20505 research outputs found
Sort by
A Robot Locomotion Mechanism Powered by Wind Energy - WANDER-Bot
No full textThis report discusses the experimental development of WANDER-Bot: an early mechanical locomotion solution design concept for planetary exploration. It requires low-cost, low-power, low-storage, and is designed to be easily manufactured and assembled onsite using ISRU manufacturing methods. It is a design proposal which aims to address the existing limitations of space robotic missions and locomotion solutions: their high-cost, power budget constraints caused by degrading solar and RTG power sources, large storage requirement for transportation on the launch vehicle, high complexity, and difficulty to maintain and repair. Titan has been proposed as the future planetary environment exploration application for the system. The WANDER-Bot integrated system is composed of 3 main mechanisms: a Savonius turbine wind energy capture mechanism, reduction gearing mechanism to multiply torque, and a Jansen legged walking mechanism for locomotion. The robot uses simple mechanical linkages and assembly methods to lower cost and improve maintainability. Its 3D printable design allows future ISRU manufacturing potential, and its wind-powered driving mechanism offsets the electrical power demand for locomotion. Developed with consideration of planetary environments such as Titan, WANDER-Bot demonstrates the ability to walk forward in omnidirectional wind conditions, on rough, firm terrain at relatively low windspeeds, with a Titan-equivalent weight offset. This report covers the design evolution of the integrated robot system, through prototype iterations, of its individual constituent mechanisms. It documents the challenges faced, design improvements, and performance analysis through the experimental testing campaign. Experiments include walking gait analysis to ensure locomotion smoothness and stability, verifying motion simulations, turbine RPM performance at different windspeeds, and the self-start threshold windspeed for each integrated system iteration. It documents how these experimental findings informed design changes in future iterations. Successes and limitations of the design and testing is identified, along with future improvements to increase performance and environmental representation. The report concludes that this early WANDER-Bot prototype is not yet applicable for the environment in its current state, but the system demonstrates the feasibility of the low-cost, wind-powered, additively manufactured, easily maintainable locomotion solution. Future work is proposed to take this design further to develop maturity for the proposed application, such as steering mechanisms, variable speed control, and basic robot perception and autonomy for obstacle avoidance.MSc in Astronautics and Space Engineerin
Three-dimensional through-flow modelling of axial compressor rotating stall and surge.
Get PDFThe operation of an aero-engine is limited by the occurrence of compressor stall, and compressor performance is sacrificed to maintain a sufficient margin of operation. Compressor stall also plays an important part in the event of a shaft failure, and in determining if this will result in a rotor burst. In Cranfield University a tool to model the whole engine during shaft failure has been developed, but it requires the knowledge of the compressor performance during stall.
Low order 1D, 2D or 3D methods to model compressor stall exist in literature, but they are still at a low maturity level and not applicable for commercial use. The only methods available are expensive experimental testing and transient 3D CFD, which has unacceptable computational costs. The objective of this PhD project has therefore been identified in the development of a fast and robust 3D tool to model compressor stall, and in its validation with data from low-speed experimental rigs.
The tool created is a three-dimensional through-flow code which uses empirical correlations to model the blade row performance. A novel methodology has been developed to estimate the performance of blade rows in reverse flow, based on previous models of separated blade passages. The Godunov scheme has been chosen to create the 3D, unsteady, cylindrical, compressible, finite volume method Euler solver on which the code is based. Appropriate body forces and boundary conditions have been chosen and implemented.
The validation carried out on two low-speed compressors demonstrates the applicability of the proposed formulation, with successful prediction of the performance during reverse flow, rotating stall and surge. The speed, size and structure of the rotating stall cell have been successfully matched to experimental data. The developed tool can reproduce the forward flow, reverse and rotating stall regions of the map in less than 72 hours, at a computational speed unrivalled by modern commercial CFD codes.PhD in Aerospac
Multi-task deep learning for lung nodule detection and segmentation in CT scans
No full textThe early detection of pulmonary nodules in chest CT scans is critical for improving lung cancer outcomes. While existing computer-aided diagnosis (CAD) systems have shown promise, most treat detection and segmentation as separate tasks, leading to fragmented pipelines and limited representation sharing. This study proposes a 2.5D multi-task learning (MTL) framework that integrates both tasks within a unified Mask R-CNN architecture. The framework incorporates a tailored preprocessing pipeline—including Hounsfield Unit (HU) normalisation, CLAHE enhancement, and lung parenchyma masking—to improve input consistency and task-relevant contrast characteristics. To enhance sensitivity for small or ambiguous nodules, an auxiliary RoI classifier is introduced. Additionally, a nodule-level evaluation strategy aggregates slice-wise predictions across the z-axis, supporting a clinically meaningful assessment that approximates 3D diagnostic workflows. Experiments on the LUNA16 dataset demonstrate that the proposed framework achieves a favourable trade-off between detection and segmentation performance under a unified 2.5D multi-task setting. These results highlight the potential of integrated MTL approaches to advance CAD systems for early lung cancer screening.This research was funded by the Postgraduate Research & Practice Innovation Program of Jiangsu Province (grant number: SJCX24_0130).Electronic
Value co-creation in B2B industrial digital platforms: a platform-native perspective under service-dominant logic
No full textBackground Emerging B2B industrial digital platforms (IDPs) are vital enablers of digital transformation in traditional manufacturing. However, existing research has predominantly focused on IDPs led by industrial incumbents, leaving a gap in understanding IDPs led by platform-native firms. For such non-industrial network technology firms built around a platform business model from the start, value co-creation with industrial users possessing domain knowledge is essential to develop IDPs. Method Through the theoretical lens of service-dominant logic, we conducted a longitudinal, exploratory case study of Alibaba's Taofactory Platform. The study explores the dynamic process through which platform natives, as cross-boundary entrants in the B2B industrial market, develop IDPs via value co-creation. Findings Our analysis reveals that, driven by evolving service foci across stages, the platform owner orchestrates three core value co-creation strategies: resource integration, service exchange, and institutional arrangements. This orchestration fosters the evolution of value co-creation modes from connection-based to interaction-based and finally to collaboration-based, which in turn shapes the evolution of platform types from an industrial transaction platform to an industrial solution platform, and ultimately to an industrial platform ecosystem. Contributions This study offers a theoretical framework that advances research on digital platform development, value co-creation, and service-dominant logic. It also provides practical insights for platform natives aiming to enter the B2B industrial market.This research was funded by the National Natural Science Foundation of China under Grant 72192823 and Zhejiang University–The Hong Kong Polytechnic University Joint Center under Grant ZUPUC-2023-03INNTechnovatio
Using model compounds to show how a change in thinking is required for regulation of brominated haloacetic acids in drinking water
No full textBrominated disinfection by-products (Br-DBPs) are formed during the chlorination of drinking water where sufficient organic precursor molecules and bromide are present. They are of heightened concern due to their enhanced toxicity over the chlorinated analogues, yet despite this Br-DBPs are not as widely studied so their prevalence and capacity for formation is less well understood. Importantly, four brominated haloacetic acids (Br-HAAs) currently sit outside of regulatory standards. This raises the question of how thinking about DBPs will change if regulation expands to include all nine chlorinated and brominated HAAs (HAA9). The current work explored this question by utilising model compounds to establish the impacts associated with different groups of organic molecules and varying functional group on the ability to form Br-HAAs. This identified that high formation of unregulated HAAs was associated to precursors within all groups, with BDCAA (bromodichloroacetic acid) forming from the aromatic and amino acid-based groups, and BCAA (bromochloroacetic acid) forming from within the aliphatic, amino acid-based and lignin-derived groups. Critical structural insight was related to the balance of TXAA (tri-halogenated acetic acids) and DXAA (di-halogenated acetic acids) which was associated with stabilisation of the di-halogenated intermediates. Consequently, thinking about formation of the currently unregulated brominated HAAs needs to consider both hydrophobic and hydrophilic precursor types, appreciating that enhanced formation propensity can originate from small changes in molecular structure. Furthermore, variation in formation pathways and bromination preference between the THMs, DXAAs and TXAAs dictates that monitoring and management of the HAA9 will require divergence from current DBP strategies.The work was supported by UK Water Industry Research Ltd. (UKWIR) [TX/05/A/204] and Engineering and Physical Sciences Research Council (EPSRC) through their funding of the Water Infrastructure and Resilience (WIRe) Centre for Doctoral Training [EP/S023666/1].Journal of Hazardous Material
The role of clothing brands’ brand Image in fashion style perspective in building consumers’ brand preference - A research on fashion self-congruity of different brand images towards consumers’ brand preference with a mediating role of social self-esteem in Indonesian young adults
No full textThis study investigates how clothing brand image influences brand preference among Indonesian young adults, focusing on the roles of fashion self-congruity and social self-esteem. Clothing has long been recognised as both functional and symbolic, shaping identity and social belonging. Grounded in self-congruity theory, enclothed cognition, and sociometer theory, the research examines how ideal fashion self-congruity and ideal social fashion self-congruity interact with social self-esteem to predict brand preference across four fashion style categories: athleisure, streetwear, vintage academia, and minimalist. A quantitative, cross-sectional survey was conducted with 145 respondents aged 18–26, recruited through purposive and convenience sampling. Participants completed validated scales for fashion self-congruity, social self-esteem, and brand preference, all demonstrating strong reliability. Data were analysed using multiple regression and Hayes’ PROCESS macro model 4 to test direct and mediating effects. The results reveal that the influence of self-congruity varies by style. For athleisure and streetwear, self-congruity does not directly affect brand preference but operates fully through social self-esteem. In vintage academia, ideal fashion self-congruity significantly predicts brand preference directly, while ideal-social works indirectly. For minimalist fashion, both dimensions of self-congruity significantly predict brand preference, with partial mediation through social self-esteem. Across all styles, social self-esteem consistently emerges as both a mediator and a direct predictor of brand preference. Theoretically, the study refines self-congruity theory by showing style-contingent pathways and confirms the role of social self-esteem as a psychological mechanism linking clothing to consumer-brand relationships. Practically, the findings suggest that brands should tailor their positioning to the symbolic meanings of different fashion styles while leveraging clothing’s power to enhance consumers’ social confidence.MSc in Strategic Marketin
Does ESG Performance Affect the Cost of Debt? A Multi-Model Analysis Incorporating Debt Maturity, ESG Data Sources, and Moderating Governance Signals
No full textThis article examines how environmental, social and governance (ESG) performance influences the cost of debt for publicly listed firms in Europe between 2012 and 2023. Drawing on data from both Refinitiv and Bloomberg and using composite and pillar‑level scores to limit provider bias, we estimate fixed‑effects regressions, two‑stage least squares (2SLS) models and dynamic system generalised method of moments (GMM) to address endogeneity and path dependence. Our results show that static models associate higher ESG scores with higher borrowing costs, suggesting creditors view sustainability expenditure as an immediate cost. By contrast, dynamic GMM models reveal that lagged ESG performance reduces the cost of debt substantially, implying that reputational and risk‑mitigating benefits accrue over time (Cheng et al., 2014; Flammer, 2021). Pillar analyses demonstrate that social performance consistently lowers borrowing costs, while environmental and governance effects are sensitive to the estimation method: environmental initiatives increase borrowing costs in static models but become beneficial under GMM, whereas governance initiatives raise costs across all specifications. Moderation analyses indicate that transparent CSR reporting and audit independence amplify ESG’s benefits, whereas state ownership and large firm size weaken them. Robustness checks using a sovereign spread proxy confirm that the results are not an artefact of the accounting measure. These findings highlight the dynamic, dimension‑specific and context‑dependent nature of ESG’s financial materiality and offer practical insights for managers, creditors and policymakers.MSc in Financ
AI-driven 5G networks for autonomous positioning system platform
No full textUnmanned Aerial Vehicles (UAVs) are becoming essential for various urban applications, such as surveillance, delivery, logistics, disaster management, and traffic
monitoring. However, their positioning performance in urban environments can be
limited due to challenges such as non-line-of-sight (NLOS) propagation, multipath
interference, and signal blockage caused by tall buildings, trees, and other obstacles.
These factors lead to reduced positioning accuracy and unreliable communication. To
address these issues, this thesis introduces three key and novel contributions. First,
it presents one of the first real-world evaluations of the 5G network performance for
UAV operations at altitudes between 50 and 110 meters, using XCAL-based field trials. This provides new insights into the altitude-dependent Quality of Service (QoS)
parameters such as latency, throughput, and handover (HO) efficiency and provides
practical recommendations for UAV-specific connectivity protocols. Second, a novel
hybrid positioning framework is proposed that integrates the observed time difference of arrival (OTDOA) of the new 5G radio (NR) with the fusion of sensor and
barometric pressure sensor through an Extended Kalman Filter (EK). This combination significantly improves positioning accuracy (2.8–7 m) in GNSS GNSS-challenged
urban environment, which has not been demonstrated in prior UAV studies. Third,
the thesis introduces a lightweight feedforward neural network (FNN) for mitigating
NLOS errors in 5G-based UAV positioning. Trained on simulated MATLAB data, the
model corrects time-of-arrival (TOA) measurements in real time, reducing positioning error to 1.3 m in LOS and 1.7 m in NLOS, outperforming conventional methods.
Unlike existing solutions, this model is designed for real-time deployment on UAV
platforms with limited resources. Overall, this research strengthens UAV navigation
and connectivity in urban airspace by combining 5G advancements, sensor fusion,
and AI-powered error correction. The novelty lies in the integration of real-world 5G
performance analysis, a hybrid OTDOA sensor fusion framework, and an AI-based
NLOS correction model into a unified solution for reliable, accurate, and scalable Urban Air Mobility (UAM), opening the door to future improvements in AI-driven 5G
networks for autonomous system platforms.PhD in Aerospac
Dataset: Using model compounds to show how a change in thinking is required for regulation of brominated haloacetic acids in drinking water
No full textThis project looks at the Disinfection By-Products (DBPs) formed from model compounds during their chlorination with bromide present. This data set is the results obtained from the analysis for trihalomethanes (THMs) and haloacetic acids (HAAs).UK Water Industry Research (UKWIR)EP/5023666/
Dataset: Cladding fragment impact prediction using material properties
No full textTerrorist attacks involving improvised explosive devices (IEDs) have more recently included fragmentation such as nuts, bolts and ball bearings. Under theTerrorism (Protection of Premises) Act 2025, building owners must incorporate public protection procedures and to do so correctly they must understand how the building would perform against an attack. As more external cladding products come onto the market ranging from ceramics and timber products tofibre cement boards and high-pressure laminates, there is a risk that not all cladding will protect occupants from blast driven fragmentation. This research assesses the suitability of standard laboratory material tests to predict the V50 ballistic limit, a measure of penetration resistance. Plastic impact testing wasnot suitable for the full set of materials, flexural strength showed moderate correlation and two types of hardness testing Mohs and Shore D were the most accurate predicters when combined with material groupings (ceramic, fibrous and non-fibrous). A validation exercise with seven additional non-ceramic products demonstrated Shore D prediction errors of up to 26%. This supports the potential utility of the simple test for comparing and estimating penetration resistance of external cladding products. It is recommended that further testing on a wider variety of materials is completed to build the dataset and refine the regression equations.British Arm