individual variation between firings, the average luminosity for any propellant loading density generated a reproducible and consistent rise, indicating the correlation hypothesised exists.
Despite making useful observations, the luminosity technique requires improvements to more accurately describe the correlations. The usefulness of this technique does rely on the ability of the propellant to produce a smokeless burn, generally a requirement for several military weapon systems for safety purposes
Restoration of ecological interactions: the influence of site and landscape factors
No full textRestoration has been extensively used in agricultural landscapes as a mitigation measure to reduce biodiversity loss in response to historic habitat destruction. Trophic interactions between insects and plants underpin key ecosystem processes and contribute to system robustness, which is a critical outcome for habitat restoration. We evaluate how restoration age, site size and landscape proximity to similar habitats impact the re-establishment of trophic linkages between empirically measured grassland plant-pollinator (60 sites; 1–76 years) and woodland plant-herbivore networks (60 sites; 13–67 years). In each case, sites were selected along a chronosequence with the goal of maximising variation along these temporal and spatial gradients. For both grassland and woodlands, older and larger sites typically support higher levels of connectance, nestedness and generality of the networks. In contrast, landscape proximity promotes these metrics for woodland webs but has the reverse effect for grassland webs. The similarities show common characteristics of community trophic re-establishment in response to local environmental drivers for these different ecosystems. Focusing on interactions rather than species identity highlights opportunities for targeted policies to restore ecosystem function in wider agricultural landscapes; for example, through increasing site size as well as the need for continuity of older sites.UK Research and Innovation Natural Environment Research CouncilThe research was funded under the NERC consortium award ‘Restoring Resilient Ecosystems’. Grants were NE/V006525/1 to JMB and BAW, NE/V006444/1 to JH and RC, NE/V006460/1 to EFM, KP and KW.Agriculture, Ecosystems & Environmen
Visible-near infrared spectroscopy and near-infrared hyperspectral imaging for the detection of T-2 and HT-2 toxins in individual oat grains
No full textOat grains are increasingly consumed worldwide due to their health benefits, yet they are highly susceptible to contamination by Fusarium toxins, particularly T-2 and HT-2 toxins (T-2+HT-2). These toxins pose serious health risks and are unevenly distributed, with a few highly contaminated grains often driving a batch over legal safety limits. Current detection methods are destructive, slow, or inadequate for detecting contamination at the individual grain level. This study is the first to demonstrate the potential of visible–near-infrared (Vis-NIR) spectroscopy and near-infrared hyperspectral imaging (NIR-HSI) to detect T-2+HT-2 in individual oat grains non-destructively. 200 grains were scanned, and their toxin content quantified by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Classification models were developed to identify grains exceeding both the European Union (EU) legal threshold (1250 μg/kg) and a higher risk level (10,000 μg/kg). Both techniques achieved high accuracy (up to 94.5 %) in identifying contaminated grains. Key wavelengths were identified (e.g., 1203, 1419, 1424 and 1476 nm in NIR; 440–455 nm in Vis), and reducing the model to 20 wavelengths preserved performance while simplifying computation. Critically, removing just 21.5 % of the most contaminated grains could reduce overall toxin levels by over 95 %. Moreover, sampling simulations revealed that analysing 30 % of grains guarantees detection of contamination above legal limits, whereas 0.5 % sampling yields only a 25–33 % detection chance. These findings highlight a feasible path for integrating spectroscopic screening into industrial oat sorting lines, improving food safety, reducing economic losses, and overcoming key limitations of conventional mycotoxin monitoring.This work was supported by the Spanish Ministry of Science and Innovation (predoctoral grant FPU21/00073 and Project PID2020-114836RB-I00 funded by MCIN/AEI/10.13039/501100011033) and Cranfield University. The authors would like to thank Derek Croucher from Morning Foods for providing the contaminated oats samples for analysis.Food Contro
Design and optimisation of a high performance lightweight monoblock cast iron brake disc
No full textIn the past few decades, part weight reduction that would yield lower fuel consumption and decrease CO₂ emissions has been one of the main challenges in the automotive industry. In brake rotors, the downsizing is even more difficult because mass reductions may result in lower thermal capacity which will compromise rotors performance. This research work focuses on developing a design methodology for a novel concept of a lightweight monoblock brake disc with fingered hub. Discs of such a design can feature reduced mass, lower cost, improved cooling and are particularly suitable for Electric and Hybrid Electric Vehicles.
The described methodology comprises a number of steps, each investigating a different attribute of this novel disc design and ensuring its feasibility. The use of Finite Element Analysis was incorporated to study the mechanical and thermomechanical loading on the disc with the results indicating acceptable behaviour. Physical testing with an actual prototype disc confirmed the findings of the modelling work. Thereafter, structural optimisation was conducted where significant reduction in mechanical stress was achieved and also further weight reductions. Finally, in a newly developed experimental facility cooling tests were conducted that revealed substantially improved cooling characteristics for the brake disc with fingered hub when compared to the baseline solid hub disc.PhD in Transport System
A lived experience assessment of public–private partnerships delivering rural water services in Rwanda
No full textPerformance measures of public–private partnerships (PPPs) for water services provision have tended to focus on countable features of the physical infrastructure, ignoring the lived experience of using the water. Using mixed methods, we explore how the PPP model has performed across two managed systems in rural Rwanda using both technical functionality measures and metrics that reflect the experiences of the intended beneficiaries. Findings evidence mixed system performance against both sets of metrics, underpinned by a lack of strong accountability measures. The paper offers actionable evidence in support of policymakers' efforts to design more equitable and impactful PPP schemes.Journal of Water, Sanitation and Hygiene for Developmen
Individual resilience in a volatile work environment
No full textThis chapter is set out to understand the way in which individuals build and maintain their resilience when facing a volatile work environment. We studied this question qualitatively in an S&P 500 company where the company faced major business changes during a time when employees faced work-related restrictions. We described practices that individuals do to maintain or inhibit their psychological resilience. Interestingly, we found that individuals were able to adopt positive framing of the situation across tenure, seniority levels, and involvement in the business change. We also found that the inhibitor for individuals to reframe their situation was related to being unable to rest. Increased working hours and a lack of work-life boundaries inhibited individuals’ pathways to resilience. Building on the findings, the chapter will conclude with managerial implications. First, the findings highlight that both direct managers and senior managers play an important role in shaping individuals’ meaning-making processes, and their narratives could encourage subordinates to demonstrate optimism around the challenge. Second, organizations can consider interventions such as instituting long and short breaks, and mindfulness training to mitigate the negative impact from lack of rest
FinTech and financial inclusion: unpacking the links to inequality
No full textThis investigation analyses how financial inclusion, dimensions of inequality and financial technology (FinTech), are related across 113 countries using data from the 2011–2021 Global Findex survey. The study finds both direct and indirect associations between the different constructs employing structural equation modelling methodologies. The primary findings reveal that FinTech simultaneously worsens income inequality while narrowing gender gaps through enhanced financial inclusion. Additionally, these technologies promote broader financial participation among excluded sections of the population. Greater financial inclusion lowers both economic and gender-based inequalities. The research also explains that effective regulation, educational opportunities, and availability of credit can contribute to more equitable outcomes through financial inclusion. These findings add to emerging scholarships on how financial technological innovation and robust financial inclusion initiatives can improve resource allocation and promote more inclusive economic development globally. This represents the first comprehensive examination of multiple inequality types and their complex relationships within this framework. The study provides preliminary evidence of the varying distributional implications of technology-based finance and financial inclusion on several types of inequalities.Artha: Journal of Business and Financ
Pathway to non-intrusive in-flight flow diagnostics with filtered Rayleigh scattering
Get PDFA pathway to in-flight application of Filtered Rayleigh Scattering (FRS) is herein presented, including a viable concept, based on recently published related work. The proposed pathway considers the key technical, operational and regulatory challenges to enable in-flight measurements using FRS for inlet flow distortion characterisation ahead of the aero-engine. Solutions to these challenges are proposed, in particular methods for light delivery, flow imaging and integration of the measurement system in the in-flight environment. This complements the experimental lab-scale demonstration of a FRS concept for flow distortion measurements and provide a route for further exploitation as a diagnostic tool for the next aircraft generations.The SINATRA project leading to this publication has received funding from the Clean Sky 2 Joint Undertaking (JU) under grant agreement No. 886521. The JU receives support from the European Union’s Horizon 2020 research and innovation programme and the Clean Sky 2 JU members other than the Union.SAE International Journal of Aerospac